Heterogeneity Among Patients Discontinuing Medical Nutrition Therapy in Primary Care: A Latent Class Analysis

Heterogeneity Among Patients Discontinuing Medical Nutrition Therapy in Primary Care: A Latent Class Analysis | 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 Heterogeneity Among Patients Discontinuing Medical Nutrition Therapy in Primary Care: A Latent Class Analysis Machiko Ukai, Kazuki Ohashi, Kazuhiro Abe, Soki Maeda, Katsuhiko Ogasawara This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8998867/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 5 You are reading this latest preprint version Abstract Background Medical nutrition therapy (MNT) is a key component of chronic disease management in primary care and is effective in managing noncommunicable diseases (NCDs), obesity, and malnutrition; however, discontinuation is common in routine clinical practice. While previous studies have examined individual factors associated with discontinuation, MNT disengagement likely reflects heterogeneous background patterns shaped by the interplay of multiple factors. From a primary care perspective emphasising comprehensiveness and continuity, this study aimed to identify heterogeneity among patients who discontinue MNT. Methods This retrospective observational study was conducted at a primary care clinic in Japan. Outpatients who received dietitian-led MNT and subsequently discontinued therapy between January 2019 and December 2024 were included (n = 332). Latent class analysis (LCA) using maximum likelihood estimation was conducted on categorical variables representing demographic characteristics, clinical factors, lifestyle behaviours, and psychosocial factors. The optimal model was selected based on information criteria, classification accuracy, and clinical interpretability. Results The final analytical sample comprised 300 patients. LCA identified four distinct subgroups with differing background patterns. These subgroups broadly reflected patterns involving older adults with relatively preserved self-care; working-age individuals with low clinical and lifestyle risk; individuals with high BMI and coexisting medical conditions receiving pharmacotherapy; and individuals exhibiting prominent lifestyle risk behaviours and low illness awareness. Conclusions Discontinuation of MNT was not attributable to a single determinant but reflected diverse, multifactorial background patterns encompassing demographic, clinical, behavioural, and psychosocial characteristics. These findings provide a foundation for developing tailored strategies to support sustained engagement with MNT in primary care, grounded in comprehensive patient profiling. Medical Nutrition Therapy Registered Dietitian Primary Care Latent Class Analysis Continuity Dropout Patient Background Figures Figure 1 Figure 2 Figure 3 Background Medical nutrition therapy (MNT) has demonstrated efficacy in enhancing health outcomes across a wide spectrum of conditions, including noncommunicable diseases (NCDs), obesity, and malnutrition [ 1 , 2 ]. In recent years, the global escalation of the double burden of malnutrition—defined by the simultaneous presence of undernutrition and overnutrition—has further emphasised the critical role of MNT in addressing complex nutritional challenges [ 3 , 4 ]. This dual burden underscores the pressing need for effective and adaptable nutritional interventions that respond to the evolving demands of population health [ 2 , 5 – 7 ]. In primary care settings, where the delivery of continuous and comprehensive care for community-dwelling individuals is a foundational principle, the integration of MNT into routine clinical practice is increasingly regarded as a pivotal strategy to support behavioural change and improve disease trajectories [ 8 – 11 ]. Sustained patient engagement is essential to fully realise the potential benefits of MNT; however, in real-world clinical contexts, a substantial proportion of patients discontinue MNT prematurely. Ongoing support within the framework of chronic disease management has been associated with improved adherence, better quality of life, and reductions in complications, hospital admissions, and healthcare costs [ 12 , 13 ]. Consequently, premature discontinuation of MNT presents a significant clinical challenge, as it limits the effectiveness of nutritional interventions and impedes optimal health outcomes [ 14 ]. Previous studies on the discontinuation of MNT have predominantly focused on hospital-based specialist outpatient clinics or structured weight-loss intervention programmes targeting obesity treatment and weight management [ 15 – 18 ]. These studies have reported associations between discontinuation and several individual factors, including younger age, lower baseline adiposity or poorer early weight loss response, employment—particularly full-time work, fewer obesity-related comorbidities, and psychological characteristics such as higher levels of psychological distress or hostility. Associations with BMI and travel distance to clinics, however, have been inconsistent across studies. In contrast, evidence concerning MNT delivered by registered dietitians within routine primary care remains scarce. The background factors underpinning discontinuation, as well as the challenges of sustaining patient engagement in real-world primary care settings, have yet to be fully elucidated. Moreover, most previous studies have examined the association between discontinuation and each attribute without considering interactions among patient characteristics, providing limited insight into the broader background patterns and multidimensional heterogeneity that may exist among patients who discontinue MNT. Discontinuation is unlikely to be attributable to a single risk factor; rather, it is likely shaped by the complex interplay of demographic, clinical, behavioural, and psychosocial influences. Accordingly, there is a need to move beyond treating discontinuers as a homogeneous group and instead to conceptualise them as comprising distinct subpopulations with divergent background profiles. Therefore, this study aimed to identify latent subgroups among patients who discontinued MNT in a primary care outpatient setting by applying latent class analysis (LCA) [ 19 – 21 ] to multidimensional background data, including demographic characteristics, clinical factors, lifestyle behaviours, and psychosocial factors. Grounded in the core primary care principles of comprehensiveness and continuity, the study sought to elucidate the heterogeneity within the population of MNT discontinuers and to clarify the subgroup-specific support needs that may inform clinical practice. By characterising this heterogeneity, the findings aim to provide a foundation for the development of tailored strategies to promote sustained engagement with MNT in routine primary care settings. Methods Study Design and Participants This retrospective observational study was conducted at a primary care clinic in Japan. The study period spanned from January 2019 to December 2024 and included patients who received outpatient MNT provided by registered dietitians upon physician referral, but who subsequently discontinued the therapy. The primary conditions for which MNT was prescribed included diabetes mellitus, dyslipidaemia, hypertension, obesity, liver disease, chronic kidney disease, cardiovascular disease, hyperuricaemia, malnutrition, and anaemia. This study is reported in accordance with the STROBE statement for observational studies. Discontinuation of MNT was defined as the absence of any subsequent MNT sessions for a period exceeding three months following the final documented nutrition counseling visit, with no further appointments scheduled. During the study period, 332 patients met the criteria for MNT discontinuation. Of these, 25 were excluded due to death (n = 13), transfer to another healthcare facility or hospitalisation (n = 8), or completion of treatment with formal discharge (n = 4). Additionally, seven patients who were under 18 years of age at the time of their initial MNT session were excluded. Accordingly, a total of 300 patients were included in the final analysis (Fig. 1). The sample size was determined by the number of eligible patients in the existing dataset and met commonly recommended criteria for latent class analysis. Among the 300 included patients, missing data were identified for residential area in two cases (0.7%) and for complications in four cases (1.3%). Although the overall proportion of missing data was low (< 5%), all available cases, including those with incomplete data, were retained in the analysis to minimise information loss. Missing data were addressed using a full-information maximum likelihood (FIML) approach. Figure 1. Flow diagram of participant selection Flow diagram illustrating the identification, screening, and inclusion of patients who discontinued outpatient medical nutrition therapy (MNT) at a primary care clinic. Data Collection and Variable Selection Patient data were retrospectively extracted from the electronic medical records of the primary care clinic. Information from physicians’ clinical notes, registered dietitians’ nutrition counseling records, and patient-completed questionnaires was integrated for analysis. The selection of observed variables was informed by previous studies investigating factors associated with MNT discontinuation, as well as by a conceptual framework that recognises adherence and discontinuation in chronic disease management as multifactorial phenomena. Demographic variables included age (< 50, 50–64, 65–74, and ≥ 75 years), sex (male or female), residential area (within or outside the ward in which the clinic was located), occupation status (including students), and whether the individual living with family (yes or no). Clinical variables comprised body mass index (BMI; <25 or ≥ 25 kg/m²), presence of complications, medication use, and presence of subjective symptoms. Lifestyle variables included meal regularity (skipping meals or consuming three meals per day), smoking status, alcohol consumption, and physical activity. Psychosocial variables included stage of change (precontemplation/contemplation, preparation, action/maintenance), perceived stress, illness awareness, financial concerns, and past nutrition therapy. For the LCA, all observed variables were treated as categorical and coded as sequential integers starting from 1, in accordance with the requirements of the poLCA package. Unless otherwise specified, binary variables were coded as No = 1 and Yes = 2. Statistical Analysis LCA was conducted to identify unobserved subgroups within the multidimensional dataset [ 22 ]. Model estimation was performed using maximum likelihood methods with 200 sets of randomly generated starting values to reduce the risk of convergence to local maxima. The maximum number of iterations was set to 10,000, and the convergence criterion was defined as 1 × 10⁻⁶ to ensure stable convergence to the optimal log-likelihood. A fixed seed for random number generation was specified to ensure reproducibility. Although the proportion of missing data was small (six cases; 2% of the sample), missing values were addressed using FIML to minimise information loss. Models comprising one to seven latent classes were estimated. The optimal number of classes was determined through a comprehensive assessment of model fit indices, including the Akaike Information Criterion (AIC), sample-size-adjusted Bayesian Information Criterion (SABIC), entropy, estimated class proportions, and clinical interpretability. Classification accuracy and model reliability were further evaluated using the bootstrap likelihood ratio test (BLRT), which compared models with K and K–1classes, and the average posterior probability of assignment (APPA) for each class. In the final model, individuals were assigned to classes based on the highest posterior probability. To support interpretation and facilitate comparisons of class-specific characteristics, heatmaps were generated to visualise item-response probabilities across latent classes. Additionally, given the clinical relevance of age and BMI as core indicators of life stage and nutritional–metabolic risk, bubble charts were constructed using these variables as axes to intuitively depict the clinical positioning and relative prevalence of each latent class. All statistical analyses were conducted in R (version 4.5.1), with LCA implemented using the poLCA package (version 1.6.0.1). Results Baseline Characteristics A total of 300 patients were included in the analysis. The median age was 62 years (interquartile range [IQR]: 51–71), and 43.3% of participants were male (Table 1). The median BMI was 24.9 kg/m² (IQR: 22.4–28.4). Regarding residential area, 75.8% of patients resided within the same ward as the clinic. In terms of sociodemographic factors, 46.0% of participants were employed (including students), and 84.7% were living with family members. With respect to clinical background, 27.0% had complications, and 37.7% reported medication use. As for lifestyle behaviours, 18.0% reported skipping meals, 31.7% reported alcohol consumption, and 23.7% were current smokers. Regular physical activity was reported by 25.3% of patients. From a psychosocial perspective, 14.0% of participants were in the precontemplation or contemplation stage of change, while 15.0% were in the action or maintenance stage. Perceived stress was reported by 87.0%, and 21.7% reported experiencing financial concerns. Table1. Baseline characteristics of participants (n = 300) Baseline demographic, clinical, lifestyle, and psychosocial characteristics of participants at study entry are shown. Values are presented as median (interquartile range) or number (%), unless otherwise indicated. Variables n （ %) Age, years, median (IQR) 62.0 (51.0–71.0) Sex, male 130 (43.3%) Body mass index (BMI), kg/m², median (IQR) 24.9 (22.4–28.4) Residential area, local 226 (75.8%) Occupation, yes 162 (54.0%) Living with family, yes 254 (84.7%) Complications, yes 80 (27.0%) Medication use, yes 113 (37.7%) Subjective symptoms present, yes 248 (82.7%) Meal regularity, regular 81 (27.0%) Alcohol consumption, yes 95 (31.7%) Smoking, yes 52 (17.3%) Physical activity, yes 90 (30.0%) Stage of change Precontemplation / Contemplation 42 (14.0%) Preparation 213 (71.0%) Action / Maintenance 45 (15.0%) Stress, yes 261 (87.0%) Illness awareness, yes 272 (90.7%) Financial concerns, yes 65 (21.7%) Past nutrition therapy, yes 108 (36.0%) Model Fit and Selection of Latent Classes Models comprising one to seven latent classes were evaluated. The AIC decreased monotonically as the number of classes increased. In contrast, the SABIC reached its lowest value with the five-class solution, although differences between the four- to seven-class models were modest. In terms of classification quality, entropy values met or exceeded the commonly accepted threshold of 0.80 for models with four or more classes (Supplementary Table S1). However, examination of class size distributions revealed that models with five or more classes included at least one class representing less than 10% of the sample. In the four-class model, the APPA ranged from 0.867 to 0.923, exceeding the recommended threshold of 0.70 and indicating good separation between classes (Supplementary Figure S1). The BLRT indicated statistically significant improvements in model fit across all comparisons from one- to five-class models (K–1 vs. K; all p < 0.005). In contrast, no significant improvements were observed when comparing the five- and six-class models (p = 0.100) or the six- and seven-class models (p = 0.085), suggesting that additional classes beyond five did not provide meaningful gains in model fit (Supplementary Figure S2). Based on these findings, and taking into account statistical stability, classification accuracy, and clinical interpretability, the four-class model was selected as the optimal solution. Characteristics of Identified Latent Classes In the four-class model, the distribution of participants across classes was as follows: Class 1 included 106 patients (35.3%), Class 2 included 106 patients (35.3%), Class 3 included 56 patients (18.7%), and Class 4 included 32 patients (10.7%). Class-specific item-response probabilities for each observed variable are shown in Figure 2. Based on the patterns of these response probabilities, each latent class demonstrated a distinct profile comprising a unique combination of demographic characteristics, clinical factors, lifestyle behaviours, and psychosocial factors. Figure 2. Item-response probabilities across four latent classes. Class-specific item-response probabilities for each observed variable estimated using latent class analysis with full-information maximum likelihood are shown as a heatmap. Darker colours indicate higher probabilities, and percentage values are displayed within cells. Abbreviations and selected variable definitions: residential area (Local = within the ward in which the clinic was located; Non-local = outside the ward); meal regularity (regular = consuming three meals per day; irregular = skipping meals); stage of change (Pre/Con = precontemplation/contemplation; Prep = preparation; Act/Man = action/maintenance). Latent Class Profiles Class 1 was predominantly composed of older adults, with 87% aged 65 years or older. Women accounted for 79% of this class, and the majority resided within the ward in which the clinic was located (90%), with a low proportion being employed (5%). Most patients had a BMI <25 kg/m² (65%), and the prevalence of complications (23%), alcohol consumption (22%), and smoking (10%) was low. Consuming three meals per day was common, reported by 92% of patients. Nearly all patients were in the preparation or maintenance stages of behaviour change (97%), and a similarly high proportion demonstrated illness awareness (97%). Class 2 consisted mainly of younger individuals, with 85% aged under 65 years, and was characterised by a high proportion of employed participants (96%). A BMI <25 kg/m² was observed in 57% of patients, and none had complications. Consuming three meals per day was reported by 76%, and only 8% were smokers. This class exhibited the highest prevalence of regular physical activity (40%) among all classes. Most individuals were in the preparation or maintenance stages of behaviour change (91%), and almost all demonstrated illness awareness (98%). In addition, 43% had past nutrition therapy. Class 3 was characterised by a predominance of men (68%) and a high prevalence of overweight or obesity, with 84% having a BMI ≥25 kg/m². This class exhibited high rates of complications (86%) and medication use (80%). The prevalence of subjective symptoms (39%) and perceived stress (98%) was also elevated. Notably, none of the individuals in this class were in the action or maintenance stages of behaviour change. Compared with other classes, a lower proportion demonstrated illness awareness, while a higher proportion reported financial concerns. Approximately half of the patients had past nutrition therapy. Class 4 was defined by a high prevalence of lifestyle-related risk behaviours. A BMI ≥25 kg/m² was observed in 63% of patients; meal skipping was common, with only 19% reporting consumption of three meals per day. High rates of smoking (62%) and alcohol consumption (56%) were also noted. Nearly half of the patients were in the precontemplation or contemplation stages of behaviour change (47%). All individuals in this class reported perceived stress (100%), and illness awareness was low. In contrast, no patients had complications or reported medication use. Visualization of Class Distribution and Cluster Separation Figure 3 depicts the distribution of the four latent classes using age and BMI as the two axes. Each bubble represents the class-specific median age and median BMI, with bubble size corresponding to the sample size of each class. This visualization illustrates distinct differences in age- and BMI-related profiles across the latent classes. Class 1 consisted predominantly of older adults, with a median age of 71.5 years (interquartile range [IQR]: 67.0–76.0) and a median BMI of 23.8 kg/m² (IQR: 21.4–26.6). In contrast, Classes 2 to 4 included younger to middle-aged adults with similar median ages (approximately 55 years), but exhibited marked variation in median BMI. These differences reflect the presence of three distinct latent subgroups with heterogeneous profiles, ranging from normal weight to overweight and obesity. Each bubble represents the median age and median BMI for a given latent class. Bubble size is proportional to the number of participants in each class. Discussion Identification of Latent Classes and Overview of the Discontinuation Population In this study, LCA was applied to patients who discontinued MNT delivered by registered dietitians in a primary care outpatient setting. The analysis identified four distinct latent classes with differing background characteristics. By integrating demographic characteristics, clinical factors, lifestyle behaviours, and psychosocial factors in a multidimensional framework, the findings demonstrated that patients who discontinue MNT represent a heterogeneous population that cannot be adequately explained by any single risk factor. Through the comprehensive characterisation of these subgroups, this study offers a foundation for developing more individualised strategies to support continuity of care in MNT. When the class structure was examined in relation to age and BMI, clear differences were observed between older adults and younger to middle-aged adults. Among older adults, patients were predominantly grouped into a single class, characterised by median BMI values within the normal range. In contrast, multiple latent classes emerged among younger to middle-aged adults, differentiated by varying median BMI values. These findings suggest that strategies to promote sustained engagement with MNT should be tailored to the patient’s life stage, taking into account the distinct characteristics and needs of each subgroup. Consistency with Previous Research and Novelty of the Present Study Previous studies on the discontinuation of MNT have predominantly employed multivariable analytical approaches to examine associations between discontinuation and individual factors such as age, BMI, complications, psychological characteristics, and socioeconomic status [15-18]. While these studies have provided valuable insights into correlates of discontinuation, they have been limited in their ability to capture the heterogeneity of the discontinuation population—particularly in terms of how patients with differing background characteristics cluster into distinct subgroups. To the best of our knowledge, this is the first study to apply LCA to patients who discontinued MNT delivered by registered dietitians in a primary care outpatient setting. The novelty of this study lies in its integrative approach, which simultaneously incorporates demographic, clinical, lifestyle, and psychosocial variables to reveal the structural heterogeneity underlying MNT discontinuation. This analytical framework provides a practical foundation for developing more individualised support strategies tailored to the needs of patients who discontinue MNT in primary care contexts. Interpretation of Each Latent Class and Clinical Implications Class 1: Community-Dwelling Older Women with Maintained Self-Care This class consisted predominantly of older adults who exhibited relatively favourable lifestyle behaviours and a high level of illness awareness, suggesting that self-management practices were generally well maintained. For these individuals, discontinuation of MNT may be less attributable to behavioural factors and more closely associated with non-behavioural challenges linked to aging—such as declining physical function, the burden of clinic visits, and the initiation or intensification of pharmacotherapy. Accordingly, rather than intensifying motivational or lifestyle counseling, interventions aimed at reducing access-related barriers—such as home-based MNT delivery—and enhancing social support networks may be more effective in promoting continuity of care. In this context, collaboration with community comprehensive support centres, public health nurses, and medical social workers may play a key role in sustaining MNT engagement [23-26]. Class 2: Low-Risk Working-Age Adults Patients in this class demonstrated high levels of illness awareness and were predominantly in the advanced stages of behaviour change. However, despite these favorable psychosocial characteristics, substantial time constraints related to employment may have impeded continued engagement in MNT. Additionally, a subjective perception of mild or absent symptoms may have reduced the perceived urgency of continuing therapy. In this group, discontinuation is likely driven by practical and perceptual barriers, rather than a lack of motivation. Therefore, support strategies that prioritise efficiency and convenience—such as tele-nutrition services, brief follow-up sessions, and flexible appointment scheduling—may be particularly appropriate for promoting sustained engagement with MNT among working-age adults [27,28]. Class 3: Men with High BMI and Complications Using Medication This class comprised patients with a relatively high medical burden, characterised by high BMI, the presence of complications, and medication use. The substantial physical, psychological, and social challenges associated with chronic conditions may have made sustained engagement with MNT particularly difficult for this group. In such cases, MNT delivered in isolation may be insufficient to address the complexity of patients’ needs. A more effective approach may involve the implementation of a comprehensive support system based on multidisciplinary collaboration, incorporating input from physicians, nurses, pharmacists, and physiotherapists. Such coordinated care models may be better positioned to address both the medical and behavioral dimensions of care and support long-term adherence to MNT [29]. Class 4: High-Risk Individuals with Low Illness Awareness Although patients in this class did not present with complications or report medication use, they exhibited multiple unhealthy lifestyle behaviours alongside psychosocial vulnerabilities. Discontinuation of MNT in this group is likely influenced by low illness awareness and high levels of perceived stress, both of which may impede readiness to change. For this subgroup, engagement may be better supported through behavioural science–informed approaches, such as motivational interviewing and stepwise goal setting. These techniques may help enhance intrinsic motivation, build self-efficacy, and facilitate progressive engagement with MNT over time [30]. Integrative Implications for Primary Care In primary care, clinicians are required to adopt a holistic perspective that extends beyond disease-specific features to encompass the patient as a whole, with an emphasis on comprehensive and continuous care. Registered dietitians likewise need to provide flexible, long-term nutritional support, adapting their approaches to patients’ life stages and evolving personal contexts. Within this framework, the present findings—demonstrating distinct support needs across latent classes—offer a practical basis for developing personalised strategies to deliver sustained and integrated support for patients who discontinue MNT in primary care settings. Limitations This study has several limitations. First, it was a retrospective, single-centre study conducted at a primary care clinic in Japan, which may limit the generalisability of the findings to other regions or healthcare systems. Second, some variables were based on self-reported information and may therefore have been subject to recall bias, particularly with respect to lifestyle behaviours and psychosocial factors. Furthermore, variability in the quality and consistency of clinical documentation—especially regarding illness awareness and stages of change—across healthcare providers may have influenced the accuracy of the data. Third, a fundamental limitation of LCA is that latent classes are probabilistic constructs derived from observed variables, and individual patients are not definitively assigned to a single class. Additionally, the labeling and interpretation of latent classes may vary depending on the researcher’s perspective and analytical decisions. While the identified latent subgroup structure offers clinically meaningful insights at the population level, cautious interpretation is warranted when applying these findings to individualised nutritional support. Finally, as the analysis was based solely on baseline characteristics at the point of discontinuation, it did not capture temporal trajectories or causal pathways leading to MNT dropout. Future research incorporating longitudinal comparisons with patients who maintain MNT, survival analyses, and latent transition analysis may provide a more comprehensive understanding of the dynamics underlying MNT discontinuation. Strengths This study has several notable strengths. First, it addresses an underexplored yet clinically relevant issue—the discontinuation of MNT in primary care outpatient settings—through the application of LCA, a method well suited to capturing population heterogeneity. Second, the analysis comprehensively integrated multidimensional data encompassing demographic characteristics, clinical factors, lifestyle behaviours, and psychosocial factors. Third, potential bias due to missing data was minimised through the use of an approach consistent with FIML. The statistical robustness and clinical plausibility of the model were rigorously evaluated using multiple criteria, including the AIC, SABIC, entropy, class size estimates, the BLRT, and APPA. Finally, the four identified latent classes were clinically interpretable and provide meaningful insights that may inform the development of tailored, evidence-based strategies to support continuity of care for patients receiving MNT in real-world primary care settings. Conclusion This study identified four latent classes among patients who discontinued dietitian-led MNT in a primary care outpatient setting, using LCA. The findings indicate that MNT discontinuation is not attributable to a single determinant but rather reflects diverse background patterns shaped by the complex interplay of demographic characteristics, clinical factors, lifestyle behaviours, and psychosocial factors. By elucidating the heterogeneity underlying MNT discontinuation, the study provides a foundation for developing tailored strategies to promote sustained engagement with MNT in primary care. Further longitudinal research is needed to generate evidence that informs the design of more comprehensive and continuous nutritional support systems. Abbreviations MNT: Medical nutrition therapy NCDs: noncommunicable diseases LCA: Latent class analysis FIML: full-information maximum likelihood BMI: body mass index AIC: Akaike Information Criterion SABIC: sample-size-adjusted Bayesian Information Criterion BLRT: bootstrap likelihood ratio test APPA: average posterior probability of assignment Declarations Ethics approval and consent to participate This study was conducted in accordance with the Declaration of Helsinki. The study protocol was reviewed and approved by the Ethics Committee of Teine Keijinkai Hospital (approval number: 4-024437-00). All data used in the analyses were fully anonymised and contained no personally identifiable information. As this was a retrospective observational study based on existing clinical records, the requirement for informed consent was waived by the Ethics Committee of Teine Keijinkai Hospital in accordance with the Ethical Guidelines for Life Science and Medical Research Involving Human Subjects in Japan. Consent for publication Not applicable Availability of data and materials The datasets generated and/or analysed during the current study are not publicly available due to privacy and ethical restrictions but are available from the corresponding author on reasonable request. Competing interests The authors declare that they have no competing interests Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. Authors' contributions Conceptualization: MA, Kazuki O, SM, Katsuhiko O Data Curation: MA Formal Analysis: MA Investigation: MA Methodology: MA, KA Project Administration: MU, Katsuhiko O Software: MA, Kazuki O, SM Supervision: Kazuki O, KA, SM, Katsuhiko O Visualization: MA Writing – Original Draft Preparation: MA Writing – Review & Editing: Kazuki O, KA, SM, Katsuhiko O All the authors have read and approved the final manuscript. Acknowledgements The authors would like to thank all those who supported this research. References Moloney L, Rozga M, Steiber A, Handu D. The Effectiveness of Medical Nutrition Therapy in Prevention and Treatment of Chronic Disease: A Position Paper of the Academy of Nutrition and Dietetics. J Acad Nutr Diet. 2026;126(2):156219. Moloney L, Chacón V, Devarakonda SLS, et al. Effectiveness of Medical Nutrition Therapy Provided by Registered Dietitian Nutritionists on Nutrition and Health Outcomes in Adults with Protein-Energy Malnutrition: A Systematic Review and Meta-Analysis. J Acad Nutr Diet. 2025;125(8):1144–e116120. Wells JC, Sawaya AL, Wibaek R, et al. The double burden of malnutrition: aetiological pathways and consequences for health. Lancet. 2020;395(10217):75–88. Popkin BM, Corvalan C, Grummer-Strawn LM. Dynamics of the double burden of malnutrition and the changing nutrition reality. Lancet. 2020;395(10217):65–74. Dent E, Wright ORL, Woo J, Hoogendijk EO. Malnutrition in older adults. Lancet. 2023;401(10380):951–66. Mohr AE, Hatem C, Sikand G, et al. Effectiveness of medical nutrition therapy in the management of adult dyslipidemia: A systematic review and meta-analysis. J Clin Lipidol. 2022;16(5):547–61. Dudzik JM, Senkus KE, Evert AB, et al. The effectiveness of medical nutrition therapy provided by a dietitian in adults with prediabetes: a systematic review and meta-analysis. Am J Clin Nutr. 2023;118(5):892–910. Papoutsakis C, Sundar C, Woodcock L, Abram JK, Lamers-Johnson E. Translating malnutrition care from the hospital to the community setting. Nutr Clin Pract. 2024;39(6):1292–8. Stratton RJ. Managing malnutrition and multimorbidity in primary care: dietary approaches to reduce treatment burden. Proc Nutr Soc. 2025;84(4):381–9. Schumacher TL, Jansson A, Herbert J, et al. The effectiveness of medical nutrition therapy for people at moderate to high risk of cardiovascular disease in an Australian rural primary care setting: 12-month results from a pragmatic cluster randomised controlled trial. BMC Health Serv Res. 2025;25(1):956. Mitchell LJ, Ball LE, Ross LJ, Barnes KA, Williams LT. Effectiveness of Dietetic Consultations in Primary Health Care: A Systematic Review of Randomized Controlled Trials. J Acad Nutr Diet. 2017;117(12):1941–62. Chan KS, Wan EY, Chin WY, et al. Effects of continuity of care on health outcomes among patients with diabetes mellitus and/or hypertension: a systematic review. BMC Fam Pract. 2021;22(1):145. Hussey PS, Schneider EC, Rudin RS, Fox DS, Lai J, Pollack CE. Continuity and the costs of care for chronic disease. JAMA Intern Med. 2014;174(5):742–8. American Diabetes Association Professional Practice Committee. Facilitating positive health behaviors and well-being to improve health outcomes: Standards of care in diabetes—2026. Diabetes Care. 2026;49(Suppl 1):S89–131. Colombo O, Ferretti VV, Ferraris C, et al. Is drop-out from obesity treatment a predictable and preventable event? Nutr J. 2014;13:13. Inelmen EM, Toffanello ED, Enzi G, et al. Predictors of drop-out in overweight and obese outpatients. Int J Obes (Lond). 2005;29(1):122–8. Everitt JD, Battista-Dowds EM, Heggs D, Hewlett P, Squire ALM. Determinants of completion and early dropout in an adult weight management service: a prospective observational study. J Hum Nutr Diet. 2023;36(5):1931–41. Følling IS, Klöckner C, Devle MT, Kulseng B. Preventing type 2 diabetes, overweight and obesity in the Norwegian primary healthcare: a longitudinal design with 60 months follow-up results and a cross-sectional design with comparison of dropouts versus completers. BMJ Open. 2022;12(3):e054841. Wendel Garcia PD, Caccioppola A, Coppola S, et al. Latent class analysis to predict intensive care outcomes in Acute Respiratory Distress Syndrome: a proposal of two pulmonary phenotypes. Crit Care. 2021;25(1):154. Grant RW, McCloskey J, Hatfield M, et al. Use of Latent Class Analysis and k-Means Clustering to Identify Complex Patient Profiles. JAMA Netw Open. 2020;3(12):e2029068. Hutchinson JM, Raffoul A, Pepetone A et al. Advances in methods for characterising dietary patterns: a scoping review. Br J Nutr Published online March 10, 2025. Sinha P, Calfee CS, Delucchi KL. Practitioner's Guide to Latent Class Analysis: Methodological Considerations and Common Pitfalls. Crit Care Med. 2021;49(1):e63–79. Gan DRY, Wister AV, Best JR. Environmental Influences on Life Satisfaction and Depressive Symptoms Among Older Adults With Multimorbidity: Path Analysis Through Loneliness in the Canadian Longitudinal Study on Aging. Gerontologist. 2022;62(6):855–64. Burnett J, Beauchamp JES. Public Health Strategies and Community Resources to Address Social Isolation and Loneliness in Older Adults. Clin Geriatr Med. 2025;41(3):369–89. Crocker TF, Ensor J, Lam N, et al. Community based complex interventions to sustain independence in older people: systematic review and network meta-analysis. BMJ. 2024;384:e077764. Kloppe T, Tetzlaff B, Mews C, Zimmermann T, Scherer M. Interprofessional collaboration to support patients with social problems in general practice-a qualitative focus group study. BMC Prim Care. 2022;23(1):169. Murray M, Bodenheimer T, Rittenhouse D, Grumbach K. Improving timely access to primary care: case studies of the advanced access model. JAMA. 2003;289(8):1042–6. Matulis JC, McCoy R. Patient-Centered Appointment Scheduling: a Call for Autonomy, Continuity, and Creativity. J Gen Intern Med. 2021;36(2):511–4. Zhou Y, Dai X, Ni Y, et al. Interventions and management on multimorbidity: An overview of systematic reviews. Ageing Res Rev. 2023;87:101901. Yanovski SZ, Yanovski JA. Approach to Obesity Treatment in Primary Care: A Review. JAMA Intern Med. 2024;184(7):818–29. Additional Declarations No competing interests reported. Supplementary Files BMCPCLCASupplementary20260228.pdf Cite Share Download PDF Status: Under Review Version 1 posted Reviewers invited by journal 03 Apr, 2026 Editor assigned by journal 31 Mar, 2026 Editor invited by journal 05 Mar, 2026 Submission checks completed at journal 03 Mar, 2026 First submitted to journal 03 Mar, 2026 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-8998867","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":619787366,"identity":"41223224-71ab-4772-83ab-066ad2533ad8","order_by":0,"name":"Machiko Ukai","email":"","orcid":"","institution":"Hokkaido University","correspondingAuthor":false,"prefix":"","firstName":"Machiko","middleName":"","lastName":"Ukai","suffix":""},{"id":619787367,"identity":"ec55068d-944b-4c89-8de8-831cb4238206","order_by":1,"name":"Kazuki Ohashi","email":"","orcid":"","institution":"Hokkaido University","correspondingAuthor":false,"prefix":"","firstName":"Kazuki","middleName":"","lastName":"Ohashi","suffix":""},{"id":619787368,"identity":"551494f0-b122-4131-be45-5716be07b9b2","order_by":2,"name":"Kazuhiro Abe","email":"","orcid":"","institution":"Hokkaido University","correspondingAuthor":false,"prefix":"","firstName":"Kazuhiro","middleName":"","lastName":"Abe","suffix":""},{"id":619787369,"identity":"25f515e3-a0cc-42f5-b330-45050d731fa8","order_by":3,"name":"Soki Maeda","email":"","orcid":"","institution":"Mitsubishi UFJ Information Technology, Ltd.","correspondingAuthor":false,"prefix":"","firstName":"Soki","middleName":"","lastName":"Maeda","suffix":""},{"id":619787370,"identity":"90cab60d-c5f7-41c5-a530-fa144c7d677e","order_by":4,"name":"Katsuhiko Ogasawara","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA60lEQVRIiWNgGAWjYHACxsdAIoGBh4FBgsEALJJASAuzMcla2KQRWogB/PyLn1UX/GHI4+c5Y3jjRwGDPH8Dw7MH+LRIznhmdntmG0OxZG+PsWWPAYPhjAMM6Qb4tBjcOGB2m7eBIXHDeR4zCR4DBsYNDAxpeF1ocOP4t2KePxAtkn8MGOwJaznfY8bMwwbUcrbHTBpoSyJBLZIzeIqledskEmf2HCu2ljGQSJ5xmIBf+PmPb/zM88cmsZ8neePNN39sbPvbe9Ie4NPCIJEAJuFcYNTypOHVwcB/AEOI/Rh+LaNgFIyCUTDSAAAT50NALXk0iAAAAABJRU5ErkJggg==","orcid":"","institution":"Hokkaido University","correspondingAuthor":true,"prefix":"","firstName":"Katsuhiko","middleName":"","lastName":"Ogasawara","suffix":""}],"badges":[],"createdAt":"2026-03-01 03:08:43","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8998867/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8998867/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106724763,"identity":"97fe66ee-e775-4f94-ae34-e59252150cbf","added_by":"auto","created_at":"2026-04-12 18:29:37","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":72698,"visible":true,"origin":"","legend":"\u003cp\u003eFlow diagram of participant selection\u003c/p\u003e\n\u003cp\u003eFlow diagram illustrating the identification, screening, and inclusion of patients who discontinued outpatient medical nutrition therapy (MNT) at a primary care clinic.\u003c/p\u003e","description":"","filename":"BMCPCLCAFigure120260228.png","url":"https://assets-eu.researchsquare.com/files/rs-8998867/v1/abdec22ad7eef316462c3b61.png"},{"id":106544873,"identity":"1a7fbc0d-c130-4723-999f-115041a35238","added_by":"auto","created_at":"2026-04-09 16:42:45","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":114865,"visible":true,"origin":"","legend":"\u003cp\u003eItem-response probabilities across four latent classes. Class-specific item-response probabilities for each observed variable estimated using latent class analysis with full-information maximum likelihood are shown as a heatmap. Darker colours indicate higher probabilities, and percentage values are displayed within cells.\u003c/p\u003e\n\u003cp\u003eAbbreviations and selected variable definitions: residential area (Local = within the ward in which the clinic was located; Non-local = outside the ward); meal regularity (regular = consuming three meals per day; irregular = skipping meals); stage of change (Pre/Con = precontemplation/contemplation; Prep = preparation; Act/Man = action/maintenance).\u003c/p\u003e","description":"","filename":"BMCPCLCAFigure220260228.png","url":"https://assets-eu.researchsquare.com/files/rs-8998867/v1/dc1b8c2c770bf65d84d9a902.png"},{"id":106544874,"identity":"46bb01e6-ac1d-47b9-8fe9-80408307a3b0","added_by":"auto","created_at":"2026-04-09 16:42:45","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":32243,"visible":true,"origin":"","legend":"\u003cp\u003eAge and BMI distribution across latent classes\u003c/p\u003e\n\u003cp\u003eEach bubble represents the median age and median BMI for a given latent class. Bubble size is proportional to the number of participants in each class.\u003c/p\u003e","description":"","filename":"BMCPCLCAFigure320260228.png","url":"https://assets-eu.researchsquare.com/files/rs-8998867/v1/775f6b1c2d8831ca01f36a3b.png"},{"id":106726351,"identity":"8cb4d92b-7717-437d-90bb-de0556dd0a35","added_by":"auto","created_at":"2026-04-12 18:35:54","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1027254,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8998867/v1/0ca05e4a-ca1c-491b-8796-f2622dfd3478.pdf"},{"id":106544871,"identity":"0ac835a0-fefc-45cf-8218-8a2697c57a21","added_by":"auto","created_at":"2026-04-09 16:42:45","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":154492,"visible":true,"origin":"","legend":"","description":"","filename":"BMCPCLCASupplementary20260228.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8998867/v1/72f532e473dff8d57829bd04.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Heterogeneity Among Patients Discontinuing Medical Nutrition Therapy in Primary Care: A Latent Class Analysis","fulltext":[{"header":"Background","content":"\u003cp\u003eMedical nutrition therapy (MNT) has demonstrated efficacy in enhancing health outcomes across a wide spectrum of conditions, including noncommunicable diseases (NCDs), obesity, and malnutrition [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. In recent years, the global escalation of the double burden of malnutrition\u0026mdash;defined by the simultaneous presence of undernutrition and overnutrition\u0026mdash;has further emphasised the critical role of MNT in addressing complex nutritional challenges [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. This dual burden underscores the pressing need for effective and adaptable nutritional interventions that respond to the evolving demands of population health [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan additionalcitationids=\"CR6\" citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIn primary care settings, where the delivery of continuous and comprehensive care for community-dwelling individuals is a foundational principle, the integration of MNT into routine clinical practice is increasingly regarded as a pivotal strategy to support behavioural change and improve disease trajectories [\u003cspan additionalcitationids=\"CR9 CR10\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Sustained patient engagement is essential to fully realise the potential benefits of MNT; however, in real-world clinical contexts, a substantial proportion of patients discontinue MNT prematurely. Ongoing support within the framework of chronic disease management has been associated with improved adherence, better quality of life, and reductions in complications, hospital admissions, and healthcare costs [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Consequently, premature discontinuation of MNT presents a significant clinical challenge, as it limits the effectiveness of nutritional interventions and impedes optimal health outcomes [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e].\u003c/p\u003e \u003cp\u003ePrevious studies on the discontinuation of MNT have predominantly focused on hospital-based specialist outpatient clinics or structured weight-loss intervention programmes targeting obesity treatment and weight management [\u003cspan additionalcitationids=\"CR16 CR17\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. These studies have reported associations between discontinuation and several individual factors, including younger age, lower baseline adiposity or poorer early weight loss response, employment\u0026mdash;particularly full-time work, fewer obesity-related comorbidities, and psychological characteristics such as higher levels of psychological distress or hostility. Associations with BMI and travel distance to clinics, however, have been inconsistent across studies. In contrast, evidence concerning MNT delivered by registered dietitians within routine primary care remains scarce. The background factors underpinning discontinuation, as well as the challenges of sustaining patient engagement in real-world primary care settings, have yet to be fully elucidated. Moreover, most previous studies have examined the association between discontinuation and each attribute without considering interactions among patient characteristics, providing limited insight into the broader background patterns and multidimensional heterogeneity that may exist among patients who discontinue MNT. Discontinuation is unlikely to be attributable to a single risk factor; rather, it is likely shaped by the complex interplay of demographic, clinical, behavioural, and psychosocial influences. Accordingly, there is a need to move beyond treating discontinuers as a homogeneous group and instead to conceptualise them as comprising distinct subpopulations with divergent background profiles.\u003c/p\u003e \u003cp\u003eTherefore, this study aimed to identify latent subgroups among patients who discontinued MNT in a primary care outpatient setting by applying latent class analysis (LCA) [\u003cspan additionalcitationids=\"CR20\" citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e] to multidimensional background data, including demographic characteristics, clinical factors, lifestyle behaviours, and psychosocial factors. Grounded in the core primary care principles of comprehensiveness and continuity, the study sought to elucidate the heterogeneity within the population of MNT discontinuers and to clarify the subgroup-specific support needs that may inform clinical practice. By characterising this heterogeneity, the findings aim to provide a foundation for the development of tailored strategies to promote sustained engagement with MNT in routine primary care settings.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy Design and Participants\u003c/h2\u003e \u003cp\u003eThis retrospective observational study was conducted at a primary care clinic in Japan. The study period spanned from January 2019 to December 2024 and included patients who received outpatient MNT provided by registered dietitians upon physician referral, but who subsequently discontinued the therapy. The primary conditions for which MNT was prescribed included diabetes mellitus, dyslipidaemia, hypertension, obesity, liver disease, chronic kidney disease, cardiovascular disease, hyperuricaemia, malnutrition, and anaemia. This study is reported in accordance with the STROBE statement for observational studies.\u003c/p\u003e \u003cp\u003eDiscontinuation of MNT was defined as the absence of any subsequent MNT sessions for a period exceeding three months following the final documented nutrition counseling visit, with no further appointments scheduled. During the study period, 332 patients met the criteria for MNT discontinuation. Of these, 25 were excluded due to death (n\u0026thinsp;=\u0026thinsp;13), transfer to another healthcare facility or hospitalisation (n\u0026thinsp;=\u0026thinsp;8), or completion of treatment with formal discharge (n\u0026thinsp;=\u0026thinsp;4). Additionally, seven patients who were under 18 years of age at the time of their initial MNT session were excluded. Accordingly, a total of 300 patients were included in the final analysis (Fig.\u0026nbsp;1). The sample size was determined by the number of eligible patients in the existing dataset and met commonly recommended criteria for latent class analysis. Among the 300 included patients, missing data were identified for residential area in two cases (0.7%) and for complications in four cases (1.3%). Although the overall proportion of missing data was low (\u0026lt;\u0026thinsp;5%), all available cases, including those with incomplete data, were retained in the analysis to minimise information loss. Missing data were addressed using a full-information maximum likelihood (FIML) approach.\u003c/p\u003e \u003cp\u003eFigure 1. Flow diagram of participant selection\u003c/p\u003e \u003cp\u003eFlow diagram illustrating the identification, screening, and inclusion of patients who discontinued outpatient medical nutrition therapy (MNT) at a primary care clinic.\u003c/p\u003e \u003cp\u003e \u003cdiv description=\"\" class=\"Drawing\" id=\"1709459627\" name=\"図 1\"\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eData Collection and Variable Selection\u003c/h3\u003e\n\u003cp\u003ePatient data were retrospectively extracted from the electronic medical records of the primary care clinic. Information from physicians\u0026rsquo; clinical notes, registered dietitians\u0026rsquo; nutrition counseling records, and patient-completed questionnaires was integrated for analysis. The selection of observed variables was informed by previous studies investigating factors associated with MNT discontinuation, as well as by a conceptual framework that recognises adherence and discontinuation in chronic disease management as multifactorial phenomena. Demographic variables included age (\u0026lt;\u0026thinsp;50, 50\u0026ndash;64, 65\u0026ndash;74, and \u0026ge;\u0026thinsp;75 years), sex (male or female), residential area (within or outside the ward in which the clinic was located), occupation status (including students), and whether the individual living with family (yes or no). Clinical variables comprised body mass index (BMI; \u0026lt;25 or \u0026ge;\u0026thinsp;25 kg/m\u0026sup2;), presence of complications, medication use, and presence of subjective symptoms. Lifestyle variables included meal regularity (skipping meals or consuming three meals per day), smoking status, alcohol consumption, and physical activity. Psychosocial variables included stage of change (precontemplation/contemplation, preparation, action/maintenance), perceived stress, illness awareness, financial concerns, and past nutrition therapy. For the LCA, all observed variables were treated as categorical and coded as sequential integers starting from 1, in accordance with the requirements of the poLCA package. Unless otherwise specified, binary variables were coded as No\u0026thinsp;=\u0026thinsp;1 and Yes\u0026thinsp;=\u0026thinsp;2.\u003c/p\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003eLCA was conducted to identify unobserved subgroups within the multidimensional dataset [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Model estimation was performed using maximum likelihood methods with 200 sets of randomly generated starting values to reduce the risk of convergence to local maxima. The maximum number of iterations was set to 10,000, and the convergence criterion was defined as 1 \u0026times; 10⁻⁶ to ensure stable convergence to the optimal log-likelihood. A fixed seed for random number generation was specified to ensure reproducibility. Although the proportion of missing data was small (six cases; 2% of the sample), missing values were addressed using FIML to minimise information loss.\u003c/p\u003e \u003cp\u003eModels comprising one to seven latent classes were estimated. The optimal number of classes was determined through a comprehensive assessment of model fit indices, including the Akaike Information Criterion (AIC), sample-size-adjusted Bayesian Information Criterion (SABIC), entropy, estimated class proportions, and clinical interpretability. Classification accuracy and model reliability were further evaluated using the bootstrap likelihood ratio test (BLRT), which compared models with K and K\u0026ndash;1classes, and the average posterior probability of assignment (APPA) for each class. In the final model, individuals were assigned to classes based on the highest posterior probability.\u003c/p\u003e \u003cp\u003eTo support interpretation and facilitate comparisons of class-specific characteristics, heatmaps were generated to visualise item-response probabilities across latent classes. Additionally, given the clinical relevance of age and BMI as core indicators of life stage and nutritional\u0026ndash;metabolic risk, bubble charts were constructed using these variables as axes to intuitively depict the clinical positioning and relative prevalence of each latent class. All statistical analyses were conducted in R (version 4.5.1), with LCA implemented using the poLCA package (version 1.6.0.1).\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eBaseline Characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA total of 300 patients were included in the analysis. The median age was 62 years (interquartile range [IQR]: 51\u0026ndash;71), and 43.3% of participants were male (Table 1). The median BMI was 24.9 kg/m\u0026sup2; (IQR: 22.4\u0026ndash;28.4). Regarding residential area, 75.8% of patients resided within the same ward as the clinic. In terms of sociodemographic factors, 46.0% of participants were employed (including students), and 84.7% were living with family members. With respect to clinical background, 27.0% had complications, and 37.7% reported medication use. As for lifestyle behaviours, 18.0% reported skipping meals, 31.7% reported alcohol consumption, and 23.7% were current smokers. Regular physical activity was reported by 25.3% of patients. From a psychosocial perspective, 14.0% of participants were in the precontemplation or contemplation stage of change, while 15.0% were in the action or maintenance stage. Perceived stress was reported by 87.0%, and 21.7% reported experiencing financial concerns.\u003c/p\u003e\n\u003cp\u003eTable1. Baseline characteristics of participants (n = 300)\u003c/p\u003e\n\u003cp\u003eBaseline demographic, clinical, lifestyle, and psychosocial characteristics of participants at study entry are shown. Values are presented as median (interquartile range) or number (%), unless otherwise indicated.\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"562\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003eVariables\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cstrong\u003en\u003c/strong\u003e\u003cstrong\u003e（\u003c/strong\u003e\u003cstrong\u003e%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAge, years, median (IQR)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e62.0 (51.0\u0026ndash;71.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSex, male\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e130 (43.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBody mass index (BMI), kg/m\u0026sup2;, median (IQR)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e24.9 (22.4\u0026ndash;28.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eResidential area, local \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e226 (75.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eOccupation, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e162 (54.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLiving with family, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e254 (84.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eComplications, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e80 (27.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMedication use, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e113 (37.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSubjective symptoms present, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e248 (82.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMeal regularity, regular\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e81 (27.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAlcohol consumption, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e95 (31.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSmoking, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e52 (17.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePhysical activity, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e90 (30.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" colspan=\"2\"\u003e\n \u003cp\u003eStage of change\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026emsp;Precontemplation / Contemplation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e42 (14.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026emsp;Preparation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e213 (71.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026emsp;Action / Maintenance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e45 (15.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStress, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e261 (87.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIllness awareness, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e272 (90.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eFinancial concerns, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e65 (21.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePast nutrition therapy, yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e108 (36.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eModel Fit and Selection of Latent Classes\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eModels comprising one to seven latent classes were evaluated. The AIC decreased monotonically as the number of classes increased. In contrast, the SABIC reached its lowest value with the five-class solution, although differences between the four- to seven-class models were modest. In terms of classification quality, entropy values met or exceeded the commonly accepted threshold of 0.80 for models with four or more classes (Supplementary Table S1). However, examination of class size distributions revealed that models with five or more classes included at least one class representing less than 10% of the sample.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn the four-class model, the APPA ranged from 0.867 to 0.923, exceeding the recommended threshold of 0.70 and indicating good separation between classes (Supplementary Figure S1). The BLRT indicated statistically significant improvements in model fit across all comparisons from one- to five-class models (K\u0026ndash;1 vs. K; all p \u0026lt; 0.005). In contrast, no significant improvements were observed when comparing the five- and six-class models (p = 0.100) or the six- and seven-class models (p = 0.085), suggesting that additional classes beyond five did not provide meaningful gains in model fit (Supplementary Figure S2).\u003c/p\u003e\n\u003cp\u003eBased on these findings, and taking into account statistical stability, classification accuracy, and clinical interpretability, the four-class model was selected as the optimal solution.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCharacteristics of Identified Latent Classes\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn the four-class model, the distribution of participants across classes was as follows: Class 1 included 106 patients (35.3%), Class 2 included 106 patients (35.3%), Class 3 included 56 patients (18.7%), and Class 4 included 32 patients (10.7%). Class-specific item-response probabilities for each observed variable are shown in Figure 2. Based on the patterns of these response probabilities, each latent class demonstrated a distinct profile comprising a unique combination of demographic characteristics, clinical factors, lifestyle behaviours, and psychosocial factors.\u003c/p\u003e\n\u003cp\u003eFigure 2. Item-response probabilities across four latent classes. Class-specific item-response probabilities for each observed variable estimated using latent class analysis with full-information maximum likelihood are shown as a heatmap. Darker colours indicate higher probabilities, and percentage values are displayed within cells.\u003c/p\u003e\n\u003cp\u003eAbbreviations and selected variable definitions: residential area (Local = within the ward in which the clinic was located; Non-local = outside the ward); meal regularity (regular = consuming three meals per day; irregular = skipping meals); stage of change (Pre/Con = precontemplation/contemplation; Prep = preparation; Act/Man = action/maintenance).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLatent Class Profiles\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eClass 1 was predominantly composed of older adults, with 87% aged 65 years or older. Women accounted for 79% of this class, and the majority resided \u003cstrong\u003ewithin the ward in which the clinic was located\u003c/strong\u003e (90%), with a low proportion \u003cstrong\u003ebeing employed\u003c/strong\u003e (5%). Most patients had a BMI \u0026lt;25 kg/m\u0026sup2; (65%), and the prevalence of complications (23%), alcohol consumption (22%), and smoking (10%) was low. \u003cstrong\u003eConsuming three meals per day\u003c/strong\u003e was common, reported by 92% of patients. Nearly all patients were in the preparation or maintenance stages of behaviour change (97%), and a similarly high proportion demonstrated illness awareness (97%).\u003c/p\u003e\n\u003cp\u003eClass 2 consisted mainly of younger individuals, with 85% aged under 65 years, and was characterised by a high proportion of employed participants (96%). A BMI \u0026lt;25 kg/m\u0026sup2; was observed in 57% of patients, and none had complications. Consuming three meals per day was reported by 76%, and only 8% were smokers. This class exhibited the highest prevalence of regular physical activity (40%) among all classes. Most individuals were in the preparation or maintenance stages of behaviour change (91%), and almost all demonstrated illness awareness (98%). In addition, 43% had past nutrition therapy.\u003c/p\u003e\n\u003cp\u003eClass 3 was characterised by a predominance of men (68%) and a high prevalence of overweight or obesity, with 84% having a BMI \u0026ge;25 kg/m\u0026sup2;. This class exhibited high rates of complications (86%) and medication use (80%). The prevalence of subjective symptoms (39%) and perceived stress (98%) was also elevated. Notably, none of the individuals in this class were in the action or maintenance stages of behaviour change. Compared with other classes, a lower proportion demonstrated illness awareness, while a higher proportion reported financial concerns. Approximately half of the patients had past nutrition therapy.\u003c/p\u003e\n\u003cp\u003eClass 4 was defined by a high prevalence of lifestyle-related risk behaviours. A BMI \u0026ge;25 kg/m\u0026sup2; was observed in 63% of patients; meal skipping was common, with only 19% reporting consumption of three meals per day. High rates of smoking (62%) and alcohol consumption (56%) were also noted. Nearly half of the patients were in the precontemplation or contemplation stages of behaviour change (47%). All individuals in this class reported perceived stress (100%), and illness awareness was low. In contrast, no patients had complications or reported medication use.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eVisualization of Class Distribution and Cluster Separation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFigure 3 depicts the distribution of the four latent classes using age and BMI as the two axes. Each bubble represents the class-specific median age and median BMI, with bubble size corresponding to the sample size of each class. This visualization illustrates distinct differences in age- and BMI-related profiles across the latent classes.\u003c/p\u003e\n\u003cp\u003eClass 1 consisted predominantly of older adults, with a median age of 71.5 years (interquartile range [IQR]: 67.0\u0026ndash;76.0) and a median BMI of 23.8 kg/m\u0026sup2; (IQR: 21.4\u0026ndash;26.6). In contrast, Classes 2 to 4 included younger to middle-aged adults with similar median ages (approximately 55 years), but exhibited marked variation in median BMI. These differences reflect the presence of three distinct latent subgroups with heterogeneous profiles, ranging from normal weight to overweight and obesity.\u003c/p\u003e\n\u003cp\u003eEach bubble represents the median age and median BMI for a given latent class. Bubble size is proportional to the number of participants in each class.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003e\u003cstrong\u003eIdentification of Latent Classes and Overview of the Discontinuation Population\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn this study, LCA was applied to patients who discontinued MNT delivered by registered dietitians in a primary care outpatient setting. The analysis identified four distinct latent classes with differing background characteristics. By integrating demographic characteristics, clinical factors, lifestyle behaviours, and psychosocial factors in a multidimensional framework, the findings demonstrated that patients who discontinue MNT represent a heterogeneous population that cannot be adequately explained by any single risk factor. Through the comprehensive characterisation of these subgroups, this study offers a foundation for developing more individualised strategies to support continuity of care in MNT.\u003c/p\u003e\n\u003cp\u003eWhen the class structure was examined in relation to age and BMI, clear differences were observed between older adults and younger to middle-aged adults. Among older adults, patients were predominantly grouped into a single class, characterised by median BMI values within the normal range. In contrast, multiple latent classes emerged among younger to middle-aged adults, differentiated by varying median BMI values. These findings suggest that strategies to promote sustained engagement with MNT should be tailored to the patient’s life stage, taking into account the distinct characteristics and needs of each subgroup.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsistency with Previous Research and Novelty of the Present Study\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePrevious studies on the discontinuation of MNT have predominantly employed multivariable analytical approaches to examine associations between discontinuation and individual factors such as age, BMI, complications, psychological characteristics, and socioeconomic status [15-18]. While these studies have provided valuable insights into correlates of discontinuation, they have been limited in their ability to capture the heterogeneity of the discontinuation population—particularly in terms of how patients with differing background characteristics cluster into distinct subgroups.\u003c/p\u003e\n\u003cp\u003eTo the best of our knowledge, this is the first study to apply LCA to patients who discontinued MNT delivered by registered dietitians in a primary care outpatient setting. The novelty of this study lies in its integrative approach, which simultaneously incorporates demographic, clinical, lifestyle, and psychosocial variables to reveal the structural heterogeneity underlying MNT discontinuation. This analytical framework provides a practical foundation for developing more individualised support strategies tailored to the needs of patients who discontinue MNT in primary care contexts.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInterpretation of Each Latent Class and Clinical Implications\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClass 1: Community-Dwelling Older Women with Maintained Self-Care\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis class consisted predominantly of older adults who exhibited relatively favourable lifestyle behaviours and a high level of illness awareness, suggesting that self-management practices were generally well maintained. For these individuals, discontinuation of MNT may be less attributable to behavioural factors and more closely associated with non-behavioural challenges linked to aging—such as declining physical function, the burden of clinic visits, and the initiation or intensification of pharmacotherapy. Accordingly, rather than intensifying motivational or lifestyle counseling, interventions aimed at reducing access-related barriers—such as home-based MNT delivery—and enhancing social support networks may be more effective in promoting continuity of care. In this context, collaboration with community comprehensive support centres, public health nurses, and medical social workers may play a key role in sustaining MNT engagement [23-26].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClass 2: Low-Risk Working-Age Adults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePatients in this class demonstrated high levels of illness awareness and were predominantly in the advanced stages of behaviour change. However, despite these favorable psychosocial characteristics, substantial time constraints related to employment may have impeded continued engagement in MNT. Additionally, a subjective perception of mild or absent symptoms may have reduced the perceived urgency of continuing therapy. In this group, discontinuation is likely driven by practical and perceptual barriers, rather than a lack of motivation. Therefore, support strategies that prioritise efficiency and convenience—such as tele-nutrition services, brief follow-up sessions, and flexible appointment scheduling—may be particularly appropriate for promoting sustained engagement with MNT among working-age adults [27,28].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClass 3: Men with High BMI and Complications Using Medication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis class comprised patients with a relatively high medical burden, characterised by high BMI, the presence of complications, and medication use. The substantial physical, psychological, and social challenges associated with chronic conditions may have made sustained engagement with MNT particularly difficult for this group. In such cases, MNT delivered in isolation may be insufficient to address the complexity of patients’ needs. A more effective approach may involve the implementation of a comprehensive support system based on multidisciplinary collaboration, incorporating input from physicians, nurses, pharmacists, and physiotherapists. Such coordinated care models may be better positioned to address both the medical and behavioral dimensions of care and support long-term adherence to MNT [29].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClass 4: High-Risk Individuals with Low Illness Awareness\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAlthough patients in this class did not present with complications or report medication use, they exhibited multiple unhealthy lifestyle behaviours alongside psychosocial vulnerabilities. Discontinuation of MNT in this group is likely influenced by low illness awareness and high levels of perceived stress, both of which may impede readiness to change.\u0026nbsp;For this subgroup, engagement may be better supported through behavioural science–informed approaches, such as motivational interviewing and stepwise goal setting. These techniques may help enhance intrinsic motivation, build self-efficacy, and facilitate progressive engagement with MNT over time [30].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eIntegrative Implications for Primary Care\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn primary care, clinicians are required to adopt a holistic perspective that extends beyond disease-specific features to encompass the patient as a whole, with an emphasis on comprehensive and continuous care. Registered dietitians likewise need to provide flexible, long-term nutritional support, adapting their approaches to patients’ life stages and evolving personal contexts. Within this framework, the present findings—demonstrating distinct support needs across latent classes—offer a practical basis for developing personalised strategies to deliver sustained and integrated support for patients who discontinue MNT in primary care settings.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLimitations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study has several limitations. First, it was a retrospective, single-centre study conducted at a primary care clinic in Japan, which may limit the generalisability of the findings to other regions or healthcare systems.\u0026nbsp;Second, some variables were based on self-reported information and may therefore have been subject to recall bias, particularly with respect to lifestyle behaviours and psychosocial factors. Furthermore, variability in the quality and consistency of clinical documentation—especially regarding illness awareness and stages of change—across healthcare providers may have influenced the accuracy of the data. Third, a fundamental limitation of LCA is that latent classes are probabilistic constructs derived from observed variables, and individual patients are not definitively assigned to a single class. Additionally, the labeling and interpretation of latent classes may vary depending on the researcher’s perspective and analytical decisions. While the identified latent subgroup structure offers clinically meaningful insights at the population level, cautious interpretation is warranted when applying these findings to individualised nutritional support. Finally, as the analysis was based solely on baseline characteristics at the point of discontinuation, it did not capture temporal trajectories or causal pathways leading to MNT dropout. Future research incorporating longitudinal comparisons with patients who maintain MNT, survival analyses, and latent transition analysis \u0026nbsp;may provide a more comprehensive understanding of the dynamics underlying MNT discontinuation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStrengths\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study has several notable strengths. First, it addresses an underexplored yet clinically relevant issue—the discontinuation of MNT in primary care outpatient settings—through the application of LCA, a method well suited to capturing population heterogeneity. Second, the analysis comprehensively integrated multidimensional data encompassing demographic characteristics, clinical factors, lifestyle behaviours, and psychosocial factors. Third, potential bias due to missing data was minimised through the use of an approach consistent with FIML. The statistical robustness and clinical plausibility of the model were rigorously evaluated using multiple criteria, including the AIC, SABIC, entropy, class size estimates, the BLRT, and\u0026nbsp;APPA.\u003c/p\u003e\n\u003cp\u003eFinally, the four identified latent classes were clinically interpretable and provide meaningful insights that may inform the development of tailored, evidence-based strategies to support continuity of care for patients receiving MNT in real-world primary care settings.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study identified four latent classes among patients who discontinued dietitian-led MNT in a primary care outpatient setting, using LCA. The findings indicate that MNT discontinuation is not attributable to a single determinant but rather reflects diverse background patterns shaped by the complex interplay of demographic characteristics, clinical factors, lifestyle behaviours, and psychosocial factors. By elucidating the heterogeneity underlying MNT discontinuation, the study provides a foundation for developing tailored strategies to promote sustained engagement with MNT in primary care. Further longitudinal research is needed to generate evidence that informs the design of more comprehensive and continuous nutritional support systems.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eMNT: Medical nutrition therapy\u003c/p\u003e \u003cp\u003eNCDs: noncommunicable diseases\u003c/p\u003e \u003cp\u003eLCA: Latent class analysis\u003c/p\u003e \u003cp\u003eFIML: full-information maximum likelihood\u003c/p\u003e \u003cp\u003eBMI: body mass index\u003c/p\u003e \u003cp\u003eAIC: Akaike Information Criterion\u003c/p\u003e \u003cp\u003eSABIC: sample-size-adjusted Bayesian Information Criterion\u003c/p\u003e \u003cp\u003eBLRT: bootstrap likelihood ratio test\u003c/p\u003e \u003cp\u003eAPPA: average posterior probability of assignment\u003c/p\u003e "},{"header":"Declarations","content":"\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. The study protocol was reviewed and approved by the Ethics Committee of Teine Keijinkai Hospital (approval number: 4-024437-00). All data used in the analyses were fully anonymised and contained no personally identifiable information. As this was a retrospective observational study based on existing clinical records, the requirement for informed consent was waived by the Ethics Committee of Teine Keijinkai Hospital in accordance with the Ethical Guidelines for \u003cem\u003eLife Science and Medical Research Involving Human Subjects\u003c/em\u003e in Japan.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets generated and/or analysed during the current study are not publicly available due to privacy and ethical restrictions but 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\u003eThis research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConceptualization: MA, Kazuki O, SM, Katsuhiko O\u003c/p\u003e\n\u003cp\u003eData Curation: MA\u003c/p\u003e\n\u003cp\u003eFormal Analysis: MA\u003c/p\u003e\n\u003cp\u003eInvestigation: MA\u003c/p\u003e\n\u003cp\u003eMethodology: MA, KA\u003c/p\u003e\n\u003cp\u003eProject Administration: MU, Katsuhiko O\u003c/p\u003e\n\u003cp\u003eSoftware: MA, Kazuki O, SM\u003c/p\u003e\n\u003cp\u003eSupervision: Kazuki O, KA, SM, Katsuhiko O\u003c/p\u003e\n\u003cp\u003eVisualization: MA\u003c/p\u003e\n\u003cp\u003eWriting \u0026ndash; Original Draft Preparation: MA\u003c/p\u003e\n\u003cp\u003eWriting \u0026ndash; Review \u0026amp; Editing: Kazuki O, KA, SM, Katsuhiko O\u003c/p\u003e\n\u003cp\u003eAll the authors have read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors would like to thank all those who supported this research.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eMoloney L, Rozga M, Steiber A, Handu D. The Effectiveness of Medical Nutrition Therapy in Prevention and Treatment of Chronic Disease: A Position Paper of the Academy of Nutrition and Dietetics. J Acad Nutr Diet. 2026;126(2):156219.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMoloney L, Chac\u0026oacute;n V, Devarakonda SLS, et al. Effectiveness of Medical Nutrition Therapy Provided by Registered Dietitian Nutritionists on Nutrition and Health Outcomes in Adults with Protein-Energy Malnutrition: A Systematic Review and Meta-Analysis. J Acad Nutr Diet. 2025;125(8):1144\u0026ndash;e116120.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWells JC, Sawaya AL, Wibaek R, et al. The double burden of malnutrition: aetiological pathways and consequences for health. Lancet. 2020;395(10217):75\u0026ndash;88.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePopkin BM, Corvalan C, Grummer-Strawn LM. Dynamics of the double burden of malnutrition and the changing nutrition reality. Lancet. 2020;395(10217):65\u0026ndash;74.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDent E, Wright ORL, Woo J, Hoogendijk EO. Malnutrition in older adults. Lancet. 2023;401(10380):951\u0026ndash;66.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMohr AE, Hatem C, Sikand G, et al. Effectiveness of medical nutrition therapy in the management of adult dyslipidemia: A systematic review and meta-analysis. J Clin Lipidol. 2022;16(5):547\u0026ndash;61.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDudzik JM, Senkus KE, Evert AB, et al. The effectiveness of medical nutrition therapy provided by a dietitian in adults with prediabetes: a systematic review and meta-analysis. Am J Clin Nutr. 2023;118(5):892\u0026ndash;910.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePapoutsakis C, Sundar C, Woodcock L, Abram JK, Lamers-Johnson E. Translating malnutrition care from the hospital to the community setting. Nutr Clin Pract. 2024;39(6):1292\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStratton RJ. Managing malnutrition and multimorbidity in primary care: dietary approaches to reduce treatment burden. Proc Nutr Soc. 2025;84(4):381\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchumacher TL, Jansson A, Herbert J, et al. The effectiveness of medical nutrition therapy for people at moderate to high risk of cardiovascular disease in an Australian rural primary care setting: 12-month results from a pragmatic cluster randomised controlled trial. BMC Health Serv Res. 2025;25(1):956.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMitchell LJ, Ball LE, Ross LJ, Barnes KA, Williams LT. Effectiveness of Dietetic Consultations in Primary Health Care: A Systematic Review of Randomized Controlled Trials. J Acad Nutr Diet. 2017;117(12):1941\u0026ndash;62.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChan KS, Wan EY, Chin WY, et al. Effects of continuity of care on health outcomes among patients with diabetes mellitus and/or hypertension: a systematic review. BMC Fam Pract. 2021;22(1):145.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHussey PS, Schneider EC, Rudin RS, Fox DS, Lai J, Pollack CE. Continuity and the costs of care for chronic disease. JAMA Intern Med. 2014;174(5):742\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAmerican Diabetes Association Professional Practice Committee. Facilitating positive health behaviors and well-being to improve health outcomes: Standards of care in diabetes\u0026mdash;2026. Diabetes Care. 2026;49(Suppl 1):S89\u0026ndash;131.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eColombo O, Ferretti VV, Ferraris C, et al. Is drop-out from obesity treatment a predictable and preventable event? Nutr J. 2014;13:13.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eInelmen EM, Toffanello ED, Enzi G, et al. Predictors of drop-out in overweight and obese outpatients. Int J Obes (Lond). 2005;29(1):122\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEveritt JD, Battista-Dowds EM, Heggs D, Hewlett P, Squire ALM. Determinants of completion and early dropout in an adult weight management service: a prospective observational study. J Hum Nutr Diet. 2023;36(5):1931\u0026ndash;41.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eF\u0026oslash;lling IS, Kl\u0026ouml;ckner C, Devle MT, Kulseng B. Preventing type 2 diabetes, overweight and obesity in the Norwegian primary healthcare: a longitudinal design with 60 months follow-up results and a cross-sectional design with comparison of dropouts versus completers. BMJ Open. 2022;12(3):e054841.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWendel Garcia PD, Caccioppola A, Coppola S, et al. Latent class analysis to predict intensive care outcomes in Acute Respiratory Distress Syndrome: a proposal of two pulmonary phenotypes. Crit Care. 2021;25(1):154.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGrant RW, McCloskey J, Hatfield M, et al. Use of Latent Class Analysis and k-Means Clustering to Identify Complex Patient Profiles. JAMA Netw Open. 2020;3(12):e2029068.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHutchinson JM, Raffoul A, Pepetone A et al. Advances in methods for characterising dietary patterns: a scoping review. Br J Nutr Published online March 10, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSinha P, Calfee CS, Delucchi KL. Practitioner's Guide to Latent Class Analysis: Methodological Considerations and Common Pitfalls. Crit Care Med. 2021;49(1):e63\u0026ndash;79.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGan DRY, Wister AV, Best JR. Environmental Influences on Life Satisfaction and Depressive Symptoms Among Older Adults With Multimorbidity: Path Analysis Through Loneliness in the Canadian Longitudinal Study on Aging. Gerontologist. 2022;62(6):855\u0026ndash;64.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBurnett J, Beauchamp JES. Public Health Strategies and Community Resources to Address Social Isolation and Loneliness in Older Adults. Clin Geriatr Med. 2025;41(3):369\u0026ndash;89.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCrocker TF, Ensor J, Lam N, et al. Community based complex interventions to sustain independence in older people: systematic review and network meta-analysis. BMJ. 2024;384:e077764.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKloppe T, Tetzlaff B, Mews C, Zimmermann T, Scherer M. Interprofessional collaboration to support patients with social problems in general practice-a qualitative focus group study. BMC Prim Care. 2022;23(1):169.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMurray M, Bodenheimer T, Rittenhouse D, Grumbach K. Improving timely access to primary care: case studies of the advanced access model. JAMA. 2003;289(8):1042\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMatulis JC, McCoy R. Patient-Centered Appointment Scheduling: a Call for Autonomy, Continuity, and Creativity. J Gen Intern Med. 2021;36(2):511\u0026ndash;4.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou Y, Dai X, Ni Y, et al. Interventions and management on multimorbidity: An overview of systematic reviews. Ageing Res Rev. 2023;87:101901.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYanovski SZ, Yanovski JA. Approach to Obesity Treatment in Primary Care: A Review. JAMA Intern Med. 2024;184(7):818\u0026ndash;29.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-primary-care","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"famp","sideBox":"Learn more about [BMC Primary Care](https://bmcprimcare.biomedcentral.com/)","snPcode":"","submissionUrl":"https://author-welcome.nature.com/12875","title":"BMC Primary Care","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Medical Nutrition Therapy, Registered Dietitian, Primary Care, Latent Class Analysis, Continuity, Dropout, Patient Background","lastPublishedDoi":"10.21203/rs.3.rs-8998867/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8998867/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMedical nutrition therapy (MNT) is a key component of chronic disease management in primary care and is effective in managing noncommunicable diseases (NCDs), obesity, and malnutrition; however, discontinuation is common in routine clinical practice. While previous studies have examined individual factors associated with discontinuation, MNT disengagement likely reflects heterogeneous background patterns shaped by the interplay of multiple factors. From a primary care perspective emphasising comprehensiveness and continuity, this study aimed to identify heterogeneity among patients who discontinue MNT.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis retrospective observational study was conducted at a primary care clinic in Japan. Outpatients who received dietitian-led MNT and subsequently discontinued therapy between January 2019 and December 2024 were included (n = 332). Latent class analysis (LCA) using maximum likelihood estimation was conducted on categorical variables representing demographic characteristics, clinical factors, lifestyle behaviours, and psychosocial factors. The optimal model was selected based on information criteria, classification accuracy, and clinical interpretability.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe final analytical sample comprised 300 patients. LCA identified four distinct subgroups with differing background patterns. These subgroups broadly reflected patterns involving older adults with relatively preserved self-care; working-age individuals with low clinical and lifestyle risk; individuals with high BMI and coexisting medical conditions receiving pharmacotherapy; and individuals exhibiting prominent lifestyle risk behaviours and low illness awareness.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDiscontinuation of MNT was not attributable to a single determinant but reflected diverse, multifactorial background patterns encompassing demographic, clinical, behavioural, and psychosocial characteristics. These findings provide a foundation for developing tailored strategies to support sustained engagement with MNT in primary care, grounded in comprehensive patient profiling.\u003c/p\u003e","manuscriptTitle":"Heterogeneity Among Patients Discontinuing Medical Nutrition Therapy in Primary Care: A Latent Class Analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-09 16:42:35","doi":"10.21203/rs.3.rs-8998867/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewersInvited","content":"","date":"2026-04-03T15:16:17+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-31T10:29:07+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-03-05T10:29:06+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-03T12:18:43+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Primary Care","date":"2026-03-03T08:40:17+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-primary-care","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"famp","sideBox":"Learn more about [BMC Primary Care](https://bmcprimcare.biomedcentral.com/)","snPcode":"","submissionUrl":"https://author-welcome.nature.com/12875","title":"BMC Primary Care","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"c0ee4be1-33e2-4b16-894c-eaa8a73212dc","owner":[],"postedDate":"April 9th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-09T16:42:35+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-09 16:42:35","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8998867","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8998867","identity":"rs-8998867","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}