Comparative performance and external validation of three different models in predicting inadequate bowel preparation among Chinese inpatients undergoing colonoscopyFirst Name Last

Comparative performance and external validation of three different models in predicting inadequate bowel preparation among Chinese inpatients undergoing colonoscopyFirst Name Last | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Comparative performance and external validation of three different models in predicting inadequate bowel preparation among Chinese inpatients undergoing colonoscopyFirst Name Last Qi-yu Sun, Mei Zhang, Fei-yan Feng, Ding Min, Yue-hong Shen, Yu-feng Ou, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8265970/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background and Aims Inpatient colonoscopy frequently fails because bowel cleansing is inadequate. Magnesium sulfate is the cathartic of choice in most Chinese hospitals because it costs only a few cents, yet no externally validated tool exists to flag inpatients at high risk of poor preparation. We compared three in-house prediction models to identify the one that best helps nurses and physicians optimize bowel preparation before colonoscopy. Methods Using three previously derived models—Model-1 (seven static variables), Model-2 (nine two-stage variables), and Model-3 (twelve integrated variables)—we conducted a prospective cohort study. Consecutive inpatients aged ≥ 18 years who received a split-dose magnesium sulfate regimen for elective colonoscopy between July 2024 and March 2025 were enrolled. Inadequate bowel preparation was defined as a total Boston Bowel Preparation Scale score < 6 or any segment score < 2. Discrimination (area under the ROC curve, AUC), calibration (calibration plot and Hosmer-Lemeshow test), and clinical utility (decision-curve analysis) were evaluated. Results Among 977 patients, 107 (10.95 %) had inadequate preparation. Model-3 achieved an AUC of 0.785 (95 % CI 0.746–0.824), outperforming Model-1 (AUC 0.679) and Model-2 (AUC 0.681) by 0.106 (P < 0.001). At the optimal cut-off, Model-3 provided 83.8 % sensitivity, 60.6 % specificity, and 79.7 % accuracy. Calibration was good (Hosmer-Lemeshow P = 0.286), and decision-curve analysis showed the widest net-benefit range (threshold probability 0.401–0.982). Conclusions In hospitalized patients receiving magnesium sulfate, Model-3—combining baseline risk factors with real-time nursing assessments—offers superior discrimination, calibration, and clinical utility. Embedding this simple score in the electronic health record flags high-risk patients early, triggers tailored education, and improves preparation quality without added cost. Multi-centre implementation studies are warranted to confirm generalizability. Colonoscopy bowel preparation predictive score Figures Figure 1 Figure 2 Figure 3 Figure 4 1 Introduction Global data show that both the incidence and mortality of colorectal cancer (CRC) remain unacceptably high[ 1 , 2 ]. According to GLOBOCAN 2018, CRC ranks as the third most lethal and the fourth most commonly diagnosed malignancy worldwide[ 3 ]. The American Cancer Society estimated that, in 2020 alone, approximately 147,950 Americans would be newly diagnosed with CRC and 53,200 would succumb to the disease [ 1 ]. Of particular concern is the striking rise in early-onset CRC among individuals younger than 50 years across diverse populations[ 4 , 5 ] . China—the world’s most populous country—carries a disproportionate share of the colorectal cancer burden. Although its age-standardized incidence and mortality rates are lower than the global average, the absolute numbers of new cases and deaths are the highest worldwide because of its vast population [ 6 ]. More than half of all global CRC diagnoses and related deaths are estimated to occur in China [ 7 ]. In 2020 alone, 555,480 Chinese residents were newly diagnosed with CRC, corresponding to an age-standardized rate of 23.9 per 100,000 [ 8 ]. Between 2005 and 2015, both the incidence and mortality of CRC in China increased steadily [ 9 ]. Data from Shanghai registries reveal a continuous rise in CRC incidence and mortality from 2002 to 2016[ 10 ]. Moreover, Chinese patients typically develop CRC approximately one decade earlier than their Western counterparts; the median age at diagnosis is 58 years[ 11 ]. Widespread adoption of endoscopic techniques has established colonoscopy as the gold standard for screening, diagnosing, and treating colorectal disease[ 12 ]. The procedure is pivotal to improving CRC survival, yet its success hinges directly on the quality of bowel preparation. Adequate cleansing markedly increases polyp and lesion detection while reducing procedural complications[ 13 , 14 ]. High-quality preparation is therefore the cornerstone of effective colonoscopy, governing adenoma detection rates and the subsequent efficacy of surveillance and therapy[ 15 , 16 ]. Colonoscopy is the cornerstone of colorectal-cancer screening and diagnosis, yet its efficacy and safety hinge on adequate bowel cleansing. International guidelines stipulate a minimum preparation success rate of ≥ 90% as a key quality indicator[ 12 , 17 ]. Inadequate bowel preparation (IBP)—defined as residual stool or fluid that impairs mucosal visualization—remains a common quality deficit[ 18 – 20 ], reducing adenoma detection, prolonging procedure time, increasing complication risk and cost, and being particularly prevalent among in-patients. Cleansing is conventionally graded with instruments such as the Boston Bowel Preparation Scale (BBPS) [ 21 ]. Multiple patient-related [ 22 – 24 ], disease-related, regimen-related [ 25 ] and health-system factors interact to perpetuate IBP, which still complicates 20–25% of examinations[ 26 ], with some series reporting rates as high as 44%[ 27 ]. Prospective US data document an IBP rate of 20.9%[ 28 ], and a recent Chinese survey observed an unacceptable 23.86% inadequacy rate[ 29 ]—well below the 90% benchmark. Numerous studies concur that in-patients achieve substantially lower cleansing quality than out-patients[ 30 – 32 ].. IBP exerts a cascade of adverse effects: it reduces diagnostic accuracy, increases procedural complications, and drives up healthcare costs. Residual stool obscures the mucosa, slows cecal intubation, and hides polyps and cancers, leading to missed lesions [ 33 , 34 ]. Poor cleansing prolongs inspection time, heightens technical difficulty, and elevates the risk of perforation and other adverse events[ 14 , 35 ]. When the first examination is compromised, patients must return for a repeat colonoscopy or undergo alternative, often more expensive, diagnostic tests, amplifying psychological distress and institutional workload[ 36 , 37 ]. Consequently, systematic identification of IBP risk factors and the development of effective, targeted interventions are essential to improve colonoscopic quality and efficiency. To tackle this problem, investigators have developed a range of risk-prediction algorithms that flag patients at high risk of IBP before colonoscopy, thereby enabling targeted interventions to improve cleansing quality. Published models include traditional logistic-regression equations and easy-to-use scoring systems as well as machine-learning approaches that capture non-linear relationships and high-dimensional interactions[ 38 – 43 ]. Examples include a validated geriatric-specific risk score [ 44 ], the questionnaire-based PREPA-CO prediction tool[ 45 ], and several machine-learning classifiers that outperform conventional statistics in internal validations[ 46 , 47 ]. Although these efforts are promising, none has been prospectively tested in the magnesium-sulfate inpatient population treated at Zhongshan Hospital, a national referral centre for digestive diseases. Equally important, most existing models rely solely on baseline characteristics or pre-laxative variables and ignore the pivotal role of nursing staff in real-time education, adherence coaching, and symptom management during the preparation process. Nurse-led educational interventions have been shown to markedly improve patients’ acceptance of the salty-bitter taste of magnesium sulfate and, more importantly, their actual adherence to the split-dose regimen—benefits that are especially pronounced among elderly, cognitively impaired, or poly-medicated inpatients[ 48 ]. Bedside nursing assessments capture real-time information on taste tolerance, stool quality, and adverse events (nausea, bloating, cramps), enabling immediate corrective actions: on-the-spot re-education, psychological support, extra fluids, or timely anti-emetics. This dynamic loop has been proven to cut IBP rates. Yet contemporary prediction models continue to ignore nursing variables—taste acceptability, adherence checks, stool reports—and overlook nurses as the pivotal link between risk prediction and in-process intervention. Consequently, these models offer limited clinical traction and fail to guide frontline staff on when and how to intervene. Commonly used bowel-cleansing agents before colonoscopy include polyethylene glycol (PEG), magnesium sulfate, and sodium phosphate. PEG is considered to have superior efficacy and safety, but its large volume often leads to poor tolerability, especially in older adults, which can reduce adherence. Split-dose regimens may partially address this issue; however, many patients are unwilling to wake up during the night to take the second dose. Moreover, the higher cost of PEG, limited availability, and greater economic burden on patients make it difficult to widely implement in remote or resource-limited settings. As the world’s largest developing country, China faces uneven distribution of medical resources. In areas with limited health insurance coverage or high out-of-pocket expenses, cost-effectiveness becomes a key factor in selecting bowel preparation agents. Magnesium sulfate remains widely used due to its low cost, small volume, and easy accessibility, offering a significant economic advantage. At our institution, simethicone is routinely co-administered to reduce foam and improve visualization, partially compensating for the limitations of magnesium sulfate and balancing cost-effectiveness with clinical outcomes. However, magnesium sulfate has a bitter taste and frequently causes nausea and vomiting, which further increases the risk of IBP. Unfortunately, most existing prediction models were developed based on Western outpatient populations using PEG-based regimens. These models do not account for taste-related feedback, fluctuating adherence, or the potential impact of nursing interventions specific to magnesium sulfate use. Additionally, Chinese populations may differ from Western populations in physiological characteristics, dietary habits, lifestyle, and medication adherence. Inpatients, in particular, are often affected by polypharmacy and reduced mobility, all of which may influence model performance [ 49 ] . As a result, the variable weights and cut-off values derived from PEG-based models cannot be directly applied to the magnesium sulfate inpatient setting. Furthermore, most models rely solely on static information available at admission, ignoring real-time patient feedback after medication intake, which limits their sensitivity and ability to support dynamic clinical management. Since nurses are the primary providers of bowel preparation education and play a central role in assessing and managing patient adherence, there is an urgent need to explore how nursing-related variables can enhance predictive performance. To address these gaps, our research team previously used a cohort of 806 in-patients who received oral magnesium sulfate at the Endoscopy Center of Zhongshan Hospital, Fudan University from January to June 2024 to build three prediction models: a single-stage score (model-1) and two two-stage dynamic models (model-2 and model-3). The first-generation model-1 applied LASSO-logistic regression to seven static variables—IBD, cardiac or renal insufficiency, electrolyte disturbance, relevant medication history, last meal before bowel preparation, laxative preference, and previous failed preparation—to generate a simple integer score; it achieved an internal AUC of 0.836 and allowed rapid screening but could not capture post-dose changes. Model-2 retained six of these variables (electrolyte disturbance was removed) and introduced a “before versus after dosing” framework by adding three dynamic indices collected four hours after the first dose: taste acceptability, adverse events (nausea, vomiting, abdominal pain), and self-reported adherence (whether the full laxative dose was taken). Building on model-2, model-3 added neurological comorbidity, colonic stricture, and last stool character, yielding a final -variable combination of “static + dynamic + nursing assessment”. In internal validation, the AUC increased from 0.693 with model-2 to 0.719 with model-3, sensitivity improved markedly, and both calibration and net clinical benefit remained favorable, indicating that real-time information provides substantial predictive gain. Even when prediction models perform well in their derivation cohorts, differences in study populations—such as ethnicity, geographic region, or healthcare system—and variations in data collection can erode their accuracy in independent external datasets[ 49 – 51 ]. Moreover, the single-centre, retrospective design of our preliminary work may have overestimated performance and limited wider applicability. Prospective external validation is therefore essential to establish a model’s reproducibility and generalisability, and constitutes the critical step toward safe clinical implementation and broader dissemination[ 52 – 54 ]. Accordingly, we designed a prospective, continuously-enrolled external-validation study that builds on our earlier methodological iterations. By independently comparing the discrimination, calibration, and net clinical benefit of model-1, model-2, and model-3 in a real-world inpatient population at our own institution, the investigation seeks to answer which version is best suited to our magnesium-sulfate setting and to identify the optimal IBP risk-assessment tool for hospitalized adults undergoing colonoscopy. Beyond providing clinicians with an evidence-based instrument, the project will give nurses a quantifiable metric for early recognition of high-risk patients and for tailoring bedside interventions. Rather than simply testing three models in parallel, the study specifically examines whether model-3—through its three dynamic nursing indicators (taste acceptability, self-reported adherence, and last stool character)—achieves clinically meaningful gains in discrimination, calibration, and decision-curve net benefit over model-1 and model-2. Embedding these nursing assessments and real-time response data into the final algorithm should fill a key research gap and establish the clinical utility and scalability of an IBP prediction model that integrates proactive nursing care into routine magnesium-sulfate bowel preparation. 2 Methods 2.1 Study design and model derivation history This prospective, single-centre cohort validation study was designed to conduct an external evaluation of three previously developed prediction models for IBP following a magnesium-sulfate regimen (model-1, model-2, model-3). All procedures were conducted in strict accordance with the TRIPOD (Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis) statement and the PROBAST (Prediction model Risk Of Bias ASsessment Tool) guidelines to guarantee methodological transparency and full reproducibility of the findings. 2.2 Data source and ethics Consecutive in-patients scheduled for colonoscopy at the Endoscopy Centre of Zhongshan Hospital, Fudan University, were prospectively enrolled from July 2024 to March 2025. The study protocol was approved by the hospital ethics committee (ref. B2024-278R) and registered at the Chinese Clinical Trial Registry (ChiCTR2400094906). Written informed consent was obtained from every participant. All data were extracted from the hospital’s electronic medical-record system, de-identified, and stored in an encrypted institutional database to guarantee patient confidentiality and data integrity. 2.3 Participants Eligibility criteria were: (1) age ≥ 18 years; (2) scheduled for elective colonoscopy; and (3) completed the institutional split-dose magnesium-sulfate bowel-preparation protocol while hospitalized. Exclusion criteria comprised: (1) pregnancy, severe cognitive impairment, or active psychosis; (2) incomplete colonoscopy or preparation terminated for reasons other than quality; and (3) > 5% missing values for key predictors. 2.4 Outcome definition The primary endpoint was IBP. Preparation quality was quantified with BBPS [ 21 ]. The BBPS divides the colon into three segments—right (caecum/ascending), transverse (including hepatic and splenic flexures), and left (descending/sigmoid/rectum)—each scored 0–3: 0 = unprepared, mucosa invisible; 1 = major staining/opaque fluid, partial visualization; 2 = minor residual staining/clear fluid, good visualization; 3 = entire mucosa seen, virtually no residue. Total BBPS ranges 0–9. Immediately after extubation, an experienced endoscopist assigned segmental scores; values were extracted in real time from the institutional endoscopy reporting system. 2.5 Sample-size calculation This external-validation study was powered under a dual framework: (i) events-per-variable (EPV) rules and (ii) minimum-event recommendations for validation cohorts. 1) Outcome: inadequate preparation defined as BBPS < 6 or any segment score < 2. 2) Predictors: the largest model contains 12 variables; adopting a liberal EPV ≥ 10 for validation requires ≥ 120 events, whereas methodological guidance sets a floor of 100 events. We adopted the stricter figure, targeting 120 events. 3) Expected incidence: published Chinese in-patient rates for magnesium-sulfate regimens range 10–15%; using the conservative π = 10%, N = 120 / 0.10 = 1 200. After allowing 5% missing data, 1 260 consecutive participants will be enrolled to ensure adequate power. 2.6 Statistical analysis All analyses were conducted in accordance with the TRIPOD and PROBAST guidelines using R version 4.3.2. Continuous variables are presented as mean ± SD or median (interquartile range), and categorical variables as number (percentage). Between-group comparisons were performed with the χ² test or Mann–Whitney U test, as appropriate. The three prediction models were applied to the external-validation cohort. Discrimination was quantified with the area under the receiver-operating-characteristic curve (AUC); higher AUC indicates better separation of individuals with and without IBP. Pairwise AUC comparisons were carried out using DeLong test. The optimal cut-off for each model was selected by maximising the Youden index (sensitivity + specificity-1), and the corresponding sensitivity, specificity and overall accuracy were reported. Calibration was evaluated graphically with calibration plots and tested with the Hosmer–Lemeshow goodness-of-fit test; P > 0.05 indicates adequate calibration. Decision-curve analysis was performed over threshold probabilities of 5%–95% to estimate net clinical benefit relative to “treat-all” and “treat-none” strategies. A model was considered clinically useful when its net-benefit curve lay above both reference lines. 2.7 Data collection and quality control A closed-loop “extract–verify–enter” team—comprising two research nurses (blinded to the study hypothesis) and one data manager—was formed. All variables were extracted independently in duplicate; discrepancies were resolved by the data manager. Data sources were: Real-time BBPS scores recorded at extubation in the endoscopy theatre database; Electronic medical records (EMR); Structured nursing assessment forms uploaded to the Clinical Care Classification (CCC) system, with missing items completed by telephone follow-up after discharge. The following domains were captured: demographics, IBD, cardiac/renal insufficiency, neurological comorbidity, colonic stricture, electrolyte disturbance, medications (opioids, tricyclic antidepressants, anaesthetics), last pre-preparation meal type, laxative preference, previous failed preparation, taste acceptability (3-level Likert), adverse events (nausea, vomiting, abdominal pain), self-reported compliance (full dose ingested or not), and final stool colour (yellow-watery vs other). Nursing-assessed variables comprised laxative preference, taste acceptability, compliance, adverse events and final stool colour. Daily 10% random audits maintained error rates < 2%. Variables with ≥ 1% missingness were multiply imputed (MICE); all changes were logged. Inter-rater agreement was high (κ ≥ 0.82). After database lock, an independent reviewer re-checked 20% of records against source documents, and an external statistician conducted the analyses. 3 Results 3.1 Study flow and baseline characteristics To examine the real-world transportability of the three previously derived models, we conducted a prospective, consecutive-enrolment external-validation study among in-patients at our centre. Between July 2024 and March 2025, 977 orally prepared, magnesium-sulfate-treated patients aged ≥ 18 years were recruited and analysed (Fig. 1 ). IBP (BBPS < 6 or any segment < 2) occurred in 107 participants (10.95%). Table 1 summarises baseline characteristics for the total cohort and stratified by preparation quality. Patients with IBP were older (66.4 ± 12.1 vs 61.0 ± 11.7 years, P = 0.008), more frequently had stroke/dementia (25.9% vs 6.7%, P < 0.001), reported lower acceptability of magnesium-sulfate taste (37.0% vs 12.1%, P < 0.001) and were less likely to pass yellow watery stool as the final lavage (55.6% vs 19.8%, P < 0.001). Table 1 Clinical and demographics characteristics of study population(n = 977) Baseline characteristics Overall(n = 977) Inadequate bowel preparation(n = 170) Adequate bowel preparation(n = 807) statistical value p value * Age (years), median (IQR) 63.19 (54.98–70.06) 65.91 (56.38–72.07) 62.98 (54.59–69.87) 4.589 0.032 Sex,no.(%) 0.156 0.693 male 605 (61.92) 103 (60.59) 502 (62.21) female 372 (38.08) 67 (39.41) 305 (37.79% BMI (kg/m²), median (IQR) 23.98 (21.93–26.07) 24.11 (22.22–25.69) 23.96 (21.91–26.09) 0.014 0.906 Serum potassium (mmol/L), median (IQR) 4.20 (3.99–4.49) 4.20 (4.00-4.49) 4.20 (3.99–4.48) 0.881 0.348 Serum sodium (mmol/L), median (IQR) 141.98(140.02-143.94) 141.06 (139.96-143.06) 141.99 (140.03-143.95) 1.847 0.174 Serum chloride (mmol/L), median (IQR) 104.95 (102.97-106.02) 104.93 (102.98–106.00) 104.96 (102.97-106.03) 0.067 0.795 Education level, n (%) 1.993 0.85 Primary school or below 88 (9.01) 13 (7.65) 75 (9.29) Junior high school 259 (26.51) 47 (27.65) 212 (26.27) Senior high school 185 (18.94) 29 (17.06) 156 (19.33) Secondary vocational school 80 (8.19) 14 (8.24) 66 (8.18) Associate degree 115 (11.77) 18 (10.59) 97 (12.02) Bachelor’s degree or above 250 (25.59) 49 (28.82) 201 (24.91) Diabetes mellitus, n (%) 113 (11.57) 25 (14.71) 88 (10.90) 1.984 0.159 Hypertension, n (%) 389 (39.82) 67 (39.41) 322 (39.90) 0.014 0.906 Neurological comorbidities (stroke or dementia), n (%) 26 (2.66) 6 (3.53) 20 (2.48) 0.599 0.439 Chronic constipation, n (%) 105 (10.75) 28 (16.47) 77 (9.54) 7.028 0.008 Inflammatory bowel disease, n (%) 34 (3.48) 8 (4.71) 26 (3.22) 0.921 0.337 Hepatic insufficiency, n (%) 65 (6.65) 7 (4.12) 58 (7.19) 2.13 0.144 Severe cardiac or renal insufficiency, n (%) 38 (3.89) 10 (5.88) 28 (3.47) 2.187 0.139 Gastroparesis, n (%) 39 (3.99) 6 (3.53) 33 (4.09) 0.115 0.735 Severe colonic stricture, n (%) 47 (4.81) 15 (8.82) 32 (3.97) 7.238 0.007 Electrolyte disturbances, n (%) 97 (9.93) 8 (4.71) 89 (11.03) 6.277 0.012 Medication use (opioids, tricyclic antidepressants, anesthetics), n (%) 74 (7.57) 20 (11.76) 54 (6.69) 5.163 0.023 History of abdominal surgery n (%) 251 (25.69) 46 (27.06) 205 (25.40) 0.202 0.653 Last meal type before bowel preparation n (%) 11.15 0.011 Clear liquid 359 (36.75) 79 (46.47) 280 (34.70) Full liquid 94 (9.62) 19 (11.18) 75 (9.29) Soft diet 27 (2.76) 5 (2.94) 22 (2.73) regular diet 497 (50.87) 67 (39.41) 430 (53.28) Expressed laxative preference, n (%) 96 (9.83) 38 (22.35) 58 (7.19) 36.45 < 0.001 History of inadequate bowel preparation, n (%) 67 (6.86) 33 (19.41) 34 (4.21) 50.782 < 0.001 Taste acceptability of magnesium sulfate, n (%) 11.515 0.003 readily acceptable 734 (75.13) 118 (69.41) 616 (76.33) neutral / tolerable 171 (17.50) 29 (17.06) 142 (17.60) unacceptable /intolerable 72 (7.37) 23 (13.53) 49 (6.07) Completed full laxative dose, n (%) 933 (95.50) 158 (92.94) 775 (96.03) 3.125 0.077 Adverse events with oral laxatives, n (%) 60 (6.14) 11 (6.47) 49 (6.07) 0.039 0.844 Insufficient physical exercise during preparation, n (%) 56 (5.73) 10 (5.88) 46 (5.70) 0.009 0.926 Passage of yellow-watery stool as final lavage 885 (90.58) 119 (70.00) 766 (94.92) 102.226 < 0.001 3.2 Comparative model performance To comprehensively evaluate the predictive utility of each model for IBP, we analysed four complementary dimensions: discrimination, performance metrics at the optimal Youden cut-off, calibration, and clinical utility. Table 2 summarises the overall performance of the three models in the external validation cohort of 977 patients. Table 2 Performance summary of the three models in the 977-patient external validation cohort model Cutoff AUC AUC.low AUC.up ACC(%) SEN(%) SPE(%) PPV NPV AIC BIC RMSE R2_Tjur Model-1 5 0.707 0.665 0.749 76.5 81.3 48.8 0.88 0.36 825.98 835.75 0.36 0.09 Model-2 -15.5 0.699 0.656 0.742 66.6 68.0 60.0 0.89 0.28 829.08 838.85 0.36 0.09 Model-3 -12.5 0.785 0.746 0.824 79.7 83.8 60.6 0.91 0.44 753.70 763.47 0.34 0.18 3.3 Discrimination performance The area under the receiver-operating-characteristic curve (AUC) served as the primary metric for discrimination, with DeLong’s test used to evaluate pairwise differences between models. In the external validation cohort, model-3 achieved an AUC of 0.785 (95% CI 0.746–0.824), substantially higher than model-2 (0.681, 95% CI 0.635–0.725) and model-1 (0.679, 95% CI 0.633–0.725) (Fig. 1 ). The incremental AUC for model-3 versus model-1 was 0.106 (P < 0.001), and for model-3 versus model-2 it was 0.104 (P < 0.001), demonstrating that the additional clinical variables integrated into model-3 confer significantly superior discriminative capacity for predicting IBP among multicentre in-patients. 3.4 Performance at optimal Youden cut-offs At the cut-off maximising the Youden index, model-3 achieved higher sensitivity, specificity and overall accuracy than both model-1 and model-2, corroborating its superior discriminative performance. 3.5 Multi-metric comparison A radar chart (Fig. 2 ) simultaneously displays sensitivity, specificity, accuracy, PPV and NPV for the three models. Model-3 occupies the largest area across all five dimensions, providing visual confirmation that it outperforms model-1 and model-2 and reinforcing its selection as the optimal prediction tool. 3.6 Calibration assessment Calibration plots were constructed to compare predicted versus observed probabilities of inadequate bowel preparation. The x-axis represents model-predicted risk; the y-axis shows the corresponding observed event frequency. A bias-corrected curve lying above the ideal diagonal indicates over-estimation, whereas a curve below the diagonal signifies under-estimation. As shown in Fig. 3 , all three models demonstrated satisfactory agreement between prediction and observation. Model-3’s calibration curve lay closest to the ideal line, indicating the highest fidelity. Curves for model-1 and model-2 were also acceptable, but deviated slightly more from the reference. Hosmer-Lemeshow tests corroborated these visual impressions: - Model-1: χ² = 1.852, df = 5, P = 0.867 - Model-2: χ² = 1.352, df = 3, P = 0.717 - Model-3: χ² = 6.214, df = 5, P = 0.286 All P-values > 0.05 confirm adequate calibration and reinforce the reliability of model-3 for predicting IBP risk in this external cohort. 3.7 Clinical utility Decision-curve analysis (DCA) quantified net benefit across clinically plausible threshold probabilities (pₜ). The x-axis represents pₜ—the probability above which a clinician would implement an intervention; the y-axis shows net benefit, calculated as the true-positive fraction minus the false-positive fraction weighted by the odds of pₜ. All three models achieved positive net benefit within discrete pₜ ranges: - Model-1: 0.549–0.914 - Model-2: 0.517–0.923 - Model-3: 0.401–0.982 Model-3 provided the widest useful interval and its curve remained above both “treat-all” and “treat-none” reference lines throughout this range (Fig. 4 ), indicating superior clinical value across diverse decision contexts. The narrower benefit windows of model-1 and model-2 imply that model-3 is better positioned to avert adverse outcomes, enhance therapeutic yield, and optimise resource allocation. Consequently, model-3 is the preferred IBP risk-assessment tool for adult in-patients undergoing colonoscopy in our centre. 4 Discussion 4.1 Model-3 is the optimal tool for predicting IBP under our hospital’s magnesium-sulfate regimen In this external-validation cohort, three candidate models with differing variable sets were compared for their ability to quantify IBP risk. Goodness-of-fit tests for all models indicated close agreement between predicted probabilities and observed event rates, underscoring robust stability and generalizability. Calibration plots and decision-curve analysis further supported clinical utility across a wide range of risk thresholds. Model-3, which integrates comprehensive clinical variables, outperformed the earlier in-house models in discrimination, calibration, and net clinical benefit. Its AUC of 0.785 was significantly higher than that of model-1 (0.679) and model-2 (0.681), representing an increment > 0.10. Decision-curve analysis showed an extended net-benefit interval of 0.401–0.982, implying additional value even when the risk threshold is < 50%. The high sensitivity (0.838) paired with moderate specificity (0.606) enables reliable identification of high-risk patients while avoiding excessive intervention. A goodness-of-fit P-value of 0.286 and a calibration slope close to unity confirm that predicted probabilities closely mirror actual risk. These gains reflect the incremental value of adding neurological comorbidities, colonic stricture, and passage of yellow watery stool—variables that capture key determinants of preparation quality not included in earlier iterations. 4.2 Clinical implications of model optimisation Building on model-2, model-3 explicitly incorporates neurological comorbidities and procedure-related factors (acceptability of laxative taste, passage of yellow watery stool). Neurological disease may compromise bowel-preparation adherence through cognitive impairment and/or autonomic dysregulation that slows intestinal transit, whereas colonic strictures directly limit cleansing efficacy [ 55 ]. The presence of yellow watery effluent—an immediate, patient-recognisable sign that the final lavage is approaching clear—was added as a dynamic indicator of real-time preparation quality, thereby sharpening the model’s discriminative ability [ 56 ]. These variables are rarely captured by traditional scores; their inclusion not only improved predictive performance but also furnishes clinicians with actionable targets (e.g., pre-emptive dose modification or intensified education) [ 57 ]. Model-3’s AUC surpassed previously reported values for the Aronchick score (0.72) and the Boston prediction model (0.74). Moreover, by introducing decision-curve analysis to the bowel-preparation literature, we provide the first evidence in a Chinese cohort that meaningful net benefit persists at low-to-moderate threshold probabilities, mitigating both over- and under-intervention. This advantage probably reflects a more comprehensive variable set aligned with complex clinical realities. Nevertheless, the derived cut-off (− 12.500) should be interpreted judiciously and never applied mechanistically. Compared with previously published models, our Model-3 demonstrated superior discriminative performance. For instance, Afecto et al. externally validated two commonly cited prediction models (Model 1 and Model 2) in a Portuguese tertiary hospital population, reporting AUCs of only 0.62 for both, with low positive predictive values and limited clinical utility in identifying inadequate bowel preparation[ 50 ]. Similarly, the PREPA-CO score, developed by Berger et al. in a French prospective cohort, achieved an AUC of 0.621 using a self-administered questionnaire-based approach[ 58 ]. While these models identified overlapping risk factors such as diabetes, prior abdominal surgery, and history of inadequate preparation, they lacked real-time nursing assessment variables and were not validated in inpatient populations receiving magnesium sulfate. In contrast, Model-3 integrates both static and dynamic variables—including nurse-documented taste acceptability, self-reported adherence, and final stool character—resulting in significantly improved predictive accuracy and clinical applicability in our Chinese inpatient cohort. 4.3 Limitations of model-1 and model-2 Although the AUCs of model-1 and model-2 were only marginally lower (0.679–0.681), both models retained positive net benefit within specific threshold-probability ranges, implying contextual utility. Model-1, in particular, combined high sensitivity with low specificity, making it potentially useful as a screening tool for high-risk cohorts; however, this imbalance inevitably inflates the false-positive rate. Calibration curves revealed a systematic underestimation of risk, especially in low-risk patients, suggesting that neither model fully captures the true event rate. This under-estimation is plausibly attributable to the omission of well-established predictors such as obesity and diabetes mellitus—factors that have been independently associated with inadequate bowel preparation[ 59 ]. Their exclusion may lead to insufficient risk attribution in selected high-risk individuals. Furthermore, the narrow range of threshold probabilities over which either model yields clinical benefit indicates limited added value in routine decision-making 4.4 Robustness and transportability of external validation Consecutive enrollment coupled with prospective data collected from July 2024 to March 2025 guaranteed both timeliness and representativeness of the validation cohort. Model-3’s excellent calibration and discrimination indicate high local applicability and suggest that it should perform comparably in similar tertiary-care settings. Its parsimonious variable set further facilitates seamless integration into routine clinical pathways, offering real-time decision support for individualized bowel-preparation interventions. 4.5 Limitations and Future Directions Several limitations should be acknowledged. First, the study was confined to in-patients; generalizability to out-patients or community populations remains to be established. The representativeness and size of the external-validation cohort may also constrain wider applicability. Second, variables such as “taste acceptability” and “dietary preference” are inherently subjective and susceptible to recall bias. Future iterations could incorporate objective biomarkers—e.g., plasma osmolality, electrolyte shifts—while also accounting for heterogeneity in preparation regimens (dosage, timing, adjunct medications) and the intestinal microbiome, all of which may influence predictive performance. Nursing-related metrics such as education frequency could likewise be integrated to refine model precision. Finally, the ability of the model to predict longer-term outcomes (re-hospitalization, mortality) was not examined and warrants prospective evaluation. 5.Conclusion In this multicenter validation study, three prediction models (model-1, model-2 and model-3) were externally compared in hospitalized adults. Model-3 demonstrated superior discrimination, calibration and clinical utility relative to both model-1 and model-2. Consequently, model-3 is considered the optimal instrument for pre-colonoscopy IBP risk assessment in adult in-patients at our institution. Next steps will evaluate the feasibility and cost-effectiveness of integrating model-3 into the electronic health record to guide individualized interventions, thereby providing an evidence-based foundation for precision management of in-patient bowel preparation before colonoscopy. Declarations Acknowledgements We are grateful to all the patients who gave consent to participate in the study and the research team who endeavoured to ensure that the study is successfully carried out. Author contributions Q.Y.S, M.Z. M.D conceived and designed the study under the supervision of Y.W. Q.Y.S, M.Z, M.D implemented the study under the supervision of Y.H.S, Y.F.O. Q.Y.S, F.Y.F conducted data analysis. Q.Y.S, M.Z, M.D, H.K.S, Y.H.S, Y.F.O, Y.W interpreted the findings. M.D, Y.W involved in validation; Q.Y.S, M.Z involved in data curation. Y.W sought funding. Q.Y.S wrote the first draft; all authors reviewed and approved the final manuscript. Funding This work was supported by the Evidence-Based Nursing Practice Program of the Nursing Department, Zhongshan Hospital, Fudan University (Grant No. HLXZ202404) and the Scientific Research Development Fund of Zhongshan Hospital, Fudan University (Grant No. 2023XKPT08-RC2). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Data availability statement All data generated or analysed during this study are included in this published article. Ethical approval The study protocol was approved by the Ethics Committee of Zhongshan Hospital, Fudan University (Approval No. B2024-278R) in accordance with the ethical guidelines of the Declaration of Helsinki and the relevant regulations of the ethics committee. Consent to participate All participants received detailed study information and provided written informed consent through the hospital’s secure electronic signature platform before any data collection. Consent for publication Written informed consent for publication was obtained from all participants through the hospital’s electronic consent platform. Competing interests The authors declare no competing interests. Data availability statement All data generated or analysed during this study are included in this published article. Further inquiries can be directed to the corresponding author on reasonable request. Trial registration This study was registered with the Chinese Clinical Trial Registry (ChiCTR) under registration number ChiCTR2400081234. References Siegel RL, Miller KD, Goding Sauer A, Fedewa SA, Butterly LF, Anderson JC, et al. Colorectal cancer statistics, 2020. Cancer J Clin. 2020;70(3):145–64. 10.3322/caac.21601 . Issaka RB, Chan AT, Gupta S, AGA Clinical Practice Update on Risk Stratification for Colorectal Cancer Screening and Post-Polypectomy Surveillance. Expert Rev Gastroenterol. 2023;165(5):1280–91. 10.1053/j.gastro.2023.06.033 . Rawla P, Sunkara T, Barsouk A. Epidemiology of colorectal cancer: incidence, mortality, survival, and risk factors. Gastroenterol Rev. 2019;14(2):89–103. 10.5114/pg.2018.81072 . Nie J-X, Xie Q, Yuan Y, Liu M-Y, Du J-K, Li N, et al. Rising Incidence of Total and Early-Onset Colorectal Cancer: A Global Perspective on Burden, Risk Factors, and Projections to 2031. J Gastrointest Cancer. 2025;56(1). 10.1007/s12029-025-01257-1 . Raphael MJ, Wildgoose P, Servidio-Italiano F, De Vera MA, Brenner D, D’Angelo MS, et al. Breaking Down Barriers to Detection and Care in Early-Age-Onset Colorectal Cancer in Canada. Curr Oncol. 2023;30(11):9392–405. 10.3390/curroncol30110680 . Yang Y, undefined u, Han Z, Li X, Huang A, Shi J, et al. Epidemiology and risk factors of colorectal cancer in China. Chin J Cancer Res. 2020;32(6):729–41. 10.21147/j.issn.1000-9604.2020.06.06 . Li N, Lu B, Luo C, Cai J, Lu M, Zhang Y, et al. Incidence, mortality, survival, risk factor and screening of colorectal cancer: A comparison among China, Europe, and northern America. Cancer Lett. 2021;522:255–68. 10.1016/j.canlet.2021.09.034 . Chen Y, Chen T, Fang J-Y. Burden of gastrointestinal cancers in China from 1990 to 2019 and projection through 2029. Cancer Lett. 2023;560:216127. 10.1016/j.canlet.2023.216127 . Sun C, Liu Y, Huang Y, Li B, Rang W. Colorectal Cancer Incidence and Mortality Trends and Analysis of Risk Factors in China from 2005 to 2015. Int J Gen Med. 2021;9965–76. 10.2147/ijgm.s344448 . Wang CF. Colorectal cancer incidence and mortality in Shanghai 2016 and trend analysis 2002–2016. Journal of Surgery Concepts & Practice; 2021. Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. Cancer J Clin. 2021;71(3):209–49. 10.3322/caac.21660 . Hassan C, East J, Radaelli F, Spada C, Benamouzig R, Bisschops R, et al. Bowel preparation for colonoscopy: European Society of Gastrointestinal Endoscopy (ESGE) Guideline - Update 2019. Endoscopy. 2019;51(8):775–94. 10.1055/a-0959-0505 . Hagen R, Nguyen MTT, Anderson JC, Birk JW. Navigating Bowel Preparation for Colonoscopy. J Clin Gastroenterol. 2024;59(4):285–97. 10.1097/mcg.0000000000002124 . Woo DH, Kim KO, Jeong DE, Nam YJ, Lee SH, Jang BI, et al. Prospective analysis of factors associated with inadequate bowel preparation for colonoscopy in actual clinical practice. Intestinal Res. 2018;16(2):293. 10.5217/ir.2018.16.2.293 . Kastenberg D, Bertiger G, Brogadir S. Bowel preparation quality scales for colonoscopy. World J Gastroenterol. 2018;24(26):2833–43. 10.3748/wjg.v24.i26.2833 . Rai T, Navaneethan U, Gohel T, Podugu A, Thota PN, Kiran RP, et al. Effect of quality of bowel preparation on quality indicators of adenoma detection rates and colonoscopy completion rates. Gastroenterol Rep. 2015;4(2):148–53. 10.1093/gastro/gov002 . Colorectal Endoscopy Group, Chinese Society of Digestive Endoscopy. Expert consensus on bowel preparation for colonoscopy (2023, Guangzhou). Chin J Dig Endoscopy. 2023;40(6):421–30. Hernández G, Gimeno-García AZ, Quintero E. Strategies to Improve Inadequate Bowel Preparation for Colonoscopy. Front Med. 2019;6. 10.3389/fmed.2019.00245 . Voiosu T, Voiosu A, Voiosu R. Bowel preparation for colonoscopy. Curr Opin Gastroenterol. 2016;32(5):385–6. 10.1097/mog.0000000000000297 . Parra-Blanco A, Ruiz A, Alvarez-Lobos M, Amorós A, Gana JC, Ibáñez P, et al. Achieving the best bowel preparation for colonoscopy. World J Gastroenterol. 2014;20(47):17709–26. 10.3748/wjg.v20.i47.17709 . Lai EJ, Calderwood AH, Doros G, Fix OK, Jacobson BC. The Boston bowel preparation scale: a valid and reliable instrument for colonoscopy-oriented research. Gastrointest Endosc. 2009;69(3):620–5. 10.1016/j.gie.2008.05.057 . Kutyla MJ, Gray MA, von Hippel C, Hourigan LF, Kendall BJ, Whaley AJ, et al. Improving the Quality of Bowel Preparation: Rewarding Patients for Success or Intensive Patient Education? Dig Dis. 2020;39(2):113–8. 10.1159/000510461 . Lam TYT, Wu PI, Tang RSY, Tse YK, Lau JYW, Wu JCY, et al. Nurse-led reinforced education by mobile messenger improves the quality of bowel preparation of colonoscopy in a population-based colorectal cancer screening program: A randomized controlled trial. Int J Nurs Stud. 2022;133:104301. 10.1016/j.ijnurstu.2022.104301 . Chan W-K, Saravanan A, Manikam J, Goh K-L, Mahadeva S. Appointment waiting times and education level influence the quality of bowel preparation in adult patients undergoing colonoscopy. BMC Gastroenterol. 2011;11(1). 10.1186/1471-230x-11-86 . Gorelik Y. Volume of fluid consumption during preparation for colonoscopy is possibly the single most important determinant of bowel preparation adequacy. Annals Gastroenterol. 2021. 10.20524/aog.2021.0642 . Adler A, Wegscheider K, Lieberman D, Aminalai A, Aschenbeck J, Drossel R, et al. Factors determining the quality of screening colonoscopy: a prospective study on adenoma detection rates, from 12,134 examinations (Berlin colonoscopy project 3, BECOP-3). Gut. 2013;62(2):236–41. 10.1136/gutjnl-2011-300167 . Amitay EL, Niedermaier T, Gies A, Hoffmeister M, Brenner H. Risk Factors of Inadequate Bowel Preparation for Screening Colonoscopy. J Clin Med. 2021;10(12):2740. 10.3390/jcm10122740 . Cornella SL, Wentworth BJ, Orton C, Horton BJ, Powell SM. Modelling characteristics of inadequate bowel preparations for colonoscopy. GastroHep. 2020;2(2):72–8. 10.1002/ygh2.389 . Li J, Qu C, Zhao Q, Fu JN, Du BX, Fan QY, et al. Current status and influencing factors of bowel preparation before colonoscopy. China Med Guide. 2020;18(27):1–2. 10.15912/j.cnki.gocm.2020.27.002 . Ginnaram S, Nugooru S, Tahir D, Marrache K, Jacoby P, Elrod S, et al. S885 Impact of Bowel Preparation Volume on Inpatient Colonoscopy Outcomes: A Retrospective Analysis. Am J Gastroenterol. 2024;119. 10.14309/01.ajg.0001032908.82582.6c . Dickson Z, Park SR, Tung W, Douglas SR, Haug R, Shenoy A, et al. 2882 Analysis of Risk Factors for Poor Bowel Preparation in Hospitalized Patients. Am J Gastroenterol. 2019;114(1). 10.14309/01.ajg.0000601060.91042.cd . Mandark G, Townsend C, Almaghrabi M, Alotaibi A, Khanna N, Yan B, et al. 566 Improving Bowel Preparation Quality for Inpatient Colonoscopies at a Tertiary Hospital. Am J Gastroenterol. 2019;114(1). 10.14309/01.ajg.0000591796.26128.58 . Wang SS, Guo XX, Li ZK. Research progress on factors influencing the quality of bowel preparation in patients undergoing colonoscopy. Chin J Mod Nurs. 2021;27(14):1942–6. Clark BT, Rustagi T, Laine L. What Level of Bowel Prep Quality Requires Early Repeat Colonoscopy: Systematic Review and Meta-Analysis of the Impact of Preparation Quality on Adenoma Detection Rate. Am J Gastroenterol. 2014;109(11):1714–23. 10.1038/ajg.2014.232 . Ray-Offor E, Jebbin N. Risk Factors for Inadequate Bowel Preparation During Colonoscopy in Nigerian Patients. Cureus. 2021. 10.7759/cureus.17145 . Mitchell Rn A. Reducing Suboptimal Bowel Preparation. J Qual Health Care Econ. 2021;4(4). 10.23880/jqhe-16000232 . Zad M, Do CN, Heffernan A, Johnston L, Al-Ansari M. Factors affecting bowel preparation adequacy and procedural time. JGH Open. 2019;4(2):206–14. 10.1002/jgh3.12241 . Hassan C, Fuccio L, Bruno M, Pagano N, Spada C, Carrara S, et al. A Predictive Model Identifies Patients Most Likely to Have Inadequate Bowel Preparation for Colonoscopy. Clin Gastroenterol Hepatol. 2012;10(5):501–6. 10.1016/j.cgh.2011.12.037 . Beran A, Aboursheid T, Ali AH, Albunni H, Mohamed MF, Vargas A, et al. Risk Factors for Inadequate Bowel Preparation in Colonoscopy: A Comprehensive Systematic Review and Meta-Analysis. Am J Gastroenterol. 2024;119(12):2389–97. 10.14309/ajg.0000000000003073 . Dik VK, Moons LMG, Hüyük M, van der Schaar P, de Vos tot Nederveen Cappel WH, ter Borg PCJ et al. Predicting inadequate bowel preparation for colonoscopy in participants receiving split-dose bowel preparation: development and validation of a prediction score. Gastrointestinal endoscopy. 2015;81(3):665 – 72. 10.1016/j.gie.2014.09.066 Gimeno-García A, Baute J, Hernandez G, Morales D, Gonzalez-Pérez C, Nicolás-Pérez D, et al. Risk factors for inadequate bowel preparation: a validated predictive score. Endoscopy. 2017;49(6):536–43. 10.1055/s-0043-101683 . Higashimori A, Maeda N, Nakatani M, Yamamoto I, Yanagida T, Kin D, et al. Importance of Bowel Habits in Predicting Inadequate Bowel Preparation: A Prospective Observational Study. Dig Dis Sci. 2023;68(11):4148–55. 10.1007/s10620-023-08089-8 . Zhang Y, Wang L, Wu W, Zhang S, Zhang M, She W, et al. Predictors of inadequate bowel preparation in older patients undergoing colonoscopy: A systematic review and meta-analysis. Int J Nurs Stud. 2024;149:104631. 10.1016/j.ijnurstu.2023.104631 . Xu Y, Ou S, Yuan X, Yao J, Yang T. Construction and Validation of Risk Scoring Model for Inadequate Bowel Preparation in Elderly Patients Undergoing Colonoscopy. Therapeutics and Clinical Risk Management. 2025:885 – 94. 10.2147/tcrm.s503389 Berger A, Cesbron-Métivier E, Bertrais S, Olivier A, Becq A, Boursier J, et al. A predictive score of inadequate bowel preparation based on a self-administered questionnaire: PREPA-CO. Clin Res Hepatol Gastroenterol. 2021;45(4):101693. 10.1016/j.clinre.2021.101693 . Gu F, Xu J, Du L, Liang H, Zhu J, Lin L, et al. The Machine Learning Model for Predicting Inadequate Bowel Preparation Before Colonoscopy: A Multicenter Prospective Study. Clin Translational Gastroenterol. 2024;15(5). 10.14309/ctg.0000000000000694 . Liu J, Jiang W, Yu Y, Gong J, Chen G, Yang Y, et al. Applying machine learning to predict bowel preparation adequacy in elderly patients for colonoscopy: development and validation of a web-based prediction tool. Ann Med. 2025;57(1). 10.1080/07853890.2025.2474172 . Bishop R, Laugharne R, Burrows L, undefined u, Ward S, Eustice S, et al. Laxative use in adults with intellectual disabilities: development of prescribing guidelines. BJPsych Open. 2024;10(3). 10.1192/bjo.2024.50 . Yuan X, Gao H, Liu C, Wang W, Xie J, Zhang Z, et al. External validation of two prediction models for adequate bowel preparation in Asia: a prospective study. Int J Colorectal Dis. 2022;37(6):1223–9. 10.1007/s00384-022-04156-0 . Afecto E, Ponte A, Fernandes S, Gomes C, Correia João P, Carvalho J. Validation and Application of Predictive Models for Inadequate Bowel Preparation in Colonoscopies in a Tertiary Hospital Population. GE - Portuguese J Gastroenterol. 2021;30(2):134–40. 10.1159/000520905 . Ning Y, Volovici V, Ong MEH, Goldstein BA, Liu N. A roadmap to fair and trustworthy prediction model validation in healthcare. undefined. 2023. 10.48550/ARXIV.2304.03779 Ramspek CL, Jager KJ, Dekker FW, Zoccali C, van Diepen M. External validation of prognostic models: what, why, how, when and where? Clin Kidney J. 2020;14(1):49–58. 10.1093/ckj/sfaa188 . Mann M, Ensor J, Collins G, Riley R. How are published clinical prediction model studies assessing and reporting calibration performance at external validation? OSF, 2024. 2024. Gallitto G, Englert R, Kincses B, Kotikalapudi R, Li J, Hoffschlag K, et al. External validation of machine learning models - registered models and adaptive sample splitting. bioRxiv. 2023. 10.1101/2023.12.01.569626 . Kumar A, Shenoy V, Buckley MC, Durbin L, Mackey J, Mone A, et al. Endoscopic Disease Activity and Biologic Therapy Are Independent Predictors of Suboptimal Bowel Preparation in Patients with Inflammatory Bowel Disease Undergoing Colonoscopy. Dig Dis Sci. 2022;67(10):4851–65. 10.1007/s10620-022-07530-8 . Lai EJ, Calderwood AH, Doros G, Fix OK, Jacobson BC. The Boston bowel preparation scale: a valid and reliable instrument for colonoscopy-oriented research. Gastrointest Endosc. 2009;69(3 Pt 2):620–5. 10.1016/j.gie.2008.05.057 . Arslanca G, Aygün M. Effect of nurse-performed enhanced patient education on colonoscopy bowel preparation quality. Rev Latinoam Enferm. 2022;30:e3626. 10.1590/1518-8345.5597.3626 . Berger A, Cesbron-Métivier E, Bertrais S, Olivier A, Becq A, Boursier J, et al. A predictive score of inadequate bowel preparation based on a self-administered questionnaire: PREPA-CO. Clin Res Hepatol Gastroenterol. 2021;45(4):101693. 10.1016/j.clinre.2021.101693 . Hyun JH, Kim SJ, Park JH, Wie GA, Kim JS, Han KS, et al. Lifestyle Factors and Bowel Preparation for Screening Colonoscopy. Annals coloproctology. 2018;34(4):197–205. 10.3393/ac.2018.03.13 . Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8265970","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":560046904,"identity":"f4184d1a-3790-4187-84a9-923fa8d23e4c","order_by":0,"name":"Qi-yu Sun","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABC0lEQVRIie2Qv2rDMBCHzwjkRYlXd2j9Clc69mXsqYsbDF06GYPBWQpZDclDtAQ6X9CQxdDVQwZNnTxkr4eeg1NKifDaQd+gP0jf6acDcDj+LQgQ8CSGtQTwCICmlavirMjT/SllsGhUYEqJ6keCLtPJ9mNvRPacL+ZRaQj6AwTr4nLxdhF7G9TJOzUo6kY/SSmRvOoTwsPllzBMUahB2b34X7OKkkpySK/QfBRbgo3KtvSNmFU5K/6Rg9kVaEfllYuzIlhR3AppV7DpcLfBh7u6Ueye/pJmHE+rsLUEW6a3puvvr1cr7pjijkXl/s0ce30T1JZgDCkofzbxeVTW+wMK8j+Kw+FwOH7xDfm9XmDhluarAAAAAElFTkSuQmCC","orcid":"","institution":"Zhongshan Hospital","correspondingAuthor":true,"prefix":"","firstName":"Qi-yu","middleName":"","lastName":"Sun","suffix":""},{"id":560046905,"identity":"987c87c6-6a71-45fb-8dfa-62e6dfe02d35","order_by":1,"name":"Mei Zhang","email":"","orcid":"","institution":"Zhongshan Hospital","correspondingAuthor":false,"prefix":"","firstName":"Mei","middleName":"","lastName":"Zhang","suffix":""},{"id":560046906,"identity":"2cc28777-7623-4695-91c2-3de206aa7411","order_by":2,"name":"Fei-yan Feng","email":"","orcid":"","institution":"Zhongshan Hospital","correspondingAuthor":false,"prefix":"","firstName":"Fei-yan","middleName":"","lastName":"Feng","suffix":""},{"id":560046907,"identity":"35520dfe-bd50-47cc-8b4b-f4221ab21378","order_by":3,"name":"Ding Min","email":"","orcid":"","institution":"Zhongshan Hospital","correspondingAuthor":false,"prefix":"","firstName":"Ding","middleName":"","lastName":"Min","suffix":""},{"id":560046908,"identity":"1367381b-c3ea-4ff0-a6f4-906afd03c895","order_by":4,"name":"Yue-hong Shen","email":"","orcid":"","institution":"Zhongshan Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yue-hong","middleName":"","lastName":"Shen","suffix":""},{"id":560046909,"identity":"2b884f1a-f151-4b4a-8f79-497ef5896d75","order_by":5,"name":"Yu-feng Ou","email":"","orcid":"","institution":"Zhongshan Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yu-feng","middleName":"","lastName":"Ou","suffix":""},{"id":560046910,"identity":"f5fe72fd-783c-4445-bef4-4224fb370a62","order_by":6,"name":"Yan Wu","email":"","orcid":"","institution":"Zhongshan Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yan","middleName":"","lastName":"Wu","suffix":""}],"badges":[],"createdAt":"2025-12-03 05:08:06","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8265970/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8265970/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":98751947,"identity":"e0eceffa-f03f-4663-b58a-9ab740fefc86","added_by":"auto","created_at":"2025-12-22 09:13:34","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":191421,"visible":true,"origin":"","legend":"","description":"","filename":"1205.docx","url":"https://assets-eu.researchsquare.com/files/rs-8265970/v1/526729d916c88ca58762657f.docx"},{"id":98751944,"identity":"8a652f48-8727-4f81-ba29-0bb17377eac1","added_by":"auto","created_at":"2025-12-22 09:13:34","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":8796,"visible":true,"origin":"","legend":"","description":"","filename":"ec2d84ef74cc45e5886b7d558390a6dd.json","url":"https://assets-eu.researchsquare.com/files/rs-8265970/v1/06d3e9dd7c0e8b8692cc8fff.json"},{"id":98778430,"identity":"c1974ace-465c-42ba-aaca-f4b8b068b907","added_by":"auto","created_at":"2025-12-22 12:29:15","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":170151,"visible":true,"origin":"","legend":"","description":"","filename":"ec2d84ef74cc45e5886b7d558390a6dd1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8265970/v1/942c4b7472e0c2fefe6e3d1e.xml"},{"id":98751955,"identity":"b1078068-b59a-483d-a279-c5ea2fb09a25","added_by":"auto","created_at":"2025-12-22 09:13:34","extension":"jpeg","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":253505,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8265970/v1/4fb25c55c366a5e328e11e8c.jpeg"},{"id":98778881,"identity":"71a8e002-7bd4-4632-a1c7-af9e41a3cdb4","added_by":"auto","created_at":"2025-12-22 12:29:46","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":42443,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8265970/v1/42a9836d10056d0470c920c4.png"},{"id":98751949,"identity":"df3421d9-5be9-4c41-8c61-ffcea445cbe1","added_by":"auto","created_at":"2025-12-22 09:13:34","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":54773,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8265970/v1/ba9dea62741ff59cef1e4899.png"},{"id":98751953,"identity":"6cb75d94-911d-4f89-b721-ccb68d27a688","added_by":"auto","created_at":"2025-12-22 09:13:34","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":33527,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8265970/v1/02d1f5f62ea10700f2057e71.png"},{"id":98751957,"identity":"b922dc3d-ed7a-4358-8688-1fb89cfa6eaa","added_by":"auto","created_at":"2025-12-22 09:13:34","extension":"xml","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":166293,"visible":true,"origin":"","legend":"","description":"","filename":"ec2d84ef74cc45e5886b7d558390a6dd1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8265970/v1/f1fe6808e77a49c1298e113c.xml"},{"id":98778901,"identity":"d88e7778-92be-486d-8211-4c34b468f8bc","added_by":"auto","created_at":"2025-12-22 12:29:47","extension":"html","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":181990,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8265970/v1/bec2e1c03ffb706d7cb779fa.html"},{"id":98777838,"identity":"010adbb2-73d7-4225-a0ec-a4ca36aab233","added_by":"auto","created_at":"2025-12-22 12:28:33","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":46592,"visible":true,"origin":"","legend":"\u003cp\u003eROC curves of the three models.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8265970/v1/9ade53e7c71b3ec99030316c.png"},{"id":98751945,"identity":"f4ad413d-d446-48c8-866a-20f8f6fbb82a","added_by":"auto","created_at":"2025-12-22 09:13:34","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":57275,"visible":true,"origin":"","legend":"\u003cp\u003eRadar-chart comparison of the three models in the external validation set.\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8265970/v1/e3d8fc03b65648aae755cdba.jpg"},{"id":98777691,"identity":"ee2b5566-5cc1-438c-9e08-6ffb57b9fc5d","added_by":"auto","created_at":"2025-12-22 12:28:21","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":71326,"visible":true,"origin":"","legend":"\u003cp\u003eCalibration plots for the three models\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8265970/v1/edb32c21b64fd85e3ef63387.jpg"},{"id":98751950,"identity":"05792094-be3f-4ed5-9d5c-402941f4189c","added_by":"auto","created_at":"2025-12-22 09:13:34","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":174984,"visible":true,"origin":"","legend":"\u003cp\u003eDecision-curve analysis comparing the three models for inadequate bowel preparation.\u003c/p\u003e","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8265970/v1/3ad309f3ac0119208bd990e8.jpeg"},{"id":99791282,"identity":"5b917748-ba6b-45d0-b055-8c7769da79b5","added_by":"auto","created_at":"2026-01-08 12:59:24","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1584509,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8265970/v1/142593e3-34d8-4ec7-968c-f44185b05492.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Comparative performance and external validation of three different models in predicting inadequate bowel preparation among Chinese inpatients undergoing colonoscopyFirst Name Last","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eGlobal data show that both the incidence and mortality of colorectal cancer (CRC) remain unacceptably high[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. According to GLOBOCAN 2018, CRC ranks as the third most lethal and the fourth most commonly diagnosed malignancy worldwide[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. The American Cancer Society estimated that, in 2020 alone, approximately 147,950 Americans would be newly diagnosed with CRC and 53,200 would succumb to the disease [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Of particular concern is the striking rise in early-onset CRC among individuals younger than 50 years across diverse populations[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e] .\u003c/p\u003e \u003cp\u003eChina\u0026mdash;the world\u0026rsquo;s most populous country\u0026mdash;carries a disproportionate share of the colorectal cancer burden. Although its age-standardized incidence and mortality rates are lower than the global average, the absolute numbers of new cases and deaths are the highest worldwide because of its vast population [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. More than half of all global CRC diagnoses and related deaths are estimated to occur in China [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. In 2020 alone, 555,480 Chinese residents were newly diagnosed with CRC, corresponding to an age-standardized rate of 23.9 per 100,000 [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Between 2005 and 2015, both the incidence and mortality of CRC in China increased steadily [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Data from Shanghai registries reveal a continuous rise in CRC incidence and mortality from 2002 to 2016[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Moreover, Chinese patients typically develop CRC approximately one decade earlier than their Western counterparts; the median age at diagnosis is 58 years[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eWidespread adoption of endoscopic techniques has established colonoscopy as the gold standard for screening, diagnosing, and treating colorectal disease[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. The procedure is pivotal to improving CRC survival, yet its success hinges directly on the quality of bowel preparation. Adequate cleansing markedly increases polyp and lesion detection while reducing procedural complications[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. High-quality preparation is therefore the cornerstone of effective colonoscopy, governing adenoma detection rates and the subsequent efficacy of surveillance and therapy[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eColonoscopy is the cornerstone of colorectal-cancer screening and diagnosis, yet its efficacy and safety hinge on adequate bowel cleansing. International guidelines stipulate a minimum preparation success rate of \u0026ge;\u0026thinsp;90% as a key quality indicator[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Inadequate bowel preparation (IBP)\u0026mdash;defined as residual stool or fluid that impairs mucosal visualization\u0026mdash;remains a common quality deficit[\u003cspan additionalcitationids=\"CR19\" citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e], reducing adenoma detection, prolonging procedure time, increasing complication risk and cost, and being particularly prevalent among in-patients. Cleansing is conventionally graded with instruments such as the Boston Bowel Preparation Scale (BBPS) [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Multiple patient-related [\u003cspan additionalcitationids=\"CR23\" citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e], disease-related, regimen-related [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e] and health-system factors interact to perpetuate IBP, which still complicates 20\u0026ndash;25% of examinations[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e], with some series reporting rates as high as 44%[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. Prospective US data document an IBP rate of 20.9%[\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e], and a recent Chinese survey observed an unacceptable 23.86% inadequacy rate[\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]\u0026mdash;well below the 90% benchmark. Numerous studies concur that in-patients achieve substantially lower cleansing quality than out-patients[\u003cspan additionalcitationids=\"CR31\" citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]..\u003c/p\u003e \u003cp\u003eIBP exerts a cascade of adverse effects: it reduces diagnostic accuracy, increases procedural complications, and drives up healthcare costs. Residual stool obscures the mucosa, slows cecal intubation, and hides polyps and cancers, leading to missed lesions [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. Poor cleansing prolongs inspection time, heightens technical difficulty, and elevates the risk of perforation and other adverse events[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. When the first examination is compromised, patients must return for a repeat colonoscopy or undergo alternative, often more expensive, diagnostic tests, amplifying psychological distress and institutional workload[\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. Consequently, systematic identification of IBP risk factors and the development of effective, targeted interventions are essential to improve colonoscopic quality and efficiency.\u003c/p\u003e \u003cp\u003eTo tackle this problem, investigators have developed a range of risk-prediction algorithms that flag patients at high risk of IBP before colonoscopy, thereby enabling targeted interventions to improve cleansing quality. Published models include traditional logistic-regression equations and easy-to-use scoring systems as well as machine-learning approaches that capture non-linear relationships and high-dimensional interactions[\u003cspan additionalcitationids=\"CR39 CR40 CR41 CR42\" citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e]. Examples include a validated geriatric-specific risk score [\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e], the questionnaire-based PREPA-CO prediction tool[\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e], and several machine-learning classifiers that outperform conventional statistics in internal validations[\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e]. Although these efforts are promising, none has been prospectively tested in the magnesium-sulfate inpatient population treated at Zhongshan Hospital, a national referral centre for digestive diseases. Equally important, most existing models rely solely on baseline characteristics or pre-laxative variables and ignore the pivotal role of nursing staff in real-time education, adherence coaching, and symptom management during the preparation process.\u003c/p\u003e \u003cp\u003eNurse-led educational interventions have been shown to markedly improve patients\u0026rsquo; acceptance of the salty-bitter taste of magnesium sulfate and, more importantly, their actual adherence to the split-dose regimen\u0026mdash;benefits that are especially pronounced among elderly, cognitively impaired, or poly-medicated inpatients[\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e]. Bedside nursing assessments capture real-time information on taste tolerance, stool quality, and adverse events (nausea, bloating, cramps), enabling immediate corrective actions: on-the-spot re-education, psychological support, extra fluids, or timely anti-emetics. This dynamic loop has been proven to cut IBP rates. Yet contemporary prediction models continue to ignore nursing variables\u0026mdash;taste acceptability, adherence checks, stool reports\u0026mdash;and overlook nurses as the pivotal link between risk prediction and in-process intervention. Consequently, these models offer limited clinical traction and fail to guide frontline staff on when and how to intervene.\u003c/p\u003e \u003cp\u003eCommonly used bowel-cleansing agents before colonoscopy include polyethylene glycol (PEG), magnesium sulfate, and sodium phosphate. PEG is considered to have superior efficacy and safety, but its large volume often leads to poor tolerability, especially in older adults, which can reduce adherence. Split-dose regimens may partially address this issue; however, many patients are unwilling to wake up during the night to take the second dose. Moreover, the higher cost of PEG, limited availability, and greater economic burden on patients make it difficult to widely implement in remote or resource-limited settings.\u003c/p\u003e \u003cp\u003eAs the world\u0026rsquo;s largest developing country, China faces uneven distribution of medical resources. In areas with limited health insurance coverage or high out-of-pocket expenses, cost-effectiveness becomes a key factor in selecting bowel preparation agents. Magnesium sulfate remains widely used due to its low cost, small volume, and easy accessibility, offering a significant economic advantage. At our institution, simethicone is routinely co-administered to reduce foam and improve visualization, partially compensating for the limitations of magnesium sulfate and balancing cost-effectiveness with clinical outcomes.\u003c/p\u003e \u003cp\u003eHowever, magnesium sulfate has a bitter taste and frequently causes nausea and vomiting, which further increases the risk of IBP. Unfortunately, most existing prediction models were developed based on Western outpatient populations using PEG-based regimens. These models do not account for taste-related feedback, fluctuating adherence, or the potential impact of nursing interventions specific to magnesium sulfate use. Additionally, Chinese populations may differ from Western populations in physiological characteristics, dietary habits, lifestyle, and medication adherence. Inpatients, in particular, are often affected by polypharmacy and reduced mobility, all of which may influence model performance [\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e] .\u003c/p\u003e \u003cp\u003eAs a result, the variable weights and cut-off values derived from PEG-based models cannot be directly applied to the magnesium sulfate inpatient setting. Furthermore, most models rely solely on static information available at admission, ignoring real-time patient feedback after medication intake, which limits their sensitivity and ability to support dynamic clinical management. Since nurses are the primary providers of bowel preparation education and play a central role in assessing and managing patient adherence, there is an urgent need to explore how nursing-related variables can enhance predictive performance.\u003c/p\u003e \u003cp\u003e To address these gaps, our research team previously used a cohort of 806 in-patients who received oral magnesium sulfate at the Endoscopy Center of Zhongshan Hospital, Fudan University from January to June 2024 to build three prediction models: a single-stage score (model-1) and two two-stage dynamic models (model-2 and model-3). The first-generation model-1 applied LASSO-logistic regression to seven static variables\u0026mdash;IBD, cardiac or renal insufficiency, electrolyte disturbance, relevant medication history, last meal before bowel preparation, laxative preference, and previous failed preparation\u0026mdash;to generate a simple integer score; it achieved an internal AUC of 0.836 and allowed rapid screening but could not capture post-dose changes. Model-2 retained six of these variables (electrolyte disturbance was removed) and introduced a \u0026ldquo;before versus after dosing\u0026rdquo; framework by adding three dynamic indices collected four hours after the first dose: taste acceptability, adverse events (nausea, vomiting, abdominal pain), and self-reported adherence (whether the full laxative dose was taken). Building on model-2, model-3 added neurological comorbidity, colonic stricture, and last stool character, yielding a final -variable combination of \u0026ldquo;static\u0026thinsp;+\u0026thinsp;dynamic\u0026thinsp;+\u0026thinsp;nursing assessment\u0026rdquo;. In internal validation, the AUC increased from 0.693 with model-2 to 0.719 with model-3, sensitivity improved markedly, and both calibration and net clinical benefit remained favorable, indicating that real-time information provides substantial predictive gain.\u003c/p\u003e \u003cp\u003eEven when prediction models perform well in their derivation cohorts, differences in study populations\u0026mdash;such as ethnicity, geographic region, or healthcare system\u0026mdash;and variations in data collection can erode their accuracy in independent external datasets[\u003cspan additionalcitationids=\"CR50\" citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e]. Moreover, the single-centre, retrospective design of our preliminary work may have overestimated performance and limited wider applicability. Prospective external validation is therefore essential to establish a model\u0026rsquo;s reproducibility and generalisability, and constitutes the critical step toward safe clinical implementation and broader dissemination[\u003cspan additionalcitationids=\"CR53\" citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e].\u003c/p\u003e \u003cp\u003e Accordingly, we designed a prospective, continuously-enrolled external-validation study that builds on our earlier methodological iterations. By independently comparing the discrimination, calibration, and net clinical benefit of model-1, model-2, and model-3 in a real-world inpatient population at our own institution, the investigation seeks to answer which version is best suited to our magnesium-sulfate setting and to identify the optimal IBP risk-assessment tool for hospitalized adults undergoing colonoscopy. Beyond providing clinicians with an evidence-based instrument, the project will give nurses a quantifiable metric for early recognition of high-risk patients and for tailoring bedside interventions. Rather than simply testing three models in parallel, the study specifically examines whether model-3\u0026mdash;through its three dynamic nursing indicators (taste acceptability, self-reported adherence, and last stool character)\u0026mdash;achieves clinically meaningful gains in discrimination, calibration, and decision-curve net benefit over model-1 and model-2. Embedding these nursing assessments and real-time response data into the final algorithm should fill a key research gap and establish the clinical utility and scalability of an IBP prediction model that integrates proactive nursing care into routine magnesium-sulfate bowel preparation.\u003c/p\u003e"},{"header":"2 Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Study design and model derivation history\u003c/h2\u003e \u003cp\u003eThis prospective, single-centre cohort validation study was designed to conduct an external evaluation of three previously developed prediction models for IBP following a magnesium-sulfate regimen (model-1, model-2, model-3). All procedures were conducted in strict accordance with the TRIPOD (Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis) statement and the PROBAST (Prediction model Risk Of Bias ASsessment Tool) guidelines to guarantee methodological transparency and full reproducibility of the findings.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Data source and ethics\u003c/h2\u003e \u003cp\u003eConsecutive in-patients scheduled for colonoscopy at the Endoscopy Centre of Zhongshan Hospital, Fudan University, were prospectively enrolled from July 2024 to March 2025. The study protocol was approved by the hospital ethics committee (ref. B2024-278R) and registered at the Chinese Clinical Trial Registry (ChiCTR2400094906). Written informed consent was obtained from every participant. All data were extracted from the hospital\u0026rsquo;s electronic medical-record system, de-identified, and stored in an encrypted institutional database to guarantee patient confidentiality and data integrity.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Participants\u003c/h2\u003e \u003cp\u003eEligibility criteria were: (1) age\u0026thinsp;\u0026ge;\u0026thinsp;18 years; (2) scheduled for elective colonoscopy; and (3) completed the institutional split-dose magnesium-sulfate bowel-preparation protocol while hospitalized. Exclusion criteria comprised: (1) pregnancy, severe cognitive impairment, or active psychosis; (2) incomplete colonoscopy or preparation terminated for reasons other than quality; and (3)\u0026thinsp;\u0026gt;\u0026thinsp;5% missing values for key predictors.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Outcome definition\u003c/h2\u003e \u003cp\u003eThe primary endpoint was IBP. Preparation quality was quantified with BBPS [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. The BBPS divides the colon into three segments\u0026mdash;right (caecum/ascending), transverse (including hepatic and splenic flexures), and left (descending/sigmoid/rectum)\u0026mdash;each scored 0\u0026ndash;3: 0\u0026thinsp;=\u0026thinsp;unprepared, mucosa invisible; 1\u0026thinsp;=\u0026thinsp;major staining/opaque fluid, partial visualization; 2\u0026thinsp;=\u0026thinsp;minor residual staining/clear fluid, good visualization; 3\u0026thinsp;=\u0026thinsp;entire mucosa seen, virtually no residue. Total BBPS ranges 0\u0026ndash;9. Immediately after extubation, an experienced endoscopist assigned segmental scores; values were extracted in real time from the institutional endoscopy reporting system.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Sample-size calculation\u003c/h2\u003e \u003cp\u003eThis external-validation study was powered under a dual framework: (i) events-per-variable (EPV) rules and (ii) minimum-event recommendations for validation cohorts.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003e1) Outcome: inadequate preparation defined as BBPS \u003c 6 or any segment score \u003c 2.\u003c/h3\u003e\n\u003cp\u003e2) Predictors: the largest model contains 12 variables; adopting a liberal EPV\u0026thinsp;\u0026ge;\u0026thinsp;10 for validation requires\u0026thinsp;\u0026ge;\u0026thinsp;120 events, whereas methodological guidance sets a floor of 100 events. We adopted the stricter figure, targeting 120 events.\u003c/p\u003e \u003cp\u003e3) Expected incidence: published Chinese in-patient rates for magnesium-sulfate regimens range 10\u0026ndash;15%; using the conservative π\u0026thinsp;=\u0026thinsp;10%, N\u0026thinsp;=\u0026thinsp;120 / 0.10\u0026thinsp;=\u0026thinsp;1 200. After allowing 5% missing data, 1 260 consecutive participants will be enrolled to ensure adequate power.\u003c/p\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.6 Statistical analysis\u003c/h2\u003e \u003cp\u003e All analyses were conducted in accordance with the TRIPOD and PROBAST guidelines using R version 4.3.2.\u003c/p\u003e \u003cp\u003eContinuous variables are presented as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;SD or median (interquartile range), and categorical variables as number (percentage). Between-group comparisons were performed with the χ\u0026sup2; test or Mann\u0026ndash;Whitney U test, as appropriate.\u003c/p\u003e \u003cp\u003eThe three prediction models were applied to the external-validation cohort. Discrimination was quantified with the area under the receiver-operating-characteristic curve (AUC); higher AUC indicates better separation of individuals with and without IBP. Pairwise AUC comparisons were carried out using DeLong test.\u003c/p\u003e \u003cp\u003eThe optimal cut-off for each model was selected by maximising the Youden index (sensitivity\u0026thinsp;+\u0026thinsp;specificity-1), and the corresponding sensitivity, specificity and overall accuracy were reported.\u003c/p\u003e \u003cp\u003eCalibration was evaluated graphically with calibration plots and tested with the Hosmer\u0026ndash;Lemeshow goodness-of-fit test; P\u0026thinsp;\u0026gt;\u0026thinsp;0.05 indicates adequate calibration.\u003c/p\u003e \u003cp\u003eDecision-curve analysis was performed over threshold probabilities of 5%\u0026ndash;95% to estimate net clinical benefit relative to \u0026ldquo;treat-all\u0026rdquo; and \u0026ldquo;treat-none\u0026rdquo; strategies. A model was considered clinically useful when its net-benefit curve lay above both reference lines.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e2.7 Data collection and quality control\u003c/h2\u003e \u003cp\u003eA closed-loop \u0026ldquo;extract\u0026ndash;verify\u0026ndash;enter\u0026rdquo; team\u0026mdash;comprising two research nurses (blinded to the study hypothesis) and one data manager\u0026mdash;was formed. All variables were extracted independently in duplicate; discrepancies were resolved by the data manager. Data sources were:\u003c/p\u003e \u003cp\u003e \u003col\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003e Real-time BBPS scores recorded at extubation in the endoscopy theatre database;\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003e Electronic medical records (EMR);\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003e Structured nursing assessment forms uploaded to the Clinical Care Classification (CCC) system, with missing items completed by telephone follow-up after discharge.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003c/ol\u003e \u003c/p\u003e \u003cp\u003eThe following domains were captured: demographics, IBD, cardiac/renal insufficiency, neurological comorbidity, colonic stricture, electrolyte disturbance, medications (opioids, tricyclic antidepressants, anaesthetics), last pre-preparation meal type, laxative preference, previous failed preparation, taste acceptability (3-level Likert), adverse events (nausea, vomiting, abdominal pain), self-reported compliance (full dose ingested or not), and final stool colour (yellow-watery vs other). Nursing-assessed variables comprised laxative preference, taste acceptability, compliance, adverse events and final stool colour.\u003c/p\u003e \u003cp\u003eDaily 10% random audits maintained error rates\u0026thinsp;\u0026lt;\u0026thinsp;2%. Variables with \u0026ge;\u0026thinsp;1% missingness were multiply imputed (MICE); all changes were logged. Inter-rater agreement was high (κ\u0026thinsp;\u0026ge;\u0026thinsp;0.82). After database lock, an independent reviewer re-checked 20% of records against source documents, and an external statistician conducted the analyses.\u003c/p\u003e \u003c/div\u003e"},{"header":"3 Results","content":"\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Study flow and baseline characteristics\u003c/h2\u003e \u003cp\u003eTo examine the real-world transportability of the three previously derived models, we conducted a prospective, consecutive-enrolment external-validation study among in-patients at our centre. Between July 2024 and March 2025, 977 orally prepared, magnesium-sulfate-treated patients aged\u0026thinsp;\u0026ge;\u0026thinsp;18 years were recruited and analysed (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). IBP (BBPS\u0026thinsp;\u0026lt;\u0026thinsp;6 or any segment\u0026thinsp;\u0026lt;\u0026thinsp;2) occurred in 107 participants (10.95%).\u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e summarises baseline characteristics for the total cohort and stratified by preparation quality. Patients with IBP were older (66.4\u0026thinsp;\u0026plusmn;\u0026thinsp;12.1 vs 61.0\u0026thinsp;\u0026plusmn;\u0026thinsp;11.7 years, P\u0026thinsp;=\u0026thinsp;0.008), more frequently had stroke/dementia (25.9% vs 6.7%, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001), reported lower acceptability of magnesium-sulfate taste (37.0% vs 12.1%, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and were less likely to pass yellow watery stool as the final lavage (55.6% vs 19.8%, P\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eClinical and demographics characteristics of study population(n\u0026thinsp;=\u0026thinsp;977)\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBaseline characteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOverall(n\u0026thinsp;=\u0026thinsp;977)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eInadequate bowel preparation(n\u0026thinsp;=\u0026thinsp;170)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAdequate bowel preparation(n\u0026thinsp;=\u0026thinsp;807)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003estatistical value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003ep value *\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge (years), median (IQR)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e63.19 (54.98\u0026ndash;70.06)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e65.91 (56.38\u0026ndash;72.07)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e62.98 (54.59\u0026ndash;69.87)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e4.589\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.032\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSex,no.(%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.156\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.693\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003emale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e605 (61.92)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e103 (60.59)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e502 (62.21)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003efemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e372 (38.08)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e67 (39.41)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e305 (37.79%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBMI (kg/m\u0026sup2;), median (IQR)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e23.98 (21.93\u0026ndash;26.07)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e24.11 (22.22\u0026ndash;25.69)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e23.96 (21.91\u0026ndash;26.09)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.014\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.906\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSerum potassium (mmol/L), median (IQR)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4.20 (3.99\u0026ndash;4.49)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4.20 (4.00-4.49)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4.20 (3.99\u0026ndash;4.48)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.881\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.348\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSerum sodium (mmol/L), median (IQR)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e141.98(140.02-143.94)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e141.06 (139.96-143.06)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e141.99 (140.03-143.95)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1.847\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.174\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSerum chloride (mmol/L), median (IQR)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e104.95 (102.97-106.02)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e104.93 (102.98\u0026ndash;106.00)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e104.96 (102.97-106.03)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.067\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.795\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEducation level, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1.993\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.85\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePrimary school or below\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e88 (9.01)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e13 (7.65)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e75 (9.29)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eJunior high school\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e259 (26.51)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e47 (27.65)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e212 (26.27)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSenior high school\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e185 (18.94)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e29 (17.06)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e156 (19.33)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSecondary vocational school\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e80 (8.19)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e14 (8.24)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e66 (8.18)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAssociate degree\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e115 (11.77)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e18 (10.59)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e97 (12.02)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBachelor\u0026rsquo;s degree or above\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e250 (25.59)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e49 (28.82)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e201 (24.91)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDiabetes mellitus, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e113 (11.57)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e25 (14.71)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e88 (10.90)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1.984\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.159\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHypertension, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e389 (39.82)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e67 (39.41)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e322 (39.90)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.014\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.906\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNeurological comorbidities (stroke or dementia), n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e26 (2.66)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e6 (3.53)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e20 (2.48)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.599\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.439\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eChronic constipation, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e105 (10.75)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e28 (16.47)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e77 (9.54)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e7.028\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.008\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInflammatory bowel disease, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e34 (3.48)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e8 (4.71)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e26 (3.22)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.921\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.337\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHepatic insufficiency, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e65 (6.65)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7 (4.12)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e58 (7.19)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2.13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.144\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSevere cardiac or renal insufficiency, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e38 (3.89)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e10 (5.88)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e28 (3.47)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2.187\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.139\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGastroparesis, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e39 (3.99)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e6 (3.53)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e33 (4.09)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.115\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.735\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSevere colonic stricture, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e47 (4.81)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e15 (8.82)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e32 (3.97)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e7.238\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.007\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eElectrolyte disturbances, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e97 (9.93)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e8 (4.71)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e89 (11.03)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e6.277\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.012\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedication use (opioids, tricyclic antidepressants, anesthetics), n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e74 (7.57)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e20 (11.76)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e54 (6.69)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e5.163\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.023\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHistory of abdominal surgery n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e251 (25.69)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e46 (27.06)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e205 (25.40)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.202\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.653\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLast meal type before bowel preparation n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e11.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.011\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClear liquid\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e359 (36.75)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e79 (46.47)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e280 (34.70)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFull liquid\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e94 (9.62)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e19 (11.18)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e75 (9.29)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSoft diet\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e27 (2.76)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5 (2.94)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e22 (2.73)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eregular diet\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e497 (50.87)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e67 (39.41)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e430 (53.28)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eExpressed laxative preference, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e96 (9.83)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e38 (22.35)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e58 (7.19)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e36.45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHistory of inadequate bowel preparation, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e67 (6.86)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e33 (19.41)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e34 (4.21)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e50.782\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTaste acceptability of magnesium sulfate, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e11.515\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.003\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ereadily acceptable\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e734 (75.13)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e118 (69.41)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e616 (76.33)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eneutral / tolerable\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e171 (17.50)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e29 (17.06)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e142 (17.60)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eunacceptable /intolerable\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e72 (7.37)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e23 (13.53)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e49 (6.07)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCompleted full laxative dose, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e933 (95.50)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e158 (92.94)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e775 (96.03)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e3.125\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.077\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAdverse events with oral laxatives, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e60 (6.14)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e11 (6.47)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e49 (6.07)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.039\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.844\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInsufficient physical exercise during preparation, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e56 (5.73)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e10 (5.88)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e46 (5.70)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.009\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.926\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePassage of yellow-watery stool as final lavage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e885 (90.58)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e119 (70.00)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e766 (94.92)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e102.226\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Comparative model performance\u003c/h2\u003e \u003cp\u003eTo comprehensively evaluate the predictive utility of each model for IBP, we analysed four complementary dimensions: discrimination, performance metrics at the optimal Youden cut-off, calibration, and clinical utility. Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e summarises the overall performance of the three models in the external validation cohort of 977 patients.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePerformance summary of the three models in the 977-patient external validation cohort\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"14\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c13\" colnum=\"13\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c14\" colnum=\"14\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003emodel\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCutoff\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAUC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAUC.low\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eAUC.up\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eACC(%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eSEN(%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eSPE(%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003ePPV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003eNPV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003eAIC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e \u003cp\u003eBIC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c13\"\u003e \u003cp\u003eRMSE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c14\"\u003e \u003cp\u003eR2_Tjur\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModel-1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.707\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.665\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.749\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e76.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e81.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e48.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.88\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e825.98\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e835.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e0.36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c14\"\u003e \u003cp\u003e0.09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModel-2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-15.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.699\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.656\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.742\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e66.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e68.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e60.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.89\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e829.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e838.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e0.36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c14\"\u003e \u003cp\u003e0.09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModel-3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-12.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.785\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.746\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.824\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e79.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e83.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e60.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e0.91\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e0.44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e753.70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e763.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e0.34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c14\"\u003e \u003cp\u003e0.18\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Discrimination performance\u003c/h2\u003e \u003cp\u003eThe area under the receiver-operating-characteristic curve (AUC) served as the primary metric for discrimination, with DeLong\u0026rsquo;s test used to evaluate pairwise differences between models. In the external validation cohort, model-3 achieved an AUC of 0.785 (95% CI 0.746\u0026ndash;0.824), substantially higher than model-2 (0.681, 95% CI 0.635\u0026ndash;0.725) and model-1 (0.679, 95% CI 0.633\u0026ndash;0.725) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The incremental AUC for model-3 versus model-1 was 0.106 (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and for model-3 versus model-2 it was 0.104 (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001), demonstrating that the additional clinical variables integrated into model-3 confer significantly superior discriminative capacity for predicting IBP among multicentre in-patients.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Performance at optimal Youden cut-offs\u003c/h2\u003e \u003cp\u003eAt the cut-off maximising the Youden index, model-3 achieved higher sensitivity, specificity and overall accuracy than both model-1 and model-2, corroborating its superior discriminative performance.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e3.5 Multi-metric comparison\u003c/h2\u003e \u003cp\u003eA radar chart (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) simultaneously displays sensitivity, specificity, accuracy, PPV and NPV for the three models. Model-3 occupies the largest area across all five dimensions, providing visual confirmation that it outperforms model-1 and model-2 and reinforcing its selection as the optimal prediction tool.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e3.6 Calibration assessment\u003c/h2\u003e \u003cp\u003eCalibration plots were constructed to compare predicted versus observed probabilities of inadequate bowel preparation. The x-axis represents model-predicted risk; the y-axis shows the corresponding observed event frequency. A bias-corrected curve lying above the ideal diagonal indicates over-estimation, whereas a curve below the diagonal signifies under-estimation.\u003c/p\u003e \u003cp\u003eAs shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, all three models demonstrated satisfactory agreement between prediction and observation. Model-3\u0026rsquo;s calibration curve lay closest to the ideal line, indicating the highest fidelity. Curves for model-1 and model-2 were also acceptable, but deviated slightly more from the reference.\u003c/p\u003e \u003cp\u003eHosmer-Lemeshow tests corroborated these visual impressions:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003e- Model-1: χ\u0026sup2; = 1.852, df\u0026thinsp;=\u0026thinsp;5, P\u0026thinsp;=\u0026thinsp;0.867\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e- Model-2: χ\u0026sup2; = 1.352, df\u0026thinsp;=\u0026thinsp;3, P\u0026thinsp;=\u0026thinsp;0.717\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e- Model-3: χ\u0026sup2; = 6.214, df\u0026thinsp;=\u0026thinsp;5, P\u0026thinsp;=\u0026thinsp;0.286\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eAll P-values\u0026thinsp;\u0026gt;\u0026thinsp;0.05 confirm adequate calibration and reinforce the reliability of model-3 for predicting IBP risk in this external cohort.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e3.7 Clinical utility\u003c/h2\u003e \u003cp\u003eDecision-curve analysis (DCA) quantified net benefit across clinically plausible threshold probabilities (pₜ). The x-axis represents pₜ\u0026mdash;the probability above which a clinician would implement an intervention; the y-axis shows net benefit, calculated as the true-positive fraction minus the false-positive fraction weighted by the odds of pₜ.\u003c/p\u003e \u003cp\u003eAll three models achieved positive net benefit within discrete pₜ ranges:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003e- Model-1: 0.549\u0026ndash;0.914\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e- Model-2: 0.517\u0026ndash;0.923\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003e- Model-3: 0.401\u0026ndash;0.982\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eModel-3 provided the widest useful interval and its curve remained above both \u0026ldquo;treat-all\u0026rdquo; and \u0026ldquo;treat-none\u0026rdquo; reference lines throughout this range (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e), indicating superior clinical value across diverse decision contexts. The narrower benefit windows of model-1 and model-2 imply that model-3 is better positioned to avert adverse outcomes, enhance therapeutic yield, and optimise resource allocation. Consequently, model-3 is the preferred IBP risk-assessment tool for adult in-patients undergoing colonoscopy in our centre.\u003c/p\u003e \u003c/div\u003e"},{"header":"4 Discussion","content":"\u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003e4.1 Model-3 is the optimal tool for predicting IBP under our hospital\u0026rsquo;s magnesium-sulfate regimen\u003c/h2\u003e \u003cp\u003eIn this external-validation cohort, three candidate models with differing variable sets were compared for their ability to quantify IBP risk. Goodness-of-fit tests for all models indicated close agreement between predicted probabilities and observed event rates, underscoring robust stability and generalizability. Calibration plots and decision-curve analysis further supported clinical utility across a wide range of risk thresholds.\u003c/p\u003e \u003cp\u003eModel-3, which integrates comprehensive clinical variables, outperformed the earlier in-house models in discrimination, calibration, and net clinical benefit. Its AUC of 0.785 was significantly higher than that of model-1 (0.679) and model-2 (0.681), representing an increment\u0026thinsp;\u0026gt;\u0026thinsp;0.10. Decision-curve analysis showed an extended net-benefit interval of 0.401\u0026ndash;0.982, implying additional value even when the risk threshold is \u0026lt;\u0026thinsp;50%. The high sensitivity (0.838) paired with moderate specificity (0.606) enables reliable identification of high-risk patients while avoiding excessive intervention. A goodness-of-fit P-value of 0.286 and a calibration slope close to unity confirm that predicted probabilities closely mirror actual risk. These gains reflect the incremental value of adding neurological comorbidities, colonic stricture, and passage of yellow watery stool\u0026mdash;variables that capture key determinants of preparation quality not included in earlier iterations.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003e4.2 Clinical implications of model optimisation\u003c/h2\u003e \u003cp\u003eBuilding on model-2, model-3 explicitly incorporates neurological comorbidities and procedure-related factors (acceptability of laxative taste, passage of yellow watery stool). Neurological disease may compromise bowel-preparation adherence through cognitive impairment and/or autonomic dysregulation that slows intestinal transit, whereas colonic strictures directly limit cleansing efficacy [\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e]. The presence of yellow watery effluent\u0026mdash;an immediate, patient-recognisable sign that the final lavage is approaching clear\u0026mdash;was added as a dynamic indicator of real-time preparation quality, thereby sharpening the model\u0026rsquo;s discriminative ability [\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e]. These variables are rarely captured by traditional scores; their inclusion not only improved predictive performance but also furnishes clinicians with actionable targets (e.g., pre-emptive dose modification or intensified education) [\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e]. Model-3\u0026rsquo;s AUC surpassed previously reported values for the Aronchick score (0.72) and the Boston prediction model (0.74). Moreover, by introducing decision-curve analysis to the bowel-preparation literature, we provide the first evidence in a Chinese cohort that meaningful net benefit persists at low-to-moderate threshold probabilities, mitigating both over- and under-intervention. This advantage probably reflects a more comprehensive variable set aligned with complex clinical realities. Nevertheless, the derived cut-off (\u0026minus;\u0026thinsp;12.500) should be interpreted judiciously and never applied mechanistically.\u003c/p\u003e \u003cp\u003eCompared with previously published models, our Model-3 demonstrated superior discriminative performance. For instance, Afecto et al. externally validated two commonly cited prediction models (Model 1 and Model 2) in a Portuguese tertiary hospital population, reporting AUCs of only 0.62 for both, with low positive predictive values and limited clinical utility in identifying inadequate bowel preparation[\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e]. Similarly, the PREPA-CO score, developed by Berger et al. in a French prospective cohort, achieved an AUC of 0.621 using a self-administered questionnaire-based approach[\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e]. While these models identified overlapping risk factors such as diabetes, prior abdominal surgery, and history of inadequate preparation, they lacked real-time nursing assessment variables and were not validated in inpatient populations receiving magnesium sulfate. In contrast, Model-3 integrates both static and dynamic variables\u0026mdash;including nurse-documented taste acceptability, self-reported adherence, and final stool character\u0026mdash;resulting in significantly improved predictive accuracy and clinical applicability in our Chinese inpatient cohort.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003e4.3 Limitations of model-1 and model-2\u003c/h2\u003e \u003cp\u003eAlthough the AUCs of model-1 and model-2 were only marginally lower (0.679\u0026ndash;0.681), both models retained positive net benefit within specific threshold-probability ranges, implying contextual utility. Model-1, in particular, combined high sensitivity with low specificity, making it potentially useful as a screening tool for high-risk cohorts; however, this imbalance inevitably inflates the false-positive rate. Calibration curves revealed a systematic underestimation of risk, especially in low-risk patients, suggesting that neither model fully captures the true event rate. This under-estimation is plausibly attributable to the omission of well-established predictors such as obesity and diabetes mellitus\u0026mdash;factors that have been independently associated with inadequate bowel preparation[\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e]. Their exclusion may lead to insufficient risk attribution in selected high-risk individuals. Furthermore, the narrow range of threshold probabilities over which either model yields clinical benefit indicates limited added value in routine decision-making\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec23\" class=\"Section2\"\u003e \u003ch2\u003e4.4 Robustness and transportability of external validation\u003c/h2\u003e \u003cp\u003eConsecutive enrollment coupled with prospective data collected from July 2024 to March 2025 guaranteed both timeliness and representativeness of the validation cohort. Model-3\u0026rsquo;s excellent calibration and discrimination indicate high local applicability and suggest that it should perform comparably in similar tertiary-care settings. Its parsimonious variable set further facilitates seamless integration into routine clinical pathways, offering real-time decision support for individualized bowel-preparation interventions.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003e4.5 Limitations and Future Directions\u003c/h2\u003e \u003cp\u003eSeveral limitations should be acknowledged. First, the study was confined to in-patients; generalizability to out-patients or community populations remains to be established. The representativeness and size of the external-validation cohort may also constrain wider applicability. Second, variables such as \u0026ldquo;taste acceptability\u0026rdquo; and \u0026ldquo;dietary preference\u0026rdquo; are inherently subjective and susceptible to recall bias. Future iterations could incorporate objective biomarkers\u0026mdash;e.g., plasma osmolality, electrolyte shifts\u0026mdash;while also accounting for heterogeneity in preparation regimens (dosage, timing, adjunct medications) and the intestinal microbiome, all of which may influence predictive performance. Nursing-related metrics such as education frequency could likewise be integrated to refine model precision. Finally, the ability of the model to predict longer-term outcomes (re-hospitalization, mortality) was not examined and warrants prospective evaluation.\u003c/p\u003e \u003c/div\u003e"},{"header":"5.Conclusion","content":"\u003cp\u003eIn this multicenter validation study, three prediction models (model-1, model-2 and model-3) were externally compared in hospitalized adults. Model-3 demonstrated superior discrimination, calibration and clinical utility relative to both model-1 and model-2. Consequently, model-3 is considered the optimal instrument for pre-colonoscopy IBP risk assessment in adult in-patients at our institution. Next steps will evaluate the feasibility and cost-effectiveness of integrating model-3 into the electronic health record to guide individualized interventions, thereby providing an evidence-based foundation for precision management of in-patient bowel preparation before colonoscopy.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u0026nbsp;\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe are grateful to all the patients who gave consent to participate in the study and the research team who endeavoured to ensure that the study is successfully carried out.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eQ.Y.S, M.Z. M.D conceived and designed the study under the supervision of Y.W. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eQ.Y.S, M.Z, M.D implemented the study under the supervision of Y.H.S, Y.F.O.\u003c/p\u003e\n\u003cp\u003eQ.Y.S, F.Y.F conducted data analysis. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eQ.Y.S, M.Z, M.D, H.K.S, Y.H.S, Y.F.O, Y.W interpreted the findings. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eM.D, Y.W involved in validation; Q.Y.S, M.Z involved in data curation. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eY.W sought funding. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eQ.Y.S wrote the first draft; all authors reviewed and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the Evidence-Based Nursing Practice Program of the Nursing Department, Zhongshan Hospital, Fudan University (Grant No. HLXZ202404) and the Scientific Research Development Fund of Zhongshan Hospital, Fudan University (Grant No. 2023XKPT08-RC2). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data generated or analysed during this study are included in this published article.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthical approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study protocol was approved by the Ethics Committee of Zhongshan Hospital, Fudan University (Approval No. B2024-278R) in accordance with the ethical guidelines of the Declaration of Helsinki and the relevant regulations of the ethics committee.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll participants received detailed study information and provided written informed consent through the hospital\u0026rsquo;s secure electronic signature platform before any data collection.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWritten informed consent for publication was obtained from all participants through the hospital\u0026rsquo;s electronic consent platform.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data generated or analysed during this study are included in this published article. Further inquiries can be directed to the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTrial registration\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was registered with the Chinese Clinical Trial Registry (ChiCTR) under registration number ChiCTR2400081234.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eSiegel RL, Miller KD, Goding Sauer A, Fedewa SA, Butterly LF, Anderson JC, et al. Colorectal cancer statistics, 2020. Cancer J Clin. 2020;70(3):145\u0026ndash;64. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3322/caac.21601\u003c/span\u003e\u003cspan address=\"10.3322/caac.21601\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIssaka RB, Chan AT, Gupta S, AGA Clinical Practice Update on Risk Stratification for Colorectal Cancer Screening and Post-Polypectomy Surveillance. Expert Rev Gastroenterol. 2023;165(5):1280\u0026ndash;91. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1053/j.gastro.2023.06.033\u003c/span\u003e\u003cspan address=\"10.1053/j.gastro.2023.06.033\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRawla P, Sunkara T, Barsouk A. Epidemiology of colorectal cancer: incidence, mortality, survival, and risk factors. Gastroenterol Rev. 2019;14(2):89\u0026ndash;103. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.5114/pg.2018.81072\u003c/span\u003e\u003cspan address=\"10.5114/pg.2018.81072\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNie J-X, Xie Q, Yuan Y, Liu M-Y, Du J-K, Li N, et al. Rising Incidence of Total and Early-Onset Colorectal Cancer: A Global Perspective on Burden, Risk Factors, and Projections to 2031. J Gastrointest Cancer. 2025;56(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s12029-025-01257-1\u003c/span\u003e\u003cspan address=\"10.1007/s12029-025-01257-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRaphael MJ, Wildgoose P, Servidio-Italiano F, De Vera MA, Brenner D, D\u0026rsquo;Angelo MS, et al. Breaking Down Barriers to Detection and Care in Early-Age-Onset Colorectal Cancer in Canada. Curr Oncol. 2023;30(11):9392\u0026ndash;405. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/curroncol30110680\u003c/span\u003e\u003cspan address=\"10.3390/curroncol30110680\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang Y, undefined u, Han Z, Li X, Huang A, Shi J, et al. Epidemiology and risk factors of colorectal cancer in China. Chin J Cancer Res. 2020;32(6):729\u0026ndash;41. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.21147/j.issn.1000-9604.2020.06.06\u003c/span\u003e\u003cspan address=\"10.21147/j.issn.1000-9604.2020.06.06\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi N, Lu B, Luo C, Cai J, Lu M, Zhang Y, et al. Incidence, mortality, survival, risk factor and screening of colorectal cancer: A comparison among China, Europe, and northern America. Cancer Lett. 2021;522:255\u0026ndash;68. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.canlet.2021.09.034\u003c/span\u003e\u003cspan address=\"10.1016/j.canlet.2021.09.034\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen Y, Chen T, Fang J-Y. Burden of gastrointestinal cancers in China from 1990 to 2019 and projection through 2029. Cancer Lett. 2023;560:216127. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.canlet.2023.216127\u003c/span\u003e\u003cspan address=\"10.1016/j.canlet.2023.216127\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSun C, Liu Y, Huang Y, Li B, Rang W. Colorectal Cancer Incidence and Mortality Trends and Analysis of Risk Factors in China from 2005 to 2015. Int J Gen Med. 2021;9965\u0026ndash;76. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2147/ijgm.s344448\u003c/span\u003e\u003cspan address=\"10.2147/ijgm.s344448\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang CF. Colorectal cancer incidence and mortality in Shanghai 2016 and trend analysis 2002\u0026ndash;2016. Journal of Surgery Concepts \u0026amp; Practice; 2021.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. Cancer J Clin. 2021;71(3):209\u0026ndash;49. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3322/caac.21660\u003c/span\u003e\u003cspan address=\"10.3322/caac.21660\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHassan C, East J, Radaelli F, Spada C, Benamouzig R, Bisschops R, et al. Bowel preparation for colonoscopy: European Society of Gastrointestinal Endoscopy (ESGE) Guideline - Update 2019. Endoscopy. 2019;51(8):775\u0026ndash;94. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1055/a-0959-0505\u003c/span\u003e\u003cspan address=\"10.1055/a-0959-0505\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHagen R, Nguyen MTT, Anderson JC, Birk JW. Navigating Bowel Preparation for Colonoscopy. J Clin Gastroenterol. 2024;59(4):285\u0026ndash;97. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/mcg.0000000000002124\u003c/span\u003e\u003cspan address=\"10.1097/mcg.0000000000002124\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWoo DH, Kim KO, Jeong DE, Nam YJ, Lee SH, Jang BI, et al. Prospective analysis of factors associated with inadequate bowel preparation for colonoscopy in actual clinical practice. Intestinal Res. 2018;16(2):293. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.5217/ir.2018.16.2.293\u003c/span\u003e\u003cspan address=\"10.5217/ir.2018.16.2.293\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKastenberg D, Bertiger G, Brogadir S. Bowel preparation quality scales for colonoscopy. World J Gastroenterol. 2018;24(26):2833\u0026ndash;43. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3748/wjg.v24.i26.2833\u003c/span\u003e\u003cspan address=\"10.3748/wjg.v24.i26.2833\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRai T, Navaneethan U, Gohel T, Podugu A, Thota PN, Kiran RP, et al. Effect of quality of bowel preparation on quality indicators of adenoma detection rates and colonoscopy completion rates. Gastroenterol Rep. 2015;4(2):148\u0026ndash;53. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/gastro/gov002\u003c/span\u003e\u003cspan address=\"10.1093/gastro/gov002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eColorectal Endoscopy Group, Chinese Society of Digestive Endoscopy. Expert consensus on bowel preparation for colonoscopy (2023, Guangzhou). Chin J Dig Endoscopy. 2023;40(6):421\u0026ndash;30.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHern\u0026aacute;ndez G, Gimeno-Garc\u0026iacute;a AZ, Quintero E. Strategies to Improve Inadequate Bowel Preparation for Colonoscopy. Front Med. 2019;6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fmed.2019.00245\u003c/span\u003e\u003cspan address=\"10.3389/fmed.2019.00245\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVoiosu T, Voiosu A, Voiosu R. Bowel preparation for colonoscopy. Curr Opin Gastroenterol. 2016;32(5):385\u0026ndash;6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/mog.0000000000000297\u003c/span\u003e\u003cspan address=\"10.1097/mog.0000000000000297\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eParra-Blanco A, Ruiz A, Alvarez-Lobos M, Amor\u0026oacute;s A, Gana JC, Ib\u0026aacute;\u0026ntilde;ez P, et al. Achieving the best bowel preparation for colonoscopy. World J Gastroenterol. 2014;20(47):17709\u0026ndash;26. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3748/wjg.v20.i47.17709\u003c/span\u003e\u003cspan address=\"10.3748/wjg.v20.i47.17709\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLai EJ, Calderwood AH, Doros G, Fix OK, Jacobson BC. The Boston bowel preparation scale: a valid and reliable instrument for colonoscopy-oriented research. Gastrointest Endosc. 2009;69(3):620\u0026ndash;5. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.gie.2008.05.057\u003c/span\u003e\u003cspan address=\"10.1016/j.gie.2008.05.057\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKutyla MJ, Gray MA, von Hippel C, Hourigan LF, Kendall BJ, Whaley AJ, et al. Improving the Quality of Bowel Preparation: Rewarding Patients for Success or Intensive Patient Education? Dig Dis. 2020;39(2):113\u0026ndash;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1159/000510461\u003c/span\u003e\u003cspan address=\"10.1159/000510461\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLam TYT, Wu PI, Tang RSY, Tse YK, Lau JYW, Wu JCY, et al. Nurse-led reinforced education by mobile messenger improves the quality of bowel preparation of colonoscopy in a population-based colorectal cancer screening program: A randomized controlled trial. Int J Nurs Stud. 2022;133:104301. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ijnurstu.2022.104301\u003c/span\u003e\u003cspan address=\"10.1016/j.ijnurstu.2022.104301\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChan W-K, Saravanan A, Manikam J, Goh K-L, Mahadeva S. Appointment waiting times and education level influence the quality of bowel preparation in adult patients undergoing colonoscopy. BMC Gastroenterol. 2011;11(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/1471-230x-11-86\u003c/span\u003e\u003cspan address=\"10.1186/1471-230x-11-86\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGorelik Y. Volume of fluid consumption during preparation for colonoscopy is possibly the single most important determinant of bowel preparation adequacy. Annals Gastroenterol. 2021. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.20524/aog.2021.0642\u003c/span\u003e\u003cspan address=\"10.20524/aog.2021.0642\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAdler A, Wegscheider K, Lieberman D, Aminalai A, Aschenbeck J, Drossel R, et al. Factors determining the quality of screening colonoscopy: a prospective study on adenoma detection rates, from 12,134 examinations (Berlin colonoscopy project 3, BECOP-3). Gut. 2013;62(2):236\u0026ndash;41. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/gutjnl-2011-300167\u003c/span\u003e\u003cspan address=\"10.1136/gutjnl-2011-300167\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAmitay EL, Niedermaier T, Gies A, Hoffmeister M, Brenner H. Risk Factors of Inadequate Bowel Preparation for Screening Colonoscopy. J Clin Med. 2021;10(12):2740. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/jcm10122740\u003c/span\u003e\u003cspan address=\"10.3390/jcm10122740\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCornella SL, Wentworth BJ, Orton C, Horton BJ, Powell SM. Modelling characteristics of inadequate bowel preparations for colonoscopy. GastroHep. 2020;2(2):72\u0026ndash;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/ygh2.389\u003c/span\u003e\u003cspan address=\"10.1002/ygh2.389\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi J, Qu C, Zhao Q, Fu JN, Du BX, Fan QY, et al. Current status and influencing factors of bowel preparation before colonoscopy. China Med Guide. 2020;18(27):1\u0026ndash;2. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.15912/j.cnki.gocm.2020.27.002\u003c/span\u003e\u003cspan address=\"10.15912/j.cnki.gocm.2020.27.002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGinnaram S, Nugooru S, Tahir D, Marrache K, Jacoby P, Elrod S, et al. S885\u0026emsp;Impact of Bowel Preparation Volume on Inpatient Colonoscopy Outcomes: A Retrospective Analysis. Am J Gastroenterol. 2024;119. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.14309/01.ajg.0001032908.82582.6c\u003c/span\u003e\u003cspan address=\"10.14309/01.ajg.0001032908.82582.6c\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDickson Z, Park SR, Tung W, Douglas SR, Haug R, Shenoy A, et al. 2882\u0026emsp;Analysis of Risk Factors for Poor Bowel Preparation in Hospitalized Patients. Am J Gastroenterol. 2019;114(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.14309/01.ajg.0000601060.91042.cd\u003c/span\u003e\u003cspan address=\"10.14309/01.ajg.0000601060.91042.cd\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMandark G, Townsend C, Almaghrabi M, Alotaibi A, Khanna N, Yan B, et al. 566\u0026emsp;Improving Bowel Preparation Quality for Inpatient Colonoscopies at a Tertiary Hospital. Am J Gastroenterol. 2019;114(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.14309/01.ajg.0000591796.26128.58\u003c/span\u003e\u003cspan address=\"10.14309/01.ajg.0000591796.26128.58\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang SS, Guo XX, Li ZK. Research progress on factors influencing the quality of bowel preparation in patients undergoing colonoscopy. Chin J Mod Nurs. 2021;27(14):1942\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eClark BT, Rustagi T, Laine L. What Level of Bowel Prep Quality Requires Early Repeat Colonoscopy: Systematic Review and Meta-Analysis of the Impact of Preparation Quality on Adenoma Detection Rate. Am J Gastroenterol. 2014;109(11):1714\u0026ndash;23. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/ajg.2014.232\u003c/span\u003e\u003cspan address=\"10.1038/ajg.2014.232\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRay-Offor E, Jebbin N. Risk Factors for Inadequate Bowel Preparation During Colonoscopy in Nigerian Patients. Cureus. 2021. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.7759/cureus.17145\u003c/span\u003e\u003cspan address=\"10.7759/cureus.17145\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMitchell Rn A. Reducing Suboptimal Bowel Preparation. J Qual Health Care Econ. 2021;4(4). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.23880/jqhe-16000232\u003c/span\u003e\u003cspan address=\"10.23880/jqhe-16000232\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZad M, Do CN, Heffernan A, Johnston L, Al-Ansari M. Factors affecting bowel preparation adequacy and procedural time. JGH Open. 2019;4(2):206\u0026ndash;14. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/jgh3.12241\u003c/span\u003e\u003cspan address=\"10.1002/jgh3.12241\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHassan C, Fuccio L, Bruno M, Pagano N, Spada C, Carrara S, et al. A Predictive Model Identifies Patients Most Likely to Have Inadequate Bowel Preparation for Colonoscopy. Clin Gastroenterol Hepatol. 2012;10(5):501\u0026ndash;6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cgh.2011.12.037\u003c/span\u003e\u003cspan address=\"10.1016/j.cgh.2011.12.037\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBeran A, Aboursheid T, Ali AH, Albunni H, Mohamed MF, Vargas A, et al. Risk Factors for Inadequate Bowel Preparation in Colonoscopy: A Comprehensive Systematic Review and Meta-Analysis. Am J Gastroenterol. 2024;119(12):2389\u0026ndash;97. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.14309/ajg.0000000000003073\u003c/span\u003e\u003cspan address=\"10.14309/ajg.0000000000003073\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDik VK, Moons LMG, H\u0026uuml;y\u0026uuml;k M, van der Schaar P, de Vos tot Nederveen Cappel WH, ter Borg PCJ et al. Predicting inadequate bowel preparation for colonoscopy in participants receiving split-dose bowel preparation: development and validation of a prediction score. Gastrointestinal endoscopy. 2015;81(3):665\u0026thinsp;\u0026ndash;\u0026thinsp;72. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.gie.2014.09.066\u003c/span\u003e\u003cspan address=\"10.1016/j.gie.2014.09.066\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGimeno-Garc\u0026iacute;a A, Baute J, Hernandez G, Morales D, Gonzalez-P\u0026eacute;rez C, Nicol\u0026aacute;s-P\u0026eacute;rez D, et al. Risk factors for inadequate bowel preparation: a validated predictive score. Endoscopy. 2017;49(6):536\u0026ndash;43. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1055/s-0043-101683\u003c/span\u003e\u003cspan address=\"10.1055/s-0043-101683\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHigashimori A, Maeda N, Nakatani M, Yamamoto I, Yanagida T, Kin D, et al. Importance of Bowel Habits in Predicting Inadequate Bowel Preparation: A Prospective Observational Study. Dig Dis Sci. 2023;68(11):4148\u0026ndash;55. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10620-023-08089-8\u003c/span\u003e\u003cspan address=\"10.1007/s10620-023-08089-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Y, Wang L, Wu W, Zhang S, Zhang M, She W, et al. Predictors of inadequate bowel preparation in older patients undergoing colonoscopy: A systematic review and meta-analysis. Int J Nurs Stud. 2024;149:104631. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ijnurstu.2023.104631\u003c/span\u003e\u003cspan address=\"10.1016/j.ijnurstu.2023.104631\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu Y, Ou S, Yuan X, Yao J, Yang T. Construction and Validation of Risk Scoring Model for Inadequate Bowel Preparation in Elderly Patients Undergoing Colonoscopy. Therapeutics and Clinical Risk Management. 2025:885\u0026thinsp;\u0026ndash;\u0026thinsp;94. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2147/tcrm.s503389\u003c/span\u003e\u003cspan address=\"10.2147/tcrm.s503389\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBerger A, Cesbron-M\u0026eacute;tivier E, Bertrais S, Olivier A, Becq A, Boursier J, et al. A predictive score of inadequate bowel preparation based on a self-administered questionnaire: PREPA-CO. Clin Res Hepatol Gastroenterol. 2021;45(4):101693. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.clinre.2021.101693\u003c/span\u003e\u003cspan address=\"10.1016/j.clinre.2021.101693\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGu F, Xu J, Du L, Liang H, Zhu J, Lin L, et al. The Machine Learning Model for Predicting Inadequate Bowel Preparation Before Colonoscopy: A Multicenter Prospective Study. Clin Translational Gastroenterol. 2024;15(5). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.14309/ctg.0000000000000694\u003c/span\u003e\u003cspan address=\"10.14309/ctg.0000000000000694\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu J, Jiang W, Yu Y, Gong J, Chen G, Yang Y, et al. Applying machine learning to predict bowel preparation adequacy in elderly patients for colonoscopy: development and validation of a web-based prediction tool. Ann Med. 2025;57(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1080/07853890.2025.2474172\u003c/span\u003e\u003cspan address=\"10.1080/07853890.2025.2474172\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBishop R, Laugharne R, Burrows L, undefined u, Ward S, Eustice S, et al. Laxative use in adults with intellectual disabilities: development of prescribing guidelines. BJPsych Open. 2024;10(3). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1192/bjo.2024.50\u003c/span\u003e\u003cspan address=\"10.1192/bjo.2024.50\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYuan X, Gao H, Liu C, Wang W, Xie J, Zhang Z, et al. External validation of two prediction models for adequate bowel preparation in Asia: a prospective study. Int J Colorectal Dis. 2022;37(6):1223\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00384-022-04156-0\u003c/span\u003e\u003cspan address=\"10.1007/s00384-022-04156-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAfecto E, Ponte A, Fernandes S, Gomes C, Correia Jo\u0026atilde;o P, Carvalho J. Validation and Application of Predictive Models for Inadequate Bowel Preparation in Colonoscopies in a Tertiary Hospital Population. GE - Portuguese J Gastroenterol. 2021;30(2):134\u0026ndash;40. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1159/000520905\u003c/span\u003e\u003cspan address=\"10.1159/000520905\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNing Y, Volovici V, Ong MEH, Goldstein BA, Liu N. A roadmap to fair and trustworthy prediction model validation in healthcare. undefined. 2023. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.48550/ARXIV.2304.03779\u003c/span\u003e\u003cspan address=\"10.48550/ARXIV.2304.03779\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRamspek CL, Jager KJ, Dekker FW, Zoccali C, van Diepen M. External validation of prognostic models: what, why, how, when and where? Clin Kidney J. 2020;14(1):49\u0026ndash;58. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/ckj/sfaa188\u003c/span\u003e\u003cspan address=\"10.1093/ckj/sfaa188\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMann M, Ensor J, Collins G, Riley R. How are published clinical prediction model studies assessing and reporting calibration performance at external validation? OSF, 2024. 2024.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGallitto G, Englert R, Kincses B, Kotikalapudi R, Li J, Hoffschlag K, et al. External validation of machine learning models - registered models and adaptive sample splitting. bioRxiv. 2023. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1101/2023.12.01.569626\u003c/span\u003e\u003cspan address=\"10.1101/2023.12.01.569626\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKumar A, Shenoy V, Buckley MC, Durbin L, Mackey J, Mone A, et al. Endoscopic Disease Activity and Biologic Therapy Are Independent Predictors of Suboptimal Bowel Preparation in Patients with Inflammatory Bowel Disease Undergoing Colonoscopy. Dig Dis Sci. 2022;67(10):4851\u0026ndash;65. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10620-022-07530-8\u003c/span\u003e\u003cspan address=\"10.1007/s10620-022-07530-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLai EJ, Calderwood AH, Doros G, Fix OK, Jacobson BC. The Boston bowel preparation scale: a valid and reliable instrument for colonoscopy-oriented research. Gastrointest Endosc. 2009;69(3 Pt 2):620\u0026ndash;5. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.gie.2008.05.057\u003c/span\u003e\u003cspan address=\"10.1016/j.gie.2008.05.057\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArslanca G, Ayg\u0026uuml;n M. Effect of nurse-performed enhanced patient education on colonoscopy bowel preparation quality. Rev Latinoam Enferm. 2022;30:e3626. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1590/1518-8345.5597.3626\u003c/span\u003e\u003cspan address=\"10.1590/1518-8345.5597.3626\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBerger A, Cesbron-M\u0026eacute;tivier E, Bertrais S, Olivier A, Becq A, Boursier J, et al. A predictive score of inadequate bowel preparation based on a self-administered questionnaire: PREPA-CO. Clin Res Hepatol Gastroenterol. 2021;45(4):101693. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.clinre.2021.101693\u003c/span\u003e\u003cspan address=\"10.1016/j.clinre.2021.101693\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHyun JH, Kim SJ, Park JH, Wie GA, Kim JS, Han KS, et al. Lifestyle Factors and Bowel Preparation for Screening Colonoscopy. Annals coloproctology. 2018;34(4):197\u0026ndash;205. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3393/ac.2018.03.13\u003c/span\u003e\u003cspan address=\"10.3393/ac.2018.03.13\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Colonoscopy, bowel preparation, predictive score","lastPublishedDoi":"10.21203/rs.3.rs-8265970/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8265970/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground and Aims\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eInpatient colonoscopy frequently fails because bowel cleansing is inadequate. Magnesium sulfate is the cathartic of choice in most Chinese hospitals because it costs only a few cents, yet no externally validated tool exists to flag inpatients at high risk of poor preparation. We compared three in-house prediction models to identify the one that best helps nurses and physicians optimize bowel preparation before colonoscopy.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eUsing three previously derived models—Model-1 (seven static variables), Model-2 (nine two-stage variables), and Model-3 (twelve integrated variables)—we conducted a prospective cohort study. Consecutive inpatients aged ≥ 18 years who received a split-dose magnesium sulfate regimen for elective colonoscopy between July 2024 and March 2025 were enrolled. Inadequate bowel preparation was defined as a total Boston Bowel Preparation Scale score \u0026lt; 6 or any segment score \u0026lt; 2. Discrimination (area under the ROC curve, AUC), calibration (calibration plot and Hosmer-Lemeshow test), and clinical utility (decision-curve analysis) were evaluated.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAmong 977 patients, 107 (10.95 %) had inadequate preparation. Model-3 achieved an AUC of 0.785 (95 % CI 0.746–0.824), outperforming Model-1 (AUC 0.679) and Model-2 (AUC 0.681) by 0.106 (P \u0026lt; 0.001). At the optimal cut-off, Model-3 provided 83.8 % sensitivity, 60.6 % specificity, and 79.7 % accuracy. Calibration was good (Hosmer-Lemeshow P = 0.286), and decision-curve analysis showed the widest net-benefit range (threshold probability 0.401–0.982).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn hospitalized patients receiving magnesium sulfate, Model-3—combining baseline risk factors with real-time nursing assessments—offers superior discrimination, calibration, and clinical utility. Embedding this simple score in the electronic health record flags high-risk patients early, triggers tailored education, and improves preparation quality without added cost. Multi-centre implementation studies are warranted to confirm generalizability.\u003c/p\u003e","manuscriptTitle":"Comparative performance and external validation of three different models in predicting inadequate bowel preparation among Chinese inpatients undergoing colonoscopyFirst Name Last","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-22 09:13:29","doi":"10.21203/rs.3.rs-8265970/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"eb8414ed-41f7-4f1b-ad5a-0dcbc761a29f","owner":[],"postedDate":"December 22nd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-01-05T07:39:58+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-22 09:13:29","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8265970","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8265970","identity":"rs-8265970","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}