ChatGPT-4o as an Assist Tool for LUTS Evaluation: Better Prediction of IPSS-QoL and OABSS Scores

ChatGPT-4o as an Assist Tool for LUTS Evaluation: Better Prediction of IPSS-QoL and OABSS Scores | 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 ChatGPT-4o as an Assist Tool for LUTS Evaluation: Better Prediction of IPSS-QoL and OABSS Scores Hoyoung Bae, Jung Hoon Lee, Min Chul Cho, Sangjun Yoo, Hyeon Jeong, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7468583/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 19 Jan, 2026 Read the published version in BMC Urology → Version 1 posted 17 You are reading this latest preprint version Abstract Purpose To evaluate the performance of GPT-4o in estimating International Prostate Symptom Score (IPSS) and Overactive Bladder Symptom Score (OABSS) based on patients’ natural language descriptions and full outpatient records, compared to actual questionnaire scores. Materials and Methods This study included 91 patients, of whom 52 completed IPSS and 77 completed OABSS. GPT-4o was prompted with verbatim symptom statements and full medical records written by a urologist. Predicted scores were compared to actual scores using paired t-tests, weighted Cohen’s kappa for item-level agreement, Spearman’s correlation for total scores, and Bland–Altman plots for bias. Diagnostic classifications (lower urinary tract symptoms [LUTS]: IPSS ≥8; overactive bladder [OAB]: OABSS ≥3 with urgency ≥2) were assessed using McNemar’s test and receiver operating characteristic curve analysis. Results Mean IPSS scores were 13.6 (GPT) and 11.2 (actual) (p = 0.006), while OABSS scores were not significantly different (6.99 vs. 6.86, p = 0.686). Diagnostic agreement was high: LUTS in 42 (actual) vs. 38 (GPT) patients, and OAB in 51 vs. 50 patients. Area under curve was 0.86 for IPSS and 0.91 for OABSS. Kappa values ranged from 0.23–0.81 (IPSS) and 0.44–0.71 (OABSS), with highest concordance in quality of life (QoL) and urgency incontinence. Spearman’s correlation coefficient was 0.60 (IPSS) and 0.70 (OABSS). Accuracy was lower in first-visit patients. Conclusions GPT-4o estimated IPSS and OABSS with clinically acceptable accuracy. Its performance was comparable regarding diagnostic classification, particularly for QoL and OABSS. GPT-4o may complement traditional questionnaires, particularly with missing or incomplete patient-reported data. Artificial intelligence Chat GPT IPSS Large language model Low urinary tract symptoms OABSS Overactive bladder. Figures Figure 1 Figure 2 Figure 3 Introduction Standardized questionnaires, such as the International Prostate Symptom Score (IPSS) and Overactive Bladder Symptom Score (OABSS), are widely used for evaluating lower urinary tract symptoms (LUTS)[1]. However, in real-world clinical settings, a significant proportion of elderly patients experience difficulty completing these self-administered questionnaires accurately. Particularly, patients with limited education or cognitive impairment may find the questionnaires complex or difficult to understand [5]. One study reported that approximately 70% of patients failed to complete the IPSS independently because of difficulties understanding some items [6]. These limitations highlight the challenged posed by traditional paper-based questionnaires, which can be influenced by recall bias and subjectivity. Recently, the IPSS requires patients to recall their urinary frequency and nocturia episodes, which can distort the scoring owing to inaccurate memory or daily variability. Consequently, physicians frequently assist elderly patients in completing each questionnaire item verbally or rely on caregivers, which is time consuming and often reduces the consistency of their responses [5]. Therefore, there is growing interest in automating symptom scoring using natural language-based artificial intelligence (AI). ChatGPT, a large language model (LLM) developed by OpenAI, has demonstrated human-like performance in high-stakes medical examinations, such as the USMLE [8]. Particularly, several studies have shown that the GPT-4 model can effectively interpret free-form patient narratives and generate clinically meaningful outputs [8–10]. Recent studies have evaluated whether ChatGPT can replace or augment traditional patient questionnaires. In one study, ChatGPT-4o was prompted to respond to structured mental health questionnaires (for depression and anxiety) as if it were the patient. The ChatGPT scores showed a high level of agreement with the validated questionnaire results. However, Bland–Altman analysis revealed a systematic bias and wide limits of agreement [9], highlighting the need for rigorous quantitative validation before clinical adoptation. This study used the OpenAI GPT-4o model to estimate IPSS and OABSS solely from patients’ free-form symptom descriptions and electronic medical records (EMRs) without using the questionnaires themselves. By comparing the predicted scores with the actual patient-reported scores, we aimed to assess the accuracy and limitations of this language-based AI model and explore its potential to replace or supplement traditional symptom questionnaires. This study aimed to validate the predictive accuracy of GPT-4o in estimating IPSS and OABSS and evaluate the degree of agreement for each questionnaire item in a clinical setting. Materials and Methods In May 2025, a retrospective analysis was conducted on 91 patients who visited the urology outpatient clinic of a secondary care hospital for the evaluation of LUTS or overactive bladder (OAB). Among them, 52 patients completed the IPSS questionnaire and 77 completed the OABSS questionnaire (some patients completed both questionnaires). The cohort included 55 men (60.4%) and 36 women (39.6%) with a mean age of approximately 71 years (range; 35–86). All patients in the IPSS group were men, whereas the OABSS group included 36 women (Table 1). Data collection The actual self-reported IPSS and OABSS questionnaire scores were collected for each patient. Additionally, free-text symptom descriptions provided by patients and EMRs from corresponding outpatient visits were used as inputs for the GPT-4o model. Symptom descriptions included sentences written by the attending urologist based on the patient’s spoken language during clinical interviews. The EMRs contained a chronologically organized documentation of the patients’ chief complaints, relevant diagnostic findings, and clinical impressions, all of which were used without further modifications. These unstructured text data were entered into the GPT-4o (OpenAI GPT-4 model, May 2025 version), which was used to estimate patients’ responses to eight IPSS items and four OABSS items. The model assigned scores ranging from 0 to each item’s defined maximum (0–5 depending on the item) and generated total scores and diagnostic classifications (LUTS based on IPSS ≥ 8; OAB based on OABSS ≥ 3 with urgency score ≥ 2). All the predictions were made using a zero-shot approach without task-specific fine-tuning. To assess the agreement between the GPT-predicted and patient-reported total scores, Bland–Altman analysis was performed besides correlation analysis. Statistical analysis Statistical comparisons between patient-reported and GPT-estimated sub-item scores and total scores for IPSS and OABSS were performed using paired t-tests. Spearman’s correlation coefficients were calculated to assess the relationship between the two scoring methods and interpreted according to conventional benchmarks proposed by Evans (1996), in which ρ = 0.40–0.59 indicates moderate, 0.60–0.79 indicates strong, and ≥ 0.80 indicates very strong correlation [11]. Agreement for individual items was further evaluated using weighted Cohen’s kappa coefficients, interpreted according to the criteria described by Landis and Koch (1977), in which κ = 0.41–0.60 indicates moderate agreement, 0.61–0.80 substantial agreement, and > 0.80 almost perfect agreement [12]. For diagnostic classification, LUTS was defined as an IPSS total score ≥ 8, and OAB was defined as an OABSS total score ≥ 3 with an urgency item (Q3) score ≥ 2. Discrepancies between the GPT-predicted and patient-reported diagnostic categories were assessed using McNemar’s test. Overall diagnostic performance was evaluated using receiver operating characteristic (ROC) curve analysis, and the area under the curve (AUC) was calculated. Moreover, Bland–Altman analysis examined the degree of numerical agreement between the GPT-estimated and patient-reported total scores and identified any systematic bias or clinically relevant discrepancies. All statistical analyses were conducted using IBM SPSS Statistics version 29 (IBM Corp., Armonk, NY, USA), and statistical significance was set at p < 0.05. This study was approved by the Institutional Review Board (IRB No. 30-2025-38) and conducted using a retrospective review of medical records without involving identifiable personal information. This research received no external funding. Results GPT-4o underestimated IPSS total scores compared to patient-reported values. The mean patient-reported IPSS score was 13.6 ± 7.1, whereas the mean GPT-estimated score was 11.2 ± 6.1, reflecting a mean difference of approximately -2.4 points, which was statistically significant (p = 0.006). Contrastingly, OABSS total scores were comparable between the two methods (patient-reported: 7.0 ± 3.6; GPT-estimated: 6.9 ± 3.6), and the difference was not statistically significant (p = 0.686) (Table 1). Regarding diagnostic classification, 42 patients were classified as having moderate-to-severe LUTS (IPSS ≥ 8) based on patient-reported scores, compared to 38 patients based on GPT estimates. This difference was not statistically significant (McNemar’s test, p = 0.344). The sensitivity and specificity of the GPT-based LUTS classification were 83.3% and 70.0%, respectively. For OAB, 51 patients met the diagnostic criteria (OABSS ≥ 3 and urgency score ≥ 2) based on patient responses, whereas GPT-4o identified 50 of these patients as having OAB. This classification difference was not statistically significant (McNemar’s p = 1.000). The sensitivity and specificity of the GPT-based OAB classification were 88.2% and 80.8%, respectively. Correlation analysis showed moderate to strong agreement between GPT-estimated and patient-reported total scores for both questionnaires (IPSS: ρ = 0.60*; OABSS: ρ = 0.70*; both p < 0.001) (Figures 1). ROC curve analysis confirmed the diagnostic performance of the model. The AUC for LUTS prediction using GPT-estimated IPSS scores and OAB prediction using GPT-estimated OABSS scores was 0.81 and 0.91, respectively (Figures 2). Sub-item agreement The agreement between GPT-4o the predicted and patient-reported responses for each questionnaire item was evaluated using weighted Cohen’s kappa coefficients. The interpretation of kappa values was based on the criteria proposed by Landis and Koch (1977) (moderate agreement: 0.41–0.60; substantial agreement: 0.61–0.80; almost perfect agreement: >0.80) [12]. For the IPSS, the weighted kappa values across the eight items ranged from 0.23–0.81, indicating moderate agreement overall. Among these, the QoL item (Q8) showed the highest agreement (κ = 0.81), reflecting nearly perfect concordance in satisfaction-related responses. The urgency item (Q4) demonstrated relatively strong agreement (κ ≈ 0.42). Contrastingly, the intermittency (Q3) and nocturia (Q7) items exhibited lower agreement levels (κ ≈ 0.33 and 0.32, respectively), suggesting that GPT predictions frequently diverged from patient-reported responses for these symptoms. Straining to void (Q6) showed the lowest agreement among all IPSS items (κ ≈ 0.23), indicating poor concordance (Table 2). For OABSS, weighted kappa values ranged from 0.44–0.71 across the four items. The urgency urinary incontinence item (Q4) showed the highest agreement (κ = 0.71), suggesting that GPT-4o effectively captured this distinct symptom. Conversely, daytime frequency (Q1) showed relatively low agreement (κ = 0.44), whereas nocturia (Q2) and urgency (Q3) demonstrated moderate agreement (κ ≈ 0.49 and 0.59, respectively) (Table 3). The Bland–Altman plot revealed a mean difference of -2.37 between GPT-estimated and patient-reported IPSS total scores, indicating a slight underestimation bias. Most data points were within the 95% limits of agreement, suggesting acceptable agreement [13], although some clinically relevant discrepancies were observed. The Bland–Altman analysis for OABSS total scores showed a negligible mean bias –0.13 between GPT-estimated and patient-reported values. The distribution of differences was symmetrical around 0, and most points were within the 95% limit of agreement, indicating strong overall concordance without systemic over- or underestimation (Figure 3). For LUTS classification based on the IPSS threshold (≥ 8), GPT-4o correctly identified 35 out of 42 patients with LUTS and seven out of 10 without LUTS, yielding a sensitivity of 83.3% and specificity of 70.0% (Table 4). In the classification of OAB based on OABSS criteria (total score ≥ 3 and urgency score ≥ 2), 45 out of 51 OAB cases and 21 out of 26 non-OAB cases were correctly identified by GPT-4o, corresponding to a sensitivity of 88.2% and specificity of 80.8%. A subgroup analysis was performed to compare the model performance between patients at their first visit (n = 17) and the entire cohort. Among first-time visitors, the correlation and agreement between GPT-estimated and patient-reported scores were lower than those among follow-up patients. For IPSS, the Spearman’s correlation coefficient for total scores in the first-visit subgroup was 0.5, which was lower than the overall cohort average (ρ = 0.6*), suggesting that the limited clinical information and variability in symptom descriptions at the initial visit may reduce the predictive accuracy of the model. A similar trend was observed for OABSS (first-visit ρ = 0.5 vs. follow-up ρ = 0.7*), although the sample size was insufficient to demonstrate statistically significant interaction effects. Discussion This study demonstrated that a LLM, such as the GPT-4o, can moderately estimate standardized LUTS and OAB symptom scores based on patients’ free-text symptom descriptions and medical records. Without any additional training or fine-tuning, the GPT-4o model could interpret subjective symptom narratives and generate sub-items and total scores for the IPSS and OABSS, which showed significant correlations with actual patient-reported scores. Particularly, the OABSS results demonstrated minimal mean differences and high diagnostic accuracy (AUC = 0.91), suggesting that natural-language-based AI models may have clinical use in evaluating OAB symptoms. Although GPT-4o tended to underestimate IPSS scores compared to patient responses, an average difference of 2–3 points was not considered clinically significant. Sub-item-level agreement analysis revealed the specific strengths and limitations of the model. GPT-4o was relatively accurate in identifying prominent symptoms, such as reduced QoL and urgency urinary incontinence. For example, in the QoL item (Q8), if patients described their symptoms using expressions, such as “my life is very uncomfortable,” the GPT correspondingly assigned high QoL scores that aligned well with the actual responses. Similarly, the presence of urinary incontinence is often clearly reflected in the patient’s narrative and appropriately captured by the model. These findings suggest that symptoms that are more explicitly described or emotionally salient are more likely to be accurately detected using the GPT. Contrastingly, items, such as intermittency or nocturia, were often more difficult for the model to infer unless explicitly mentioned. In the case of intermittency, patients may not be fully aware of or fail to verbalize issues, such as interrupted urinary flow, which can contribute to lower agreement. For nocturia, unless the patients clearly stated the frequency they walked up at night to void, the GPT struggled to estimate the correct score. Prediction performance for the straining item was poor, because of the lack of direct expressions, such as “I need to push to urinate,” in the source text. Thus, the GPT-4o performed well for symptoms that were obvious or emotionally emphasized (urgency, incontinence, QoL impairment) but showed limitations for symptoms that required observational or quantitative details (stream interruption, nocturnal frequency) unless explicitly documented in the clinical narrative. The reduced predictive accuracy of GPT-4o in first-visit patients can be attributed to several factors. First, the GPT input was based on EMR documentation, which, in the case of follow-up visits, often included cumulative clinical information from previous encounters. This allowed the model to reference past data and better infer the patient’s symptom trajectory and baseline IPSS or OABSS. However, first-visit patients lacked historical records, forcing the model to rely solely on information from a single visit, which provided limited contextual cues and led to uncertain estimations. Second, symptom descriptions during initial visits were often more variable and ambiguous, making it more difficult for the model to map free-text inputs to structured questionnaire items. Clinically, physicians frequently require additional clarification in interpreting first-visit symptom narratives to assign appropriate IPSS or OABSS scores. During follow-up visits, patients describe their symptoms in a more structured and familiar manner and previous standardized documentation can aid the model by providing more consistent input patterns. Hence, these findings suggest that the GPT-4o may perform less reliably if previous contextual information is lacking. To ensure safe clinical application, especially for first-visit evaluations, additional caution or complementary mechanisms may be necessary to support the model performance. Recent studies suggest that LLMs can be effectively used to generate and respond to clinical questionnaires. In a study by Zou et al., users compared emotional responses and perceived acceptability of mental health questionnaires written by humans and those generated by ChatGPT. Interestingly, a hybrid format in which human-written questions were rephrased by ChatGPT elicited the most positive emotional responses from participants, outperforming the purely human- and AI-generated versions [ 14 ]. Other investigations focused on comparing the performances of different LLM platforms. In one study evaluating ChatGPT, Gemini, and Copilot for their ability to answer clinically relevant questions about onabotulinum toxin and sacral neuromodulation (SNM) for OAB, ChatGPT consistently outperformed the others across all evaluated domains, including factual accuracy, coherence, and clinical applicability [ 15 ]. The clinical significance and potential use of the LLM-based symptom score estimation are considerable. Its key advantage lies in its applicability to situations in which patients may have difficulty completing structured questionnaires, such as older adults with limited comprehension, individuals with lower levels of education, or patients facing language barriers. Accordingly, a natural language interface enables patients to describe their symptoms in their own words and the AI model can convert this narrative into standardized scores. This allows clinicians to obtain objective and reliable assessments without relying on formal self-reporting instruments. This approach could be highly valuable in telephone consultations or telemedicine settings in which symptom tracking and treatment response evaluation can be conducted without face-to-face interaction. Additionally, by analyzing the description of patients about each domain, the model may offer supplementary insights beyond numerical scores. For example, GPT-generated outputs can help clinicians identify symptoms that are emphasized or overlooked by patients, allowing for more tailored counseling and clinical decision-making. Thus, this methodology may serve as a scoring tool and the foundation for a broader diagnostic support system. In our study, the GPT-4o produced diagnostic classifications for LUTS and OAB that showed clinically acceptable differences compared to those based on patient-reported questionnaire scores. With further refinement and large-scale training, LLMs may evolve into robust clinical adjuncts for assessing LUT dysfunction. However, this study had several limitations. First, the study was conducted at a single institution with a relatively small sample size, which may have limited the generalizability of the findings. Particularly, the cohort included female patients with OAB and male patients with LUTS, and further validation in more homogeneous subgroups is warranted to assess the model performance more precisely. Second, GPT-4o is a large language model trained primarily on English-language data. Although the model demonstrated a reasonable performance using Korean clinical records, it may have limitations in interpreting subtle or nuanced expressions in Korean. Although most patient narratives in our dataset were relatively simple, more diverse or metaphorical languages may have been misinterpreted by the model. Third, the real-world application of such models will require addressing the issues of data privacy and cost. Using the OpenAI GPT-4 API involves transmitting patient data to external servers, raising concerns regarding the handling of sensitive medical information. Finally, although the number of patients classified as having LUTS or OAB did not differ significantly between the GPT and patient-reported scores, a subset of patients showed discordant classifications. These mismatches contributed to the relatively modest sensitivity and specificity observed in diagnostic performance analysis. Conclusion This study demonstrated that the GPT-4o can moderately predict IPSS and OABSS scores based solely on patients’ free-text symptom descriptions and EMR, including previous symptom scores. GPT-4o-based natural language symptom assessment has strong potential as a complementary or alternative to conventional questionnaires. In situations in which self-reporting is challenging or impractical, such as in the elderly or low-literacy populations, this approach may enhance clinical efficiency and patient accessibility by enabling reliable assessment through natural dialogue alone. Declarations Clinical trial number: Not applicable. Ethics approval and consent to participate This study was conducted in accordance with the Declaration of Helsinki for experiments involving humans. Ethics approval for this study was obtained from the Institutional Review Board (IRB) of Seoul National University Boramae Medical Center (IRB No. 30-2025-38). The requirement for informed consent was waived by the IRB as this was a retrospective study, posed no risk to the participants, and did not involve the exposure of any personally identifiable information. Consent to publication Not applicable. Availability of data and material The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Competing interests The authors declare that they have no competing interests. Funding No funding was received for this study. References Barry MJ, Fowler FJ, Jr., O'Leary MP, Bruskewitz RC, Holtgrewe HL, Mebust WK, et al. The American Urological Association symptom index for benign prostatic hyperplasia. The Measurement Committee of the American Urological Association. J Urol. 1992;148(5):1549-57; discussion 64. Abrams P, Cardozo L, Fall M, Griffiths D, Rosier P, Ulmsten U, et al. The standardisation of terminology in lower urinary tract function: report from the standardisation sub-committee of the International Continence Society. Urology. 2003;61(1):37-49. Jeong SJ, Homma Y, Oh SJ. Korean version of the overactive bladder symptom score questionnaire: translation and linguistic validation. Int Neurourol J. 2011;15(3):135-42. Kurita N, Yamazaki S, Fukumori N, Otoshi K, Otani K, Sekiguchi M, et al. Overactive bladder symptom severity is associated with falls in community-dwelling adults: LOHAS study. BMJ Open. 2013;3(5):e002413. Okinami T. [LIMITATIONS OF QUESTIONNAIRE-BASED EVALUATION OF LOWER URINARY TRACT DYSFUNCTION IN ELDERLY PATIENTS]. Nihon Hinyokika Gakkai Zasshi. 2021;112(4):185-91. K C. Should International Prostate Symptom Score Solely Guide the Management of Benign Prostatic Hyperplasia? J Urol Ren Dis 2017: 142 2017. Sanman KN, Shetty R, Adapala RR, Patil S, Prabhu GL, Venugopal P. Can new, improvised Visual Prostate Symptom Score replace the International Prostate Symptom Score? Indian perspective. Indian J Urol. 2020;36(2):123-9. Kung TH, Cheatham M, Medenilla A, Sillos C, De Leon L, Elepaño C, et al. Performance of ChatGPT on USMLE: Potential for AI-assisted medical education using large language models. PLOS Digital Health. 2023;2(2):e0000198. Liu J, Gu J, Tong M, Yue Y, Qiu Y, Zeng L, et al. Evaluating the agreement between ChatGPT-4 and validated questionnaires in screening for anxiety and depression in college students: a cross-sectional study. BMC Psychiatry. 2025;25(1):359. Wang Y, Huang Y, Nimma IR, Pang S, Pang M, Cui T, et al. Validation of GPT-4 for clinical event classification: A comparative analysis with ICD codes and human reviewers. J Gastroenterol Hepatol. 2024;39(8):1535-43. Evans JD. Straightforward statistics for the behavioral sciences. Pacific Grove, CA Brooks/Cole Pub Co United States. 1996. Landis JR, Koch GG. The Measurement of Observer Agreement for Categorical Data. Biometrics. 1977;33(1):159-74. Martin Bland J, Altman D. STATISTICAL METHODS FOR ASSESSING AGREEMENT BETWEEN TWO METHODS OF CLINICAL MEASUREMENT. The Lancet. 1986;327(8476):307-10. Zou Z, Mubin O, Alnajjar F, Ali L. A pilot study of measuring emotional response and perception of LLM-generated questionnaire and human-generated questionnaires. Scientific Reports. 2024;14(1):2781. Hacibey I, Halis A. Assessment of artificial intelligence performance in answering questions on onabotulinum toxin and sacral neuromodulation. Investig Clin Urol. 2025;66(3):188-93. Tables Table 1. Comparison of patient-reported and GPT-estimated IPSS and OABSS total and sub-item scores. Mean scores and standard deviations (mean ± SD) are shown for each domain. Statistical significance between GPT-estimated and patient-reported values was assessed using paired t-tests for continuous scores and McNemar’s test for diagnostic classification. GPT estimations showed significant differences in certain IPSS items such as Intermittency and Weak stream, while OABSS showed generally high agreement. Patient reported IPSS GPT estimated IPSS p Patient reported OABSS GPT estimated OABSS p Statistical test Age 71.13(52-86) 71.58(35-86) Total, n 91 IPSS - total, n 52 OABSS - total, n 77 IPSS - Male, n 52 OABSS - Male, n 41 IPSS - Female, n 0 OABSS - Female, n 36 LUTS Diagnosed, n 42 38 0.344 OAB Diagnosed, n 51 50 1 McNemar’s test IPSS, Sub-items scores OABSS, Sub-items scores Q1.FIE 1.62 ± 1.48 1.51 ± 1.3 0.624 Q1.Urine/d 0.81 ± 0.65 1.1 ± 0.74 0.0004 Paired t-test Q2.Urine in 2hrs. 2.23 ± 1.37 2.15 ± 1.33 0.699 Q2.Nocturia 2.12 ± 0.83 2.09 ± 0.99 0.804 Q3.Intermittency 1.96 ± 1.63 1.48 ± 1.23 0.041 Q3.Urgency 2.19 ± 1.57 1.96 ± 1.42 0.132 Q4.Urgency 1.73 ± 1.4 1.48 ± 1.39 0.233 Q4.UUI 1.87 ± 1.61 1.7 ± 1.61 0.232 Q5.Weak stream 2.5 ± 1.66 1.96 ± 1.37 0.019 Sum 6.99 ± 3.63 6.86 ± 3.58 0.686 Q6.Straining 1.38 ± 1.32 1.02 ± 1.09 0.084 Q7.Nocturia 2.17 ± 1.13 1.65 ± 0.9 0.002 Q8.QoL 3.02 ± 1.43 2.81 ± 1.21 0.062 Sum 13.6 ± 7.09 11.23 ± 6.1 0.006 Table 2. Weighted kappa values for agreement between GPT-estimated and patient-reported IPSS sub-item scores. Kappa coefficients were calculated for each of the 8 IPSS items to assess agreement in ordinal responses. The highest agreement was observed in the QoL item (κ = 0.807), while items such as Straining and Nocturia showed lower concordance. IPSS Weighted Kappa Q1.FIE 0.488 Q2.Urine in 2hrs. 0.442 Q3.Intermittency 0.329 Q4.Urgency 0.422 Q5.Weak stream 0.421 Q6.Straining 0.226 Q7.Nocturia 0.315 Q8.QoL 0.807 Table 3. Weighted kappa values for agreement between GPT-estimated and patient-reported OABSS sub-item scores. Kappa values indicate moderate to substantial agreement across all four OABSS items, with the highest for UUI (κ = 0.705) and the lowest for Urine frequency (κ = 0.438). OABSS Weighted Kappa Q1.Urine/d 0.438 Q2.Nocturia 0.494 Q3.Urgency 0.589 Q4.UUI 0.705 Interpretation of weighted kappa values follows Landis and Koch (1977): 0.41–0.60 = moderate, 0.61–0.80 = substantial, >0.80 = almost perfect agreement. Table 4. Confusion matrix comparing GPT-estimated LUTS classification with patient-reported IPSS-based diagnosis (IPSS ≥ 8). GPT correctly identified 35 of 42 LUTS cases and 7 of 10 non-LUTS cases, resulting in a sensitivity of 83.3% and specificity of 70.0%. Additional Declarations No competing interests reported. Supplementary Files 2025.08.05v1.4Supplementtable12.docx Cite Share Download PDF Status: Published Journal Publication published 19 Jan, 2026 Read the published version in BMC Urology → Version 1 posted Editorial decision: Revision requested 10 Nov, 2025 Reviews received at journal 08 Nov, 2025 Reviews received at journal 06 Nov, 2025 Reviewers agreed at journal 04 Nov, 2025 Reviewers agreed at journal 03 Nov, 2025 Reviewers agreed at journal 01 Nov, 2025 Reviews received at journal 28 Oct, 2025 Reviewers agreed at journal 22 Oct, 2025 Reviewers agreed at journal 20 Oct, 2025 Reviews received at journal 20 Oct, 2025 Reviewers agreed at journal 17 Oct, 2025 Reviewers agreed at journal 06 Oct, 2025 Reviewers invited by journal 13 Sep, 2025 Editor invited by journal 03 Sep, 2025 Editor assigned by journal 01 Sep, 2025 Submission checks completed at journal 01 Sep, 2025 First submitted to journal 27 Aug, 2025 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-7468583","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":518260409,"identity":"aa46c4f6-a783-4a02-a79a-49ad852fcc93","order_by":0,"name":"Hoyoung Bae","email":"","orcid":"","institution":"Seoul National University Boramae Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Hoyoung","middleName":"","lastName":"Bae","suffix":""},{"id":518260410,"identity":"7732cbf0-3b60-4a75-9629-d5bc05caecb2","order_by":1,"name":"Jung Hoon Lee","email":"","orcid":"","institution":"Seoul National University Boramae Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Jung","middleName":"Hoon","lastName":"Lee","suffix":""},{"id":518260411,"identity":"a18d7799-470f-4779-9959-18babc99594c","order_by":2,"name":"Min Chul Cho","email":"","orcid":"","institution":"Seoul National University Boramae Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Min","middleName":"Chul","lastName":"Cho","suffix":""},{"id":518260412,"identity":"f0b73106-2358-47b7-9783-f47a5fd7c724","order_by":3,"name":"Sangjun Yoo","email":"","orcid":"","institution":"Seoul National University Boramae Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Sangjun","middleName":"","lastName":"Yoo","suffix":""},{"id":518260413,"identity":"1dce79d6-0366-4fe3-b6b1-374297a508ff","order_by":4,"name":"Hyeon Jeong","email":"","orcid":"","institution":"Seoul National University Boramae Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Hyeon","middleName":"","lastName":"Jeong","suffix":""},{"id":518260414,"identity":"55591de9-fb30-4a78-813c-66d91c5c55ec","order_by":5,"name":"Hwancheol Son","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAwklEQVRIiWNgGAWjYFAC5sYDDAw2MF4CMVoYG4Ba0kjXcpgELQbHGxsOfGw7b28ukcD44QdDWj5hLWcONhyc2XY7ceeMBGbJHoYcywZCWiRnJDYc5m27nWBwI4FBmoGhwoCgLWAtf9vO2QO1MP8mSgu/BFALY9sBxg03EtiAtuQQoYUH6Jeec8mJG848bLPsMUgjrIWNvfnggx9ldvYGx5MP3/hRkUxYCxgwsoHJBmAAEqcBCP4QrXIUjIJRMApGIgAAooA/N7UUpi8AAAAASUVORK5CYII=","orcid":"","institution":"Seoul National University Boramae Medical Center","correspondingAuthor":true,"prefix":"","firstName":"Hwancheol","middleName":"","lastName":"Son","suffix":""}],"badges":[],"createdAt":"2025-08-27 06:53:33","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7468583/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7468583/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12894-026-02054-z","type":"published","date":"2026-01-19T15:57:29+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":91951740,"identity":"3a091ed9-df5b-4da4-a27b-e47eb5e7bc62","added_by":"auto","created_at":"2025-09-23 06:48:30","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":815249,"visible":true,"origin":"","legend":"","description":"","filename":"2025.08.30BMCManuscriptwithaffilation.docx","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/771516f367a851575fcd05a9.docx"},{"id":91951738,"identity":"8bed6bad-7f10-4063-8506-dcc7821f84f7","added_by":"auto","created_at":"2025-09-23 06:48:30","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7716,"visible":true,"origin":"","legend":"","description":"","filename":"6104e0048d8243bdacbdcc95287d64df.json","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/9a9a4c13586486a078209e38.json"},{"id":91953849,"identity":"31204241-738a-48da-92f8-e017655913f8","added_by":"auto","created_at":"2025-09-23 07:04:30","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":789494,"visible":true,"origin":"","legend":"","description":"","filename":"2025.08.05v1.4Supplementtable12.docx","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/2ab1e117697655c9363e88d2.docx"},{"id":91953847,"identity":"2c519e8a-1678-4432-b0d5-c3508edd5e5b","added_by":"auto","created_at":"2025-09-23 07:04:30","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":76407,"visible":true,"origin":"","legend":"","description":"","filename":"6104e0048d8243bdacbdcc95287d64df1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/29b0c94f719e77dcb5b047cf.xml"},{"id":91953846,"identity":"8dd23c41-89b3-4ef2-a8db-2f508688a78b","added_by":"auto","created_at":"2025-09-23 07:04:30","extension":"jpeg","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":595595,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/6f1e521489c9206986be8722.jpeg"},{"id":91953441,"identity":"0c0a923b-6bf4-4f40-90b0-cadc9521eb76","added_by":"auto","created_at":"2025-09-23 06:56:30","extension":"jpeg","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":341741,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/5abac754f3516efed769852b.jpeg"},{"id":91953450,"identity":"f02fa31f-0981-4c4e-8658-2f9f6a8c093e","added_by":"auto","created_at":"2025-09-23 06:56:31","extension":"jpeg","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":176644,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/36f16fffa3a0426298de1852.jpeg"},{"id":91953443,"identity":"7d68eb1a-8812-43ed-b856-d763caf2a49c","added_by":"auto","created_at":"2025-09-23 06:56:30","extension":"jpeg","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":155516,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/dd40cd7951fa4641c421cf49.jpeg"},{"id":91951746,"identity":"a3ffe4e6-8748-4b4c-bd07-b5dd37f220b7","added_by":"auto","created_at":"2025-09-23 06:48:30","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":82720,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/6035ad84fa8a7746573ae959.png"},{"id":91956891,"identity":"60b56a4d-4ec1-4714-a4d1-3b76f5aab996","added_by":"auto","created_at":"2025-09-23 07:20:30","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":89540,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/39cbb5d1160b7cfa205eba4e.png"},{"id":91956362,"identity":"404ba3c2-059c-4862-a7b3-ef4d53cbf6b9","added_by":"auto","created_at":"2025-09-23 07:12:30","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":113560,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/b10c64000a3a32fc4415f43f.png"},{"id":91953448,"identity":"492e2f2e-be9b-40fc-8f8b-1ade2594859b","added_by":"auto","created_at":"2025-09-23 06:56:30","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":66667,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/373ad14bb247eca7462c1638.png"},{"id":91953444,"identity":"0b098848-e5c0-4b98-be69-c432719f5f0a","added_by":"auto","created_at":"2025-09-23 06:56:30","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":39901,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/84192c8d95055487e1f867a1.png"},{"id":91953850,"identity":"93178060-5623-4b6e-b7a0-968736fde69b","added_by":"auto","created_at":"2025-09-23 07:04:31","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":33946,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/a5738072452414abac375d79.png"},{"id":91956364,"identity":"7cc61a6c-12e9-43d5-b95f-47ced0d84f6b","added_by":"auto","created_at":"2025-09-23 07:12:31","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":23102,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/6d18da4fcec716f4799d9566.png"},{"id":91951755,"identity":"971f369d-f054-4b8b-b1b6-4121cd2e26d6","added_by":"auto","created_at":"2025-09-23 06:48:31","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":24961,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/1b542f91ff54fcfd2d87f56e.png"},{"id":91951759,"identity":"35a32d3d-0a84-4f6d-9bdd-3c8ec5bc109b","added_by":"auto","created_at":"2025-09-23 06:48:31","extension":"xml","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":75444,"visible":true,"origin":"","legend":"","description":"","filename":"6104e0048d8243bdacbdcc95287d64df1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/7e0f87df18e01998b2f2abc2.xml"},{"id":91953451,"identity":"63732ba6-2f77-4dac-9f88-f8e92c0a0037","added_by":"auto","created_at":"2025-09-23 06:56:31","extension":"html","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":84588,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/ba5057150b8f4067b285b892.html"},{"id":91951741,"identity":"2e3f1fa0-08e5-40ce-9669-0dd4d9c9a93c","added_by":"auto","created_at":"2025-09-23 06:48:30","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":286629,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003ea) Spearman correlation heatmap between GPT-estimated IPSS sub-item scores and patient-reported scores.\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eStrongest correlations were observed in the QoL and total score domains (ρ = 0.83 and 0.60, respectively). Asterisks \u003c/em\u003eindicate p \u0026lt; 0.05.*\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u003cstrong\u003e(b) Spearman correlation heatmap between GPT-estimated OABSS sub-item scores and patient-reported scores.\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e\u003cbr\u003e\nStrongest correlations were observed in the urgency and total score domains (ρ = 0.71 and 0.70, respectively). Asterisks indicate p \u0026lt; 0.05.*\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eCorrelation interpretation: ρ = 0.40–0.59 = moderate, 0.60–0.79 = strong, ≥ 0.80 = very strong (Evans, 1996).\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/fb6c70ae42e4fae320c33c30.png"},{"id":91951737,"identity":"3cf5173a-9052-4810-a56b-c3380ed114c5","added_by":"auto","created_at":"2025-09-23 06:48:30","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":105704,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003e(a) Receiver operating characteristic (ROC) curve for LUTS prediction based on GPT-estimated IPSS total scores.\u003c/strong\u003e\u003cbr\u003e\nThe ROC curve evaluates the performance of GPT-4o in predicting moderate-to-severe LUTS (defined as IPSS ≥ 8) using total predicted IPSS scores.\u003cbr\u003e\nThe area under the curve (AUC) was 0.81, indicating fair-to-good discriminative ability.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(b) Receiver operating characteristic (ROC) curve for OAB prediction based on GPT-estimated OABSS total scores.\u003c/strong\u003e\u003cbr\u003e\nThe ROC curve assesses the discriminative ability of GPT-4o in predicting clinically defined OAB (OABSS ≥ 3 and urgency score ≥ 2).\u003cbr\u003e\nThe area under the curve (AUC) was 0.91, indicating excellent predictive performance.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/bfc60da38b4c6ba712816d76.png"},{"id":91951750,"identity":"cbe462b6-616c-434a-8a52-7264d34b07c0","added_by":"auto","created_at":"2025-09-23 06:48:30","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":171062,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003e(a) Bland–Altman plot comparing GPT-estimated and patient-reported total IPSS scores.\u003c/strong\u003e\u003cbr\u003e\nEach point represents the difference between GPT-estimated and actual IPSS total scores plotted against their mean.\u003cbr\u003e\nThe red dashed line indicates the mean difference (bias = –2.37), while the upper and lower dashed lines indicate the 95% limits of agreement (LoA).\u003cbr\u003e\nMost values fall within the agreement range, suggesting moderate consistency, though GPT tended to underestimate IPSS total scores on average.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e(b) Bland–Altman plot comparing GPT-estimated and patient-reported total OABSS scores.\u003c/strong\u003e\u003cbr\u003e\nEach point represents the difference between GPT-estimated and actual OABSS total scores plotted against their mean.\u003cbr\u003e\nThe mean difference (bias) was –0.13, suggesting minimal systematic bias.\u003cbr\u003e\nThe majority of data points fell within the 95% limits of agreement (LoA), indicating overall good agreement between GPT and patient-reported OABSS.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLoA\u003c/strong\u003e: Limits of Agreement (95% agreement interval)\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/06e67edf1c1a83da37fc70f5.png"},{"id":101152043,"identity":"34fb3483-4432-49fc-9d86-450d51d88081","added_by":"auto","created_at":"2026-01-26 16:09:16","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1449572,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/5c6db033-94d6-4890-ad10-bfa0a99ba008.pdf"},{"id":91951744,"identity":"199bf2bd-ec21-4efe-9775-0a89d86faabd","added_by":"auto","created_at":"2025-09-23 06:48:30","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":789494,"visible":true,"origin":"","legend":"","description":"","filename":"2025.08.05v1.4Supplementtable12.docx","url":"https://assets-eu.researchsquare.com/files/rs-7468583/v1/bc44b3bb17d6e3047eb4b3e5.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"ChatGPT-4o as an Assist Tool for LUTS Evaluation: Better Prediction of IPSS-QoL and OABSS Scores","fulltext":[{"header":"Introduction","content":"\u003cp\u003eStandardized questionnaires, such as the International Prostate Symptom Score (IPSS) and Overactive Bladder Symptom Score (OABSS), are widely used for evaluating lower urinary tract symptoms (LUTS)[1].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eHowever, in real-world clinical settings, a significant proportion of elderly patients experience difficulty completing these self-administered questionnaires accurately. Particularly, patients with limited education or cognitive impairment may\u0026nbsp;find the questionnaires complex or difficult to understand [5].\u0026nbsp;One study reported that approximately 70% of patients failed to complete the IPSS independently because of difficulties understanding some items [6].\u0026nbsp;These limitations highlight the challenged posed by traditional paper-based questionnaires, which can be influenced by recall bias and subjectivity.\u003c/p\u003e\n\u003cp\u003eRecently, the IPSS requires patients to recall their urinary frequency and nocturia episodes, which can distort the scoring owing to inaccurate memory or daily variability. Consequently, physicians frequently assist elderly patients in completing each questionnaire item verbally or rely on caregivers, which is time consuming and often reduces the consistency of their responses [5].\u0026nbsp;Therefore, there is growing interest in automating symptom scoring using natural language-based artificial intelligence (AI).\u003c/p\u003e\n\u003cp\u003eChatGPT, a large language model (LLM) developed by OpenAI, has demonstrated human-like performance in high-stakes medical examinations, such as the USMLE [8]. Particularly, several studies have shown that the GPT-4 model can effectively interpret free-form patient narratives and generate clinically meaningful outputs [8–10].\u0026nbsp;Recent studies have evaluated whether\u0026nbsp;ChatGPT can replace or augment traditional patient questionnaires. In one study, ChatGPT-4o was prompted to respond to structured mental health questionnaires (for depression and anxiety) as if it were the patient. The ChatGPT scores showed a high level of agreement with the validated questionnaire results. However, Bland–Altman analysis revealed a systematic bias and wide limits of agreement [9],\u0026nbsp;highlighting the need for rigorous quantitative validation before clinical\u0026nbsp;adoptation.\u003c/p\u003e\n\u003cp\u003eThis study used the OpenAI GPT-4o model to estimate IPSS and OABSS solely from patients’ free-form symptom descriptions and electronic medical records (EMRs) without using the questionnaires themselves. By comparing the predicted scores with the actual patient-reported scores, we aimed to assess the accuracy and limitations of this language-based AI model and explore its potential to replace or supplement traditional symptom questionnaires.\u0026nbsp;\u003cbr\u003e\u0026nbsp;This study aimed to validate the predictive accuracy of GPT-4o in estimating IPSS and OABSS and evaluate the degree of agreement for each questionnaire item in a clinical setting.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cp\u003eIn May 2025, a retrospective analysis was conducted on 91 patients who visited the urology outpatient clinic of a secondary care hospital for the evaluation of LUTS or overactive bladder (OAB). Among them, 52 patients completed the IPSS questionnaire and 77 completed the OABSS questionnaire (some patients completed both\u0026nbsp;questionnaires). The cohort included 55 men (60.4%) and 36 women (39.6%) with a mean age of approximately 71 years (range; 35–86). All patients in the IPSS group were men, whereas the OABSS group included 36 women (Table 1).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eData collection\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe actual self-reported IPSS and OABSS questionnaire scores were collected\u0026nbsp;for each patient. Additionally, free-text symptom descriptions provided by patients and EMRs from corresponding outpatient visits were used as inputs for the GPT-4o model. Symptom descriptions included sentences written by the attending urologist based on the patient’s spoken language during clinical interviews. The EMRs contained a chronologically organized documentation of the\u0026nbsp;patients’ chief complaints, relevant diagnostic findings, and clinical impressions, all of which were used without further modifications. These unstructured text data were entered into the GPT-4o (OpenAI GPT-4 model, May 2025 version), which was used to estimate patients’ responses to eight IPSS items and four OABSS items. The model assigned scores ranging from 0 to each item’s defined maximum (0–5 depending on the item) and generated total scores and diagnostic classifications (LUTS based on IPSS ≥ 8; OAB based on OABSS ≥ 3 with urgency score ≥ 2). All the predictions were made using a zero-shot approach without\u0026nbsp;task-specific fine-tuning.\u003c/p\u003e\n\u003cp\u003eTo assess the agreement between the GPT-predicted and patient-reported total scores, Bland–Altman analysis was performed besides correlation analysis.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eStatistical analysis\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eStatistical comparisons between patient-reported and GPT-estimated sub-item scores and total scores for IPSS and OABSS were performed using paired t-tests. Spearman’s correlation coefficients were calculated to assess the relationship between the two scoring methods and interpreted according to conventional benchmarks proposed by Evans (1996), in which ρ = 0.40–0.59 indicates moderate, 0.60–0.79 indicates strong, and ≥ 0.80 indicates very strong correlation [11]. Agreement for individual items was further evaluated using weighted Cohen’s kappa coefficients, interpreted according to the criteria described by Landis and Koch (1977), in which κ = 0.41–0.60 indicates moderate agreement, 0.61–0.80 substantial agreement, and \u0026gt; 0.80 almost perfect agreement [12]. For diagnostic classification, LUTS was defined as an IPSS total score ≥ 8, and OAB was defined as an OABSS total score ≥ 3 with an urgency item (Q3) score ≥ 2. Discrepancies between the GPT-predicted and patient-reported diagnostic categories were assessed using\u0026nbsp;McNemar’s test. Overall diagnostic performance was evaluated using receiver operating characteristic (ROC) curve analysis, and the area under the curve (AUC) was calculated. Moreover, Bland–Altman analysis examined the degree of numerical agreement between the GPT-estimated and patient-reported total scores and identified any systematic bias or clinically relevant discrepancies. All statistical analyses were conducted using IBM SPSS Statistics version 29 (IBM Corp., Armonk, NY, USA), and statistical significance was set at p \u0026lt; 0.05.\u003c/p\u003e\n\u003cp\u003eThis study was approved by the Institutional Review Board (IRB No. 30-2025-38) and conducted using a retrospective review of medical records without involving identifiable personal information.\u003c/p\u003e\n\u003cp\u003e\u003cbr\u003e\u0026nbsp;This research received no external funding.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eGPT-4o underestimated IPSS total scores compared to patient-reported values. The mean patient-reported IPSS score was 13.6 \u0026plusmn; 7.1, whereas the mean GPT-estimated score was 11.2 \u0026plusmn; 6.1, reflecting a mean difference of approximately -2.4 points, which was statistically significant (p = 0.006). Contrastingly, OABSS total scores were comparable between the two methods (patient-reported: 7.0 \u0026plusmn; 3.6; GPT-estimated: 6.9 \u0026plusmn; 3.6), and the difference was not statistically significant (p = 0.686) (Table 1). Regarding diagnostic classification, 42 patients were classified as having moderate-to-severe LUTS (IPSS \u0026ge; 8) based on patient-reported scores, compared to 38 patients based on GPT estimates. This difference was not statistically significant (McNemar\u0026rsquo;s test, p = 0.344). The sensitivity and specificity of the GPT-based LUTS classification were 83.3% and 70.0%, respectively. For OAB, 51 patients met the diagnostic criteria (OABSS \u0026ge; 3 and urgency score \u0026ge; 2) based on patient responses, whereas GPT-4o identified 50 of these patients as having OAB. This classification difference was not statistically significant (McNemar\u0026rsquo;s p = 1.000). The sensitivity and specificity of the GPT-based OAB classification were 88.2% and 80.8%, respectively.\u003c/p\u003e\n\u003cp\u003eCorrelation analysis showed moderate to strong agreement between GPT-estimated and patient-reported total scores for both questionnaires (IPSS: \u0026rho; = 0.60*; OABSS: \u0026rho; = 0.70*; both p \u0026lt; 0.001) (Figures 1).\u003c/p\u003e\n\u003cp\u003eROC curve analysis confirmed the diagnostic performance\u0026nbsp;of the model.\u003c/p\u003e\n\u003cp\u003eThe AUC for LUTS prediction using GPT-estimated IPSS scores and OAB prediction using GPT-estimated OABSS scores was 0.81 and 0.91, respectively (Figures 2).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eSub-item agreement\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe agreement between GPT-4o the predicted and patient-reported responses for each questionnaire item was evaluated using weighted Cohen\u0026rsquo;s kappa coefficients. The interpretation of kappa values was based on the criteria proposed by Landis and Koch (1977) (moderate agreement: 0.41\u0026ndash;0.60; substantial agreement: 0.61\u0026ndash;0.80; almost perfect agreement: \u0026gt;0.80) [12]. For the IPSS, the weighted kappa values across the eight items ranged from 0.23\u0026ndash;0.81, indicating moderate agreement overall. Among these, the QoL item (Q8) showed the highest agreement (\u0026kappa; = 0.81), reflecting nearly perfect concordance in satisfaction-related responses. The urgency item (Q4) demonstrated relatively strong agreement (\u0026kappa; \u0026asymp; 0.42). Contrastingly, the intermittency (Q3) and nocturia (Q7) items exhibited lower agreement levels (\u0026kappa; \u0026asymp; 0.33 and 0.32, respectively), suggesting that GPT predictions frequently diverged from patient-reported responses for these symptoms. Straining to void (Q6) showed the lowest agreement among all IPSS items (\u0026kappa; \u0026asymp; 0.23), indicating poor concordance (Table 2). For OABSS, weighted kappa values ranged from 0.44\u0026ndash;0.71 across the four items. The urgency urinary incontinence item (Q4) showed the highest agreement (\u0026kappa; = 0.71), suggesting that GPT-4o effectively captured this distinct symptom. Conversely, daytime frequency (Q1) showed relatively low agreement (\u0026kappa; = 0.44), whereas nocturia (Q2) and urgency (Q3) demonstrated moderate agreement (\u0026kappa; \u0026asymp; 0.49 and 0.59, respectively) (Table 3).\u003c/p\u003e\n\u003cp\u003eThe Bland\u0026ndash;Altman plot revealed a mean difference of -2.37 between GPT-estimated and patient-reported IPSS total scores, indicating a slight underestimation bias. Most data points were within the 95% limits of agreement, suggesting acceptable agreement [13], although some clinically relevant discrepancies were observed.\u003c/p\u003e\n\u003cp\u003eThe Bland\u0026ndash;Altman analysis for OABSS total scores showed a negligible mean bias \u0026ndash;0.13 between GPT-estimated and patient-reported values. The distribution of differences was symmetrical around 0, and most points were within the 95% limit of agreement, indicating strong overall concordance without systemic over- or underestimation (Figure 3).\u003c/p\u003e\n\u003cp\u003eFor LUTS classification based on the IPSS threshold (\u0026ge; 8), GPT-4o correctly identified 35 out of 42 patients with LUTS and seven out of 10 without LUTS, yielding a sensitivity of 83.3% and specificity of 70.0% (Table 4).\u003c/p\u003e\n\u003cp\u003eIn the classification of OAB based on OABSS criteria (total score \u0026ge; 3 and urgency score \u0026ge; 2), 45 out of 51 OAB cases and 21 out of 26 non-OAB cases were correctly identified by GPT-4o, corresponding to a sensitivity of 88.2% and specificity of 80.8%.\u003c/p\u003e\n\u003cp\u003eA subgroup analysis was performed to compare the model performance between patients at their first visit (n = 17) and the entire cohort. Among first-time visitors, the correlation and agreement between GPT-estimated and patient-reported scores were lower than those among follow-up patients. For IPSS, the Spearman\u0026rsquo;s correlation coefficient for total scores in the first-visit subgroup was 0.5, which was lower than the overall cohort average (\u0026rho; = 0.6*), suggesting that the limited clinical information and variability in symptom descriptions at the initial visit may reduce the predictive accuracy of the model. A similar trend was observed for OABSS (first-visit \u0026rho; = 0.5 vs. follow-up \u0026rho; = 0.7*), although the sample size was insufficient to demonstrate statistically significant interaction effects.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study demonstrated that a LLM, such as the GPT-4o, can moderately estimate standardized LUTS and OAB symptom scores based on patients\u0026rsquo; free-text symptom descriptions and medical records. Without any additional training or fine-tuning, the GPT-4o model could interpret subjective symptom narratives and generate sub-items and total scores for the IPSS and OABSS, which showed significant correlations with actual patient-reported scores. Particularly, the OABSS results demonstrated minimal mean differences and high diagnostic accuracy (AUC\u0026thinsp;=\u0026thinsp;0.91), suggesting that natural-language-based AI models may have clinical use in evaluating OAB symptoms. Although GPT-4o tended to underestimate IPSS scores compared to patient responses, an average difference of 2\u0026ndash;3 points was not considered clinically significant.\u003c/p\u003e\u003cp\u003eSub-item-level agreement analysis revealed the specific strengths and limitations of the model. GPT-4o was relatively accurate in identifying prominent symptoms, such as reduced QoL and urgency urinary incontinence. For example, in the QoL item (Q8), if patients described their symptoms using expressions, such as \u0026ldquo;my life is very uncomfortable,\u0026rdquo; the GPT correspondingly assigned high QoL scores that aligned well with the actual responses. Similarly, the presence of urinary incontinence is often clearly reflected in the patient\u0026rsquo;s narrative and appropriately captured by the model. These findings suggest that symptoms that are more explicitly described or emotionally salient are more likely to be accurately detected using the GPT. Contrastingly, items, such as intermittency or nocturia, were often more difficult for the model to infer unless explicitly mentioned. In the case of intermittency, patients may not be fully aware of or fail to verbalize issues, such as interrupted urinary flow, which can contribute to lower agreement. For nocturia, unless the patients clearly stated the frequency they walked up at night to void, the GPT struggled to estimate the correct score. Prediction performance for the straining item was poor, because of the lack of direct expressions, such as \u0026ldquo;I need to push to urinate,\u0026rdquo; in the source text. Thus, the GPT-4o performed well for symptoms that were obvious or emotionally emphasized (urgency, incontinence, QoL impairment) but showed limitations for symptoms that required observational or quantitative details (stream interruption, nocturnal frequency) unless explicitly documented in the clinical narrative.\u003c/p\u003e\u003cp\u003eThe reduced predictive accuracy of GPT-4o in first-visit patients can be attributed to several factors. First, the GPT input was based on EMR documentation, which, in the case of follow-up visits, often included cumulative clinical information from previous encounters. This allowed the model to reference past data and better infer the patient\u0026rsquo;s symptom trajectory and baseline IPSS or OABSS. However, first-visit patients lacked historical records, forcing the model to rely solely on information from a single visit, which provided limited contextual cues and led to uncertain estimations. Second, symptom descriptions during initial visits were often more variable and ambiguous, making it more difficult for the model to map free-text inputs to structured questionnaire items. Clinically, physicians frequently require additional clarification in interpreting first-visit symptom narratives to assign appropriate IPSS or OABSS scores. During follow-up visits, patients describe their symptoms in a more structured and familiar manner and previous standardized documentation can aid the model by providing more consistent input patterns. Hence, these findings suggest that the GPT-4o may perform less reliably if previous contextual information is lacking. To ensure safe clinical application, especially for first-visit evaluations, additional caution or complementary mechanisms may be necessary to support the model performance.\u003c/p\u003e\u003cp\u003eRecent studies suggest that LLMs can be effectively used to generate and respond to clinical questionnaires. In a study by Zou et al., users compared emotional responses and perceived acceptability of mental health questionnaires written by humans and those generated by ChatGPT. Interestingly, a hybrid format in which human-written questions were rephrased by ChatGPT elicited the most positive emotional responses from participants, outperforming the purely human- and AI-generated versions [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eOther investigations focused on comparing the performances of different LLM platforms. In one study evaluating ChatGPT, Gemini, and Copilot for their ability to answer clinically relevant questions about onabotulinum toxin and sacral neuromodulation (SNM) for OAB, ChatGPT consistently outperformed the others across all evaluated domains, including factual accuracy, coherence, and clinical applicability [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eThe clinical significance and potential use of the LLM-based symptom score estimation are considerable.\u003c/p\u003e\u003cp\u003eIts key advantage lies in its applicability to situations in which patients may have difficulty completing structured questionnaires, such as older adults with limited comprehension, individuals with lower levels of education, or patients facing language barriers. Accordingly, a natural language interface enables patients to describe their symptoms in their own words and the AI model can convert this narrative into standardized scores. This allows clinicians to obtain objective and reliable assessments without relying on formal self-reporting instruments. This approach could be highly valuable in telephone consultations or telemedicine settings in which symptom tracking and treatment response evaluation can be conducted without face-to-face interaction. Additionally, by analyzing the description of patients about each domain, the model may offer supplementary insights beyond numerical scores. For example, GPT-generated outputs can help clinicians identify symptoms that are emphasized or overlooked by patients, allowing for more tailored counseling and clinical decision-making.\u003c/p\u003e\u003cp\u003eThus, this methodology may serve as a scoring tool and the foundation for a broader diagnostic support system. In our study, the GPT-4o produced diagnostic classifications for LUTS and OAB that showed clinically acceptable differences compared to those based on patient-reported questionnaire scores. With further refinement and large-scale training, LLMs may evolve into robust clinical adjuncts for assessing LUT dysfunction.\u003c/p\u003e\u003cp\u003eHowever, this study had several limitations. First, the study was conducted at a single institution with a relatively small sample size, which may have limited the generalizability of the findings. Particularly, the cohort included female patients with OAB and male patients with LUTS, and further validation in more homogeneous subgroups is warranted to assess the model performance more precisely. Second, GPT-4o is a large language model trained primarily on English-language data. Although the model demonstrated a reasonable performance using Korean clinical records, it may have limitations in interpreting subtle or nuanced expressions in Korean.\u003c/p\u003e\u003cp\u003eAlthough most patient narratives in our dataset were relatively simple, more diverse or metaphorical languages may have been misinterpreted by the model. Third, the real-world application of such models will require addressing the issues of data privacy and cost. Using the OpenAI GPT-4 API involves transmitting patient data to external servers, raising concerns regarding the handling of sensitive medical information. Finally, although the number of patients classified as having LUTS or OAB did not differ significantly between the GPT and patient-reported scores, a subset of patients showed discordant classifications. These mismatches contributed to the relatively modest sensitivity and specificity observed in diagnostic performance analysis.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study demonstrated that the GPT-4o can moderately predict IPSS and OABSS scores based solely on patients\u0026rsquo; free-text symptom descriptions and EMR, including previous symptom scores. GPT-4o-based natural language symptom assessment has strong potential as a complementary or alternative to conventional questionnaires. In situations in which self-reporting is challenging or impractical, such as in the elderly or low-literacy populations, this approach may enhance clinical efficiency and patient accessibility by enabling reliable assessment through natural dialogue alone.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eClinical trial number:\u0026nbsp;\u003c/strong\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was conducted in accordance with the Declaration of Helsinki for experiments involving humans. Ethics approval for this study was obtained from the Institutional Review Board (IRB) of Seoul National University Boramae Medical Center (IRB No. 30-2025-38). The requirement for informed consent was waived by the IRB as this was a retrospective study, posed no risk to the participants, and did not involve the exposure of any personally identifiable information.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNo funding was received for this study.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eBarry MJ, Fowler FJ, Jr., O\u0026apos;Leary MP, Bruskewitz RC, Holtgrewe HL, Mebust WK, et al. The American Urological Association symptom index for benign prostatic hyperplasia. The Measurement Committee of the American Urological Association. J Urol. 1992;148(5):1549-57; discussion 64.\u003c/li\u003e\n \u003cli\u003eAbrams P, Cardozo L, Fall M, Griffiths D, Rosier P, Ulmsten U, et al. The standardisation of terminology in lower urinary tract function: report from the standardisation sub-committee of the International Continence Society. Urology. 2003;61(1):37-49.\u003c/li\u003e\n \u003cli\u003eJeong SJ, Homma Y, Oh SJ. Korean version of the overactive bladder symptom score questionnaire: translation and linguistic validation. Int Neurourol J. 2011;15(3):135-42.\u003c/li\u003e\n \u003cli\u003eKurita N, Yamazaki S, Fukumori N, Otoshi K, Otani K, Sekiguchi M, et al. Overactive bladder symptom severity is associated with falls in community-dwelling adults: LOHAS study. BMJ Open. 2013;3(5):e002413.\u003c/li\u003e\n \u003cli\u003eOkinami T. [LIMITATIONS OF QUESTIONNAIRE-BASED EVALUATION OF LOWER URINARY TRACT DYSFUNCTION IN ELDERLY PATIENTS]. Nihon Hinyokika Gakkai Zasshi. 2021;112(4):185-91.\u003c/li\u003e\n \u003cli\u003eK C. Should International Prostate Symptom Score Solely Guide the Management of Benign Prostatic Hyperplasia? J Urol Ren Dis 2017: 142 2017.\u003c/li\u003e\n \u003cli\u003eSanman KN, Shetty R, Adapala RR, Patil S, Prabhu GL, Venugopal P. Can new, improvised Visual Prostate Symptom Score replace the International Prostate Symptom Score? Indian perspective. Indian J Urol. 2020;36(2):123-9.\u003c/li\u003e\n \u003cli\u003eKung TH, Cheatham M, Medenilla A, Sillos C, De Leon L, Elepa\u0026ntilde;o C, et al. Performance of ChatGPT on USMLE: Potential for AI-assisted medical education using large language models. PLOS Digital Health. 2023;2(2):e0000198.\u003c/li\u003e\n \u003cli\u003eLiu J, Gu J, Tong M, Yue Y, Qiu Y, Zeng L, et al. Evaluating the agreement between ChatGPT-4 and validated questionnaires in screening for anxiety and depression in college students: a cross-sectional study. BMC Psychiatry. 2025;25(1):359.\u003c/li\u003e\n \u003cli\u003eWang Y, Huang Y, Nimma IR, Pang S, Pang M, Cui T, et al. Validation of GPT-4 for clinical event classification: A comparative analysis with ICD codes and human reviewers. J Gastroenterol Hepatol. 2024;39(8):1535-43.\u003c/li\u003e\n \u003cli\u003eEvans JD. Straightforward statistics for the behavioral sciences. Pacific Grove, CA Brooks/Cole Pub Co United States. 1996.\u003c/li\u003e\n \u003cli\u003eLandis JR, Koch GG. The Measurement of Observer Agreement for Categorical Data. Biometrics. 1977;33(1):159-74.\u003c/li\u003e\n \u003cli\u003eMartin Bland J, Altman D. STATISTICAL METHODS FOR ASSESSING AGREEMENT BETWEEN TWO METHODS OF CLINICAL MEASUREMENT. The Lancet. 1986;327(8476):307-10.\u003c/li\u003e\n \u003cli\u003eZou Z, Mubin O, Alnajjar F, Ali L. A pilot study of measuring emotional response and perception of LLM-generated questionnaire and human-generated questionnaires. Scientific Reports. 2024;14(1):2781.\u003c/li\u003e\n \u003cli\u003eHacibey I, Halis A. Assessment of artificial intelligence performance in answering questions on onabotulinum toxin and sacral neuromodulation. Investig Clin Urol. 2025;66(3):188-93.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable 1. Comparison of patient-reported and GPT-estimated IPSS and OABSS total and sub-item scores.\u003c/strong\u003e\u003cbr\u003e\u0026nbsp;Mean scores and standard deviations (mean \u0026plusmn; SD) are shown for each domain. Statistical significance between GPT-estimated and patient-reported values was assessed using paired t-tests for continuous scores and McNemar\u0026rsquo;s test for diagnostic classification. GPT estimations showed significant differences in certain IPSS items such as Intermittency and Weak stream, while OABSS showed generally high agreement.\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"687\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003ePatient reported\u0026nbsp;\u003cbr\u003e\u0026nbsp;IPSS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003eGPT estimated\u0026nbsp;\u003cbr\u003e\u0026nbsp;IPSS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003ep\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 70px;\"\u003e\n \u003cp\u003ePatient reported\u0026nbsp;\u003cbr\u003e\u0026nbsp;OABSS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 63px;\"\u003e\n \u003cp\u003eGPT estimated\u0026nbsp;\u003cbr\u003e\u0026nbsp;OABSS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003ep\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003eStatistical test\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 144px;\"\u003e\n \u003cp\u003e71.13(52-86)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 11px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 133px;\"\u003e\n \u003cp\u003e71.58(35-86)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003eTotal, n\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"8\" style=\"width: 499px;\"\u003e\n \u003cp\u003e91\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003eIPSS - total, n\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 144px;\"\u003e\n \u003cp\u003e52\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 11px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003eOABSS - total, n\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 133px;\"\u003e\n \u003cp\u003e77\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003eIPSS - Male, n\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 144px;\"\u003e\n \u003cp\u003e52\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 11px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003eOABSS - Male, n\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 133px;\"\u003e\n \u003cp\u003e41\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003eIPSS - Female, n\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 144px;\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 11px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003eOABSS - Female, n\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 133px;\"\u003e\n \u003cp\u003e36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003eLUTS Diagnosed, n\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e42\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e0.344\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003eOAB Diagnosed, n\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 70px;\"\u003e\n \u003cp\u003e51\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 63px;\"\u003e\n \u003cp\u003e50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003eMcNemar\u0026rsquo;s test\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"4\" style=\"width: 302px;\"\u003e\n \u003cp\u003eIPSS, Sub-items scores\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"5\" style=\"width: 374px;\"\u003e\n \u003cp\u003eOABSS, Sub-items scores\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003eQ1.FIE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e1.62 \u0026plusmn; 1.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e1.51 \u0026plusmn; 1.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e0.624\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003eQ1.Urine/d\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 70px;\"\u003e\n \u003cp\u003e0.81 \u0026plusmn; 0.65\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 63px;\"\u003e\n \u003cp\u003e1.1 \u0026plusmn; 0.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e0.0004\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"9\" style=\"width: 86px;\"\u003e\n \u003cp\u003ePaired t-test\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003eQ2.Urine in 2hrs.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e2.23 \u0026plusmn; 1.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e2.15 \u0026plusmn; 1.33\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e0.699\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003eQ2.Nocturia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 70px;\"\u003e\n \u003cp\u003e2.12 \u0026plusmn; 0.83\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 63px;\"\u003e\n \u003cp\u003e2.09 \u0026plusmn; 0.99\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e0.804\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003eQ3.Intermittency\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e1.96 \u0026plusmn; 1.63\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e1.48 \u0026plusmn; 1.23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e0.041\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003eQ3.Urgency\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 70px;\"\u003e\n \u003cp\u003e2.19 \u0026plusmn; 1.57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 63px;\"\u003e\n \u003cp\u003e1.96 \u0026plusmn; 1.42\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e0.132\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003eQ4.Urgency\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e1.73 \u0026plusmn; 1.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e1.48 \u0026plusmn; 1.39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e0.233\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003eQ4.UUI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 70px;\"\u003e\n \u003cp\u003e1.87 \u0026plusmn; 1.61\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 63px;\"\u003e\n \u003cp\u003e1.7 \u0026plusmn; 1.61\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e0.232\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003eQ5.Weak stream\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e2.5 \u0026plusmn; 1.66\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e1.96 \u0026plusmn; 1.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e0.019\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003eSum\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 70px;\"\u003e\n \u003cp\u003e6.99 \u0026plusmn; 3.63\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 63px;\"\u003e\n \u003cp\u003e6.86 \u0026plusmn; 3.58\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e0.686\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003eQ6.Straining\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e1.38 \u0026plusmn; 1.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e1.02 \u0026plusmn; 1.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e0.084\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 70px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 63px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003eQ7.Nocturia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e2.17 \u0026plusmn; 1.13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e1.65 \u0026plusmn; 0.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 70px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 63px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003eQ8.QoL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e3.02 \u0026plusmn; 1.43\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e2.81 \u0026plusmn; 1.21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e0.062\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 70px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 63px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 102px;\"\u003e\n \u003cp\u003eSum\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 74px;\"\u003e\n \u003cp\u003e13.6 \u0026plusmn; 7.09\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 71px;\"\u003e\n \u003cp\u003e11.23 \u0026plusmn; 6.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e0.006\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 11px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 100px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 70px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 63px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 55px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2. Weighted kappa values for agreement between GPT-estimated and patient-reported IPSS sub-item scores.\u003c/strong\u003e\u003cbr\u003e\u0026nbsp;Kappa coefficients were calculated for each of the 8 IPSS items to assess agreement in ordinal responses.\u003cbr\u003e\u0026nbsp;The highest agreement was observed in the QoL item (\u0026kappa; = 0.807), while items such as Straining and Nocturia showed lower concordance.\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"312\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 147px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 165px;\"\u003e\n \u003cp\u003eIPSS Weighted Kappa\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 147px;\"\u003e\n \u003cp\u003eQ1.FIE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 165px;\"\u003e\n \u003cp\u003e0.488\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 147px;\"\u003e\n \u003cp\u003eQ2.Urine in 2hrs.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 165px;\"\u003e\n \u003cp\u003e0.442\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 147px;\"\u003e\n \u003cp\u003eQ3.Intermittency\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 165px;\"\u003e\n \u003cp\u003e0.329\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 147px;\"\u003e\n \u003cp\u003eQ4.Urgency\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 165px;\"\u003e\n \u003cp\u003e0.422\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 147px;\"\u003e\n \u003cp\u003eQ5.Weak stream\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 165px;\"\u003e\n \u003cp\u003e0.421\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 147px;\"\u003e\n \u003cp\u003eQ6.Straining\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 165px;\"\u003e\n \u003cp\u003e0.226\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 147px;\"\u003e\n \u003cp\u003eQ7.Nocturia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 165px;\"\u003e\n \u003cp\u003e0.315\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 147px;\"\u003e\n \u003cp\u003eQ8.QoL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 165px;\"\u003e\n \u003cp\u003e0.807\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3. Weighted kappa values for agreement between GPT-estimated and patient-reported OABSS sub-item scores.\u003c/strong\u003e\u003cbr\u003e\u0026nbsp;Kappa values indicate moderate to substantial agreement across all four OABSS items, with the highest for UUI (\u0026kappa; = 0.705) and the lowest for Urine frequency (\u0026kappa; = 0.438).\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"299\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 119px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003eOABSS Weighted Kappa\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 119px;\"\u003e\n \u003cp\u003eQ1.Urine/d\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e0.438\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 119px;\"\u003e\n \u003cp\u003eQ2.Nocturia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e0.494\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 119px;\"\u003e\n \u003cp\u003eQ3.Urgency\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e0.589\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 119px;\"\u003e\n \u003cp\u003eQ4.UUI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 180px;\"\u003e\n \u003cp\u003e0.705\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cem\u003eInterpretation of weighted kappa values follows Landis and Koch (1977): 0.41\u0026ndash;0.60 = moderate, 0.61\u0026ndash;0.80 = substantial, \u0026gt;0.80 = almost perfect agreement.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 4. Confusion matrix comparing GPT-estimated LUTS classification with patient-reported IPSS-based diagnosis (IPSS \u0026ge; 8).\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGPT correctly identified 35 of 42 LUTS cases and 7 of 10 non-LUTS cases, resulting in a sensitivity of 83.3% and specificity of 70.0%.\u003c/p\u003e\n\u003cp\u003e\u003cimg src=\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAhsAAACOCAYAAACYE8pkAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsMAAA7DAcdvqGQAACHJSURBVHhe7d1/cNP1/QfwZzfnGILo/EELzQeE4kedV8ZCt9PGmz8S4MzV46qOpbtt2sQpuW1yOQKd4sym7guEy+G2Q4EE53YmdqLjxsUrNP7YkdrNGpk9N/fRUuXzKVJQ5hR1zDH5/rF+snzen6RNS5Im7fNx97km78+P/KDH59n3+/15fapOnTp1CkRERERF8hmxgYiIiKiQGDaIiIioqBg2iIiIqKgYNoiIiKioGDaIiIioqBg2iIiIqKgYNoiIiKioqvKts1FVVSU2ERERUQXJ85RfcOzZICIioqIadc9GIBDA008/jYaGhvRy2WWXiZvTJFBVVQVFUcRmIiIqM7IsA+PYszHqsHHy5En88Y9/RE9PT3o5duyYIXw0NDSgpqZGPARNMAwbRESVoeLCRrbNjxw5YggfPT09mDFjhimATJkyRdyVKhjDBhFRZZgQYSObv/3tb6YAsmjRIkP4uPzyy8XdqIIwbBARVYYJGzZEJ0+eNIWPI0eOmHo/Zs+eLe5KZYphg4ioMkyasJHNO++8YwogZ511limATJ06VdyVygDDBhFRZZjUYSOb119/3RRA6uvr0dDQgMWLF6OhoQH19fXibjQOGDaIiCoDw8YIPv30U1P4OHToEL761a8aAojFYhF3pSJj2CAiqgwMG2Nw7NgxUwCZMmWKYehl8eLFmDZtmrgrFRDDRuVKpVJoaWlBIpFgUCeaBBg2CqSvr88UQC677DJDAFm4cKG4G52GsYQNt9uNZDJpaMs84eknQZG+Ta71OpfLhUAgIDYXTTweh8/nq7iTNsMG0eQy3mFjwpQrr6urg8vlQigUwr59+3DixAls3boVCxcuRFdXF771rW9h+vTpuO6669DW1oYnn3wSqqqKh6EiSaVSkGUZzc3NUBTFsLS2toqbI5FIpNeHQiHY7XZomgar1WrYFwBCoVD6eSmDBhER5WfChI1sGhoa4PV68cgjj+DVV1+FqqpYu3Ytpk+fjt/85je44oorMGfOHNx0003YsGEDnn32WRw/flw8DJ0m/a/oaDQKp9MprkZnZ6fYZOB0OmGz2bBnzx5xVUnF43E4HA5Dm9PphKIoJekdCIfDcLvdYjMRUdmb0GFDdO6552LJkiW4++67sWvXLhw6dAjPPfccbr75Zrzzzjv4yU9+gpkzZ+LLX/4ybrvtNmzbtg379+8XD0OjtGXLFvj9flitVnFV3iwWCwYGBsRmIiKqAJMqbGQzb948rFixAps2bcIf/vAHfPzxx4hEIrBarfjTn/6E7373u5g2bRquueYarFmzBk888QTeeust8TCUg6ZpSCaTWLp0qbhqVDRNQ21trdg8Jg6HA7Isp5dMmqYZ1qVSqfQ+Pp8Pqqoa9tOHh3T6c/E4mqaZ2uLxeHo/DPWcZK7PHBKSZRnBYBDJZBKyLBt6WMTjZuv9CAQChm2IiEpp0oeNbKxWK+644w5EIhH09vbi0KFDuPvuu3HuueciFovhqquugsViQXNzM/7v//4PiUQC77//vngYAnD06FFgqGdirAKBAJLJJDwej7hq1BwOB1asWJGe4+H3+w0nX7vdbpgDsnv3bmBoqCcUCkGSJMN8kVxaW1vT27lcLrS2tsJutxvmofh8vvT2mqahp6cnvT6RSCAWi6UDif5ebTYbFEVJDz1pmga73Y5oNGp4X5mBw+12o6urK71eUZRhJ9kSERUaw0YeZsyYAbvdjh/96Ed46qmnoGka9u3bB5fLhffeew8PPPAAZs+ejfr6erjdbjz88MPpv4gpO7fbbfhLW9M0w3q73W5YN9LJPR/hcBiSJBlCi8fjgSRJSKVS6feQWTRurBNOd+zYkX7sdruhqipCoVC6TZ+7ov+eWCwWw2tZLBbYbDYcPnw43ZZNIBAwDVHp4QxDx08mk4b3AwDRaNTwnIiomBg2xmju3Lm4+eabsXHjRjz33HP48MMP8eijj+JrX/saUqkU3G43pk6diq9//etYvXo12tvb0d/fLx5mwrvwwguBob/AM0UiESiKkvOkl3k1SiQSEVePycDAQHoYInNRVRWDg4OwWCxwuVyw2+1ZhyJGI1tPTnV1tdiEwcFBw/PM95VMJkecp6KqKoLBoGE/u90ODH3ng4ODkCQp6/shIioVho0CWrRoEb73ve9h+/bt+POf/4wjR47g3nvvxQUXXIAnnngC11xzDWbPno3ly5fjgQcewN69e/Hee++Jh5lQLBYLJElCb2+vuGpcuFwuw3CCvug9DYFAAIqipOdmjLVnY7TC4TBkWTaELJvNJm6WVeawT+bCgEFE5YJho4imT5+Oa6+9FmvXrsXOnTtx8OBBdHd349vf/jaOHz+ODRs2QJIkXH755bj11luxZcsW9PT0iIepeKtWrTLMTxgvtbW16OrqEpuz6uzsTM+bEHtliqG9vR2hUMgQEPKpAyNJ0rC/M9XV1VmPI/aoEBEVE8NGiUmShBtvvBHr16/HM888g+PHj+Oxxx5DY2MjXnnlFdx+++2YMmUKrrrqKvh8PsRiMfT19YmHqShOpxMulwtyxtUdulKe9DweD1RVNfVW6EMmmqYhHA6n28XJrblO3IUghoZwOGx6rZqaGlOb1+s1TCTF0OfQP6PVaoUkSYZhIU3TyiL8EdHkwbBRBhYuXAiPx4OtW7fi5Zdfxrvvvov77rsPNTU1+N3vfgeHw4GamhrccMMNuO+++9DR0YFjx46JhylrgUAA0WgULS0thvkFpS71rSgKYrGY4T3oJ2a9loferpfz1lmtVthstvT6QopEIob3BcA0jKIP9cgZl75arVZEo1H4fL70vq2trYZA1dnZabhkt7W1NedcGSKiYpgw90aZ6PRLI1966SX09PTgxRdfxOzZsw33fmloaMBnPlO6/DiWe6MQEVHp6X/EjNc5nGGjgvX29hoCSG9vryl8XHzxxeJuBcOwQURUGRg2qGA+/vhj051vP/roI1MAueCCC8Rdx4Rhg4ioMjBsUFEdOnTIFEBmzpxpCiBnnHGGuOuIGDaIiCoDwwaV3KuvvmoIH/v37zeFj0suuUTczYRhg4ioMjBs0Lg7ceKEqffj/fffNwWQmTNnGvZj2CAiqgwMG1SWDh8+bAog5513niF86DcFIyKi8lZxYYOIiIgqU56n/IIbddjgX7Kkk2UZ9957Lx577DEsWLAAdXV1pp9EEwWHDUtLlmV+3wVUcT0b/McnnSzL+PTTT/H666/jjTfeQF9fH/r6+tKP33rrLVP40H/OmzdPPBxRWWPYKC2GjcJi2KCKJcvysL+4n3zyiSmA6I/ffvttUwDRf86ZM0c8FNG4Y9goLYaNwmLYoIo1UtgYzokTJ0wBRH/8zjvvZA0hCxYsQG1trXgoopJg2Cgtho3CYtiginU6YWM4H330kSmA6I/ff/991NXVmUJIXV0dZs2aJR6KqGAYNkqLYaOwGDaoYhUrbAzngw8+yBlEPv74Y1MA0X+KNUKIRotho7QYNgqLYYMq1niEjeH84x//SAcPMYycPHkyawhZsGABzj//fPFQRCYMG6XFsFFYDBtUscotbAzn2LFjpgCiP66qqso5NPPFL35RPBRNUgwbpcWwUVgMG1SxKilsDOfo0aOmAKI/PvPMM00BRP85Y8YM8VA0gTFslBbDRmExbFDFmihhYziDg4M5h2amTZuWNYQsWLAA06ZNEw9FFS7fsBEOhxEMBuFyuRAIBMTVCAQC0DQNkUhEXDVqmqbBbrcjGo3CarWKqxEOh9Hd3Z31tVKpFFpaWqAoCuLxOHw+n7hJmt/vh8fjATI+ny7X5zxd+YaNcDiM9vZ2dHZ2iquAoePk+n7cbjeuuOIKeDweuN1uJJNJcZO0RCIBi8UCZJy4s60rV+MdNj4jNhDR/1RXV+Oqq67Crbfeip/97Gf47W9/i5dffhnHjx/HSy+9hPvuuw9XXnklPvjgAzz++OO45ZZbMHPmTNTW1uLqq6/Gbbfdhg0bNuCpp55Cb28v/vnPf4ovQRNULBZDKpUSm8uS0+mEoihQFAWJRAIAEI1G02160AgEAmhvb0+3K4qC2traivmcw4lEIunP5Pf7IUmS4XNaLBZomgZZlhEKhQzrihG2JhqGDaIxmj17Nq6++mp4PB6sX78eTz75JF555RV89NFHeOGFF/DjH/8Yixcvxrvvvotf//rXaGlpwbnnnos5c+bguuuuw+23345NmzZh165d+Mtf/oJ//etf4ktQhbLZbACAtrY2cVXF0jQNsVgMO3bsMLR7PJ6svQYTUSQSgcvlgtPpNLXT8Bg2iIpAkiRce+21uP322xEMBrFr1y68+uqrOHHiBJ5//nm0tbVh4cKFePvtt7Fjxw7cdNNNmDZtGi666CIsWbIEXq8XoVAIu3fvxmuvvYaTJ0+KL0FlLhQKQVVVhMNhcVVFO3r0qNg0qWiaJjZRHhg2iErsoosugsPhSAeK3//+93jttdfw73//G4lEAj6fD5dddhkOHjyIhx9+GMuXL8eZZ56Juro6LFu2DD/4wQ/w4IMP4umnn8brr78+bmOwlJuqqnA6nXC5XAgGg3mdoGRZNizxeFzcZFxZLBbYbLYJ1VszWk1NTUgmk2X3b1MJGDaIysj8+fOxbNkyfP/738eDDz6IeDwORVHwn//8B08//TR++MMfoq6uDn19ffj5z3+O66+/Hp/97GchyzKcTifuvPNO/PKXv0RHRwcOHDggHp5KTB/LH25MP9s8gGg0Cp/PV3YnNX24QJblCddjkw+r1YpQKASfz2eaJErDY9ggqgBVVVW4+OKLcf311+POO+/EL37xC3R0dKCvrw+ffPIJdu3ahTvuuANz5szBX//6V4RCIdjtdnzuc5/DpZdeihtuuAE+nw9btmxBZ2cn3nzzTfElqEhCodCwfw0HAgHTPACr1Qq/34/Nmzcbti0HnZ2dCIVCCAaDkGV5QkwOHQ19Mq0kSZBlGW63W9yEsmDYIKpwZ5xxBi699FI0NTWlA8XevXvx5ptv4sMPP8TOnTvR2tqKWbNm4ZVXXsH69etx9dVXY8qUKbj88suxfPly+P1+bN26Fc8++yxUVRVfgk6D0+mEzWbLeWmpqqpoaGgQm7Fo0aKy/bfQT7h+vx8tLS2TLnBgKHQpioJkMsnAkQeGDaIJ7POf/zy+9KUvYfny5Vi9ejW2bt2KZ555BgcPHsR7772HaDSK73znOzj//PPx0ksv4ac//SmuvPJKnHXWWVi4cCFuvPFGtLW1IRwO4/nnn8ehQ4fEl6A8ZA6nlPLOxTU1NTkDy+DgICRJEptHxePxwOVyYffu3eKqsiFJEgYHB8VmYCjo1dTUiM2jkkgkkEwm85qXM5kxbBBNUl/4whdQX1+P5uZmrF27Ftu3b8fzzz+PgYEBHDlyBL/61a/wzW9+E2effTZeeOEF3HPPPVi8eDGmT5+Or3zlK/jGN76Bu+66C4888gj27duX8z90+u/kSr/fj1gshu7ubsM6SZLQ09NjaAOA/fv3n3YYqK6uhqqqWU+EPT09aGxsFJsnnFzfr6ZpUFUV9fX14ioqAoYNIjKZNm0aFi1ahJtvvhl33XUXduzYgX379uHw4cMYGBjA9u3b0dzcjKlTp6Yv5a2vr8c555yDxYsXw+Vy4Z577sGjjz6KF154YdJfLomhXgBJkkxVKr1eL2KxmGFOh6ZpCAaDWL9+vWHb0bJarbDZbLDb7Yb2eDyOWCw2qu5/TdNM2+u1N5qamgzt5UT/fsUJra2trXC5XKOq/BkIBExDRpFIBDabbVTHmYxYrpzGbDKUK6fR+fvf/24o7Z5Z4v3UqVOmG97pj8877zzxUGVnNOXKc5XP1suL22w2QyEovT1TPiWws+2Hob/mM18/EAggFosZtsn1WYYrgZ6trHk+73MsRlOuPLN8ui6zxHq27ylzvWi4f0OHw2EYmipWufZCG+9y5QwbNGYMGzQa7777rimE6D/POOMMUwDRf55zzjniocZFvmGDCiPfsEH5YdigisWwQYVy5MiRrCHkjTfewNSpU7P2htTV1eHss88WD1U0DBulxbBRWAwbVLEYNqgU3n77bVMA0YdmZsyYkTOInHXWWeKhTgvDRmkxbBQWwwZVLIYNGm8DAwOmEKL/vOCCC0wBRH88ZcoU8VAjYtgoLYaNwppwYSPbZB1xstJYORwONDY2VsRknMmAYYPK2cGDB7OGkDfeeAOzZs0yBRD98ZlnnikeCmDYKDmGjcKakGFDnMUbCATQ1dV12oGDYaO8MGxQperv7zcFEP3n3LlzTQFkwYIFuOSSS0b8/48Kh2GjsCZF2MDQB812KVUuDocDq1atMtwvoFjyuVkSmem/vEREVBnyPOUXXMnCxmjDw2i3Px0MG2PDvzxoMmFPXmlVVVXx+y4g/Rw+Xt9pySqIqqqK6urq9PNAIABZltOLXj0vHo9DlmWoqpq+ja9e+c3tdpuqwIXD4azHAYBUKpX+6ztzG/0Y+q2dY7EYYrEYZOEOfrmOS0RERPkrSdhwOByw2WzpIZR4PI6GhgYoigJFURAKheDz+aBpmuH2vaFQCIqijFjlTT9OIpGAz+czBQOHw2F4rWAwiFQqBYvFAkVR4HK54HK5oChKuqqfw+GA3+9P75ettj4RERGNrChhQ1VVQ6/AihUrDKV5nU6nYXhEfzya+yfo9w7YsWNHuk2/2ZEYDDK3cTqdkCQJ+/fvN2wjUlUVixYtSj/nEAsREdHYFCVsSJKU7hHI1TOhD3HoC4ZueZwvPZjY7XbDcYLBoOkOh2Ld/nzupOj3+9HS0gKHwyGuIiIiolEoStgYidvtRltbmyGQjFXmMfQlsxdlrDweD5Sh4RxxLgcRERHlr+RhQ9M0JJNJw9CG2BORjwsvvBAY6iEppkgkAkVRkEwmi/5aREREE1HJw4Y+pNHb25tuyzYfQpIkHD58WGxOs1gscLlcaGlpMbSHw+FRh4La2lpT4Ml8T/o6PeAQERFR/koeNgAgGo2mL2uVZRler1fcBF6vF8FgEHLGpaqiQCAAl8tlmvuRb+EwncfjQTKZNAyX1NbWpo9pt9sRjUZNcz+IiIhoZAUv6kWTh8yiXjSJyCzqVVIs6lVYk6aoF01eeuG1bMNlGOqhKtQEXL1QW66htHA4nPO1MovA6cXlci2ZvW1iYblcnzMbfd9iy/xs2eR6/263O+f3RWbJZBJVVVU5l/7+fnEXKoC6ujrD90zlh2GDSiYWi+UMAeVGLy6nDBWLw9Dwn3g5dyAQMBSWUxQFtbW1eX3OVCqFYDBY1N4hPWSIc5syybKcLqCnKApisVg6TOlXdomF8ig7m82GU6dOmZaVK1di5cqVmDdvnrgLnaaNGzdi79696e96w4YNDBxliGGDSsJmswEA2traxFUVS9M0xGIxw5VVGJoDlM+8oS1btiAUConNBdXS0oJQKJTzdcLhMGw2m6HIXigUQnt7e/q51+vF5s2b089pdPr7+/HQQw9h9erV4ioqgDVr1hhC3Jo1a4ChXiYqHwwbVDKhUAiqquac8FupRlP5VqdfAl7sGw0qijLsa3R3d6O5udnQVl9fD1VV01dh6cEpn94aMtu0aRN7NUpIH6qaNWuWuIrGEcMGlYSqqnA6nXC5XFmrvGYjzpMot658i8UCm802pt6aPXv2wOVyGdoCgQACgYBpvkg+39VYJZNJww0SkePy9MbGxhFL/JMZezVKz+v1MtyVIYYNKil98uFwkyj1SZ6Z8wj0y6XLLXDocxrkYS7RzmZgYAC1tbViM7q6utDT05P+3DabDa2trYZtHA6HKYjpS6Emc4ol/RsaGtDd3W1oo5Ht3LmTJ74SyJwgum7dOmzZskXchMYZwwaVXCgUQjKZzBkc9Popmd3/VqsVfr+/LOcOdHZ2pu8mLA9zJUwmTdNQU1MjNkOSJEMQ83q9hiENDL1e5oTUzKUQpfqpcLZt2zbs5FwqjL6+vvQE0fvvv58TRMsQwwaVnNPphM1mg8/nE1cBQ0MuDQ0NYjMWLVoEVVXF5rKgX72i38Avn8CRjVg4bjyr1orDKzQ6+gRFfXI0lUZHRweWLl2atVgkjR+GDRoXmcMp2YYTiqWmpiZnYBkcHDQNH4yWx+OBy+XC7t27xVUGFotl2HL8wynkMIokSaa7LWuaBlVVxzXoTATRaBRLliwRm6kEOGxVfhg2aFxYLBb4/X7EYjHTXABJktDT02NoA4D9+/efdhiorq42DUvoenp60NjYKDYXRW1tLQYGBsTmvBRyGKWxsdH0Xff29kKSJEMvS09Pj6nXhYa3d+9eDqGMk/7+fsydO1dspnHEsEHjxuPxQJIk0/XwXq8XsVjMMKdD0zQEg0GsX7/esO1oWa1W2Gw22O12Q3s8HkcsFhtVz4Cmaabt9dobTU1NhnbRokWL0NXVJTaXnNvtNn3XPp8Pq1atMmzX1dU14mei/+nv78eBAwc4hFICy5YtMzzfuHEj9uzZk663QeWBYYPGlVgQC0OBIJFIGG7WZ7fbkUgk8iqWhaFiVpnDCw6HI70uEomYbuDn8/mgKMqo/nq3WCxobm42HCff91kutSssFovpxoihUMgwOVd/jyN9JvqfF198EfPnzxebqQjWrVtnKFW+bdu2cbv/B+XGG7HRmMm8EdtpicfjeOqpp0Y99FFqbrcbzc3NwxYHmwxk3oitpKp4I7aC0s/h4/WdsmfjNMhlWGiKKod+8haHYsqJ/t4me9AgotPDno3TkK3LeTJhzwZNJuzZKC32bBQWezaIiIhoQptUYUO/3TYRERGVzqQKG0RERFR6BZ+zoVeGrK2tRTAYhMvlSre53W5DTYVEIpG+1DAej2Pz5s3o7Ow09D5k7q8Lh8MIBoPp55IkobOzM/08lUqhpaUFiUQiXU/BZrOZ6jlkvr7D4TBUlhQ/ZzweN5TXjkajaGlpmfRzNvL89SGqeJxDUFr8vgtrQs7Z0IsVKYpiCBoWiyVd6TAUCsFut5sqOeqTDvUlFosZwkYgEEB7e7thG0mSDHUUMrfVt4lEIohGo8DQ+1Iyaio4HA6sWLEi3e73+w2BJxwOw+fzIZFIpLfhXQX/K/P6di5cJvLC3/fSLvy+C7uMt6L0bHR1dRl6GjJ7LTJlXr+v9xxk9jbo++oFlzJ7LMTiS5lXhujbRaNRQyEivT3zM4TDYXR3d5tqHTgcDqxfvx5WqzXnVSe52icLXo1Ckwl78kqrij0bBaWfw8frOy1Kz4Z4f4nDhw9DVVXIwk2jksmk4WZU4v0YAKC+vh4YKgOt3yhL3AZDwyTija3yqXg4MDCAZDJpem+qqmJwcDDd86K/DyIiIhqdooSNbGw2m2HoQ188Ho+4acm5XC7T+1IUZdL2WBARERVSScJGTU2NaXJmNtlu/Z15B8rh7tiZTCZRU1MjNo+otrZ22Bti6b0oR48eNbRnew9ERERkVpKw4XQ6IUmSqSyz+Fxs0zTNcAdK/Y6dra2tGXv8d56IzWYbsSfiwgsvBISg4PF4oKqq6YqXzPfhcrlMt4oW3wMRERFlV5KwAQCdnZ2meRter9ewjSRJ8Hq96fV2u900ATMSiaCxsdFwHL19JBaLBS6XC3a7HbIsp0OHMnTVS+YxxStgxLuEZrtbKREREZkV/GqUscp1xQqVL5lXo9AkIvNqlJLi1SiFNSGvRiEiIiLSMWwQERFRUTFsEBERUVGVTdhwOp2crzFBhcPhrOXkdbIsI5VKic3A0FVB4XA4/Thzkq64ZF5lNNy6kTgcDtPVScXidrtN300gEDC898z1mqZBlmXE43HDPlQeksmkqUx05tLf3y/uQgVQV1c3bGnuxx9/3PRvkW07Kp6yCRtEI4lEIumCa36/H5IkGYqwWSyW9Mk4FAoZ1uUbHgKBABobG/Pe/nSkUqms9We6uroM7z3z3j8WiwWJRMJwU0AqHzabDadOnTItK1euxMqVKzFv3jxxFzpNGzduxN69e9Pf9YYNG0xBQlVVLF261PTvQqXDsEETSiQSgcvlMtVcyefSaE3TTDf+K6a2tjbYbDax2dTDF4lEoKpquvdHv4Rb7/Gh8tbf34+HHnoIq1evFldRAaxZs8YQ4tasWQMM9TJlYtAbXwwbNOGMZsgk0549e+ByucTmogiHw5AkCVdccYW4Ki9NTU1ob28Xm6kMbdq0ib0aJaQPVc2aNSvd9tZbb2VsQeOBYYMmlKamJiSTyTHNaWhvb0dTU5OhTR6aT5I5XyRb5dvR0DQNwWAw7x4UPTzpFXAxVE03V+l+Kh/s1Sg9r9ebNdzNnTvX8JxKi2GDJhSr1YpQKASfzwd5qLpsvlRVNZzQdW1tbfB6vVAUBYlEAslk0jCEEY/HTRNSMxdx8mtraytCoVDWuxdnE4lEYLPZTNvbbDb09vYa2qi87Ny5M+uJjworc4LounXrsGXLFsP6/v5+rF27Nr1NXV2dYT0VH8MGTThOpzM9sTLfngi9h0A8oQPAqlWrYLVa0+tdLhe6u7vT6/XXy7Xo+2JoAqokSaY5Jbm43W7EYrG85pxQ+dm2bZvpvkpUeH19felJn/fff79pgmhHR4dhYmhdXR0DR4kxbNCE1dnZCUVRkEwm8wocuVRXVxue19bWGp7nKx6P5x0c9KtqUMRbBFBx6RMUs00CpuLp6OjA0qVLTffeytTR0YEDBw7g8ccfF1dRkTBs0LiTJAmDg4NiMzA0tFFTUyM2j4o+9DHc/Aa9R2O4bXLJdxhl8+bNgFADJBgMpm9QqM8zSaVSsNvt8Pv9eQUTKk/RaBRLliwRm6kE8hm2mj9/vthERcSwQeNOkiT09PSIzdA0Daqqor6+XlxVFJIk4ejRo2LziPIdRtF7WjKXzHoh+tBKS0sLQqEQPB6P8EpGyWTS1OtC5WPv3r0cQhkn/f39I04IPXDgwJh7KWn0GDZo3Hm9XsRiMVPdiNbWVrhcrqzzKHIJBAKmCZm5JliKGhsbsX//frG5pPT3PtKcjlQqBUmSDPNBqHz09/fjwIEDHEIpgWXLlhmeb9y4EXv27EnX28i2TV1dHebPn89/nxJi2KCS0IcKMhc9XFitViQSCQSDQcP6FStW5H15qC4QCKCtrc1wHORZ1Kscalfow0nidyXLsuG72L17N1asWJGxJ5WTF198kd30JbJu3TpDCfJt27aZqoPecssthm2WLFmCvr4+wzZUXFWnxH+VHPTZvZysRjpZlifc74Pb7UZzc/OIPQvjSdM02O32CffdlztZlk0nMSqeqqoqft8FpJ/Dx+s7Zc8GUYZAIACfzzemomCloAeNUCgkriIiKlvs2aAxm4g9G0S5sGejtNizUVjs2SAiIqIJjWGDiIiIiophg4iIiIqKczZozDiGTZMJ5xCUFr/vwhrvORujDht5bk6TAP8zoMmEv++lxe+7sMb7HD7qsEFERESVKc9TfsFxzgYREREVVd49G0RERERjwZ4NIiIiKiqGDSIiIioqhg0iIiIqKoYNIiIiKiqGDSIiIioqhg0iIiIqKoYNIiIiKiqGDSIiIioqhg0iIiIqKoYNIiIiKiqGDSIiIioqhg0iIiIqqv8HVQHt4W6KpLoAAAAASUVORK5CYII=\"\u003e\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-urology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"buro","sideBox":"Learn more about [BMC Urology](http://bmcurol.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/buro/default.aspx","title":"BMC Urology","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Artificial intelligence, Chat GPT, IPSS, Large language model, Low urinary tract symptoms, OABSS, Overactive bladder. ","lastPublishedDoi":"10.21203/rs.3.rs-7468583/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7468583/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003ePurpose\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo evaluate the performance of GPT-4o in estimating International Prostate Symptom Score (IPSS) and Overactive Bladder Symptom Score (OABSS) based on patients’ natural language descriptions and full outpatient records, compared to actual questionnaire scores.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMaterials and Methods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study included 91 patients, of whom 52 completed IPSS and 77 completed OABSS. GPT-4o was prompted with verbatim symptom statements and full medical records written by a urologist. Predicted scores were compared to actual scores using paired t-tests, weighted Cohen’s kappa for item-level agreement, Spearman’s correlation for total scores, and Bland–Altman plots for bias. Diagnostic classifications (lower urinary tract symptoms [LUTS]: IPSS ≥8; overactive bladder [OAB]: OABSS ≥3 with urgency ≥2) were assessed using McNemar’s test and receiver operating characteristic curve analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMean IPSS scores were 13.6 (GPT) and 11.2 (actual) (p = 0.006), while OABSS scores were not significantly different (6.99 vs. 6.86, p = 0.686). Diagnostic agreement was high: LUTS in 42 (actual) vs. 38 (GPT) patients, and OAB in 51 vs. 50 patients. Area under curve was 0.86 for IPSS and 0.91 for OABSS. Kappa values ranged from 0.23–0.81 (IPSS) and 0.44–0.71 (OABSS), with highest concordance in quality of life (QoL) and urgency incontinence. Spearman’s correlation coefficient was 0.60 (IPSS) and 0.70 (OABSS). Accuracy was lower in first-visit patients.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGPT-4o estimated IPSS and OABSS with clinically acceptable accuracy. Its performance was comparable regarding diagnostic classification, particularly for QoL and OABSS. GPT-4o may complement traditional questionnaires, particularly with missing or incomplete patient-reported data.\u003c/p\u003e","manuscriptTitle":"ChatGPT-4o as an Assist Tool for LUTS Evaluation: Better Prediction of IPSS-QoL and OABSS Scores","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-23 06:48:26","doi":"10.21203/rs.3.rs-7468583/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-11-10T20:24:16+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-08T20:28:07+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-06T11:32:38+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"227186507069282222141097182902687936735","date":"2025-11-04T11:18:32+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"23655656687462965984007213442223069141","date":"2025-11-03T15:55:19+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"242683359851775100459940958996300318195","date":"2025-11-01T07:51:43+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-28T12:47:46+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"157443008092678705793777621604656945035","date":"2025-10-22T06:31:06+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"137166401868731936285229983702919692272","date":"2025-10-21T02:58:51+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-20T12:37:51+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"115045585782523992934416982132505385952","date":"2025-10-18T02:33:18+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"67773418881349965051900974943530315496","date":"2025-10-06T12:43:13+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-13T10:27:13+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-09-03T05:26:15+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-01T12:09:35+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-09-01T12:08:46+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Urology","date":"2025-08-27T06:51:49+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-urology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"buro","sideBox":"Learn more about [BMC Urology](http://bmcurol.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/buro/default.aspx","title":"BMC Urology","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"48b251a2-5ac1-48ac-9926-ba5a7662fc7b","owner":[],"postedDate":"September 23rd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2026-01-26T16:06:12+00:00","versionOfRecord":{"articleIdentity":"rs-7468583","link":"https://doi.org/10.1186/s12894-026-02054-z","journal":{"identity":"bmc-urology","isVorOnly":false,"title":"BMC Urology"},"publishedOn":"2026-01-19 15:57:29","publishedOnDateReadable":"January 19th, 2026"},"versionCreatedAt":"2025-09-23 06:48:26","video":"","vorDoi":"10.1186/s12894-026-02054-z","vorDoiUrl":"https://doi.org/10.1186/s12894-026-02054-z","workflowStages":[]},"version":"v1","identity":"rs-7468583","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7468583","identity":"rs-7468583","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}