What should we ask patients living with obesity every week? Towards a phenotype-informed framework for continuous digital patient-reported data.

Since we started building digital medical devices to improve care for people living with obesity at Boli Care, one question has persisted across dozens of clinical interviews: what should we ask patients between consultations, and how often?

The global medical community is finally acknowledging obesity for what it truly is: a chronic, relapsing, neurobehavioral and metabolic disease. In 2025, the Lancet Commission on Clinical Obesity codified this transition, defining obesity as an illness in its own right rather than a mere lifestyle risk factor.² Concurrently, second-generation anti-obesity medications (AOMs), notably semaglutide and tirzepatide, have established unequivocal efficacy, generating mean body weight reductions of 15% to 21% in randomized phase 3 trials.^8,9 Their emergence has achieved more in five years to destigmatize obesity than decades of advocacy: in the collective consciousness, where an effective medicine exists, a legitimate biological disease must be acknowledged.

Yet an alarming paradox persists: while the pathology operates continuously every hour of every day, the clinical data used to understand and evaluate it are still collected like an episodic, yearly check-up.

1. The interval nobody sees

Everything clinically decisive in obesity care happens between clinic appointments. Eating behavior and its psychological triggers (stress, emotional regulation, boredom, hedonic craving); gastrointestinal tolerability week by week; changes in sleep architecture, mood, and daily vitality. Above all, the quiet drift toward treatment discontinuation unfolds entirely during this unmonitored window.

While pivotal randomized trials (STEP-1, SURMOUNT-1) reported persistence rates exceeding 80% to 90%, real-world cohorts reveal a stark access and retention gap. In an analysis of electronic health records covering 125,474 US adults initiating dual-labeled GLP-1 receptor agonists in routine clinical practice, Rodriguez and colleagues found that within twelve months, 64.8% of patients without type 2 diabetes and 46.5% of patients with diabetes had discontinued therapy.³ Randomized withdrawal trials (STEP-1 extension, SURMOUNT-4) confirm that upon drug cessation, approximately two thirds of lost weight returns, accompanied by a rapid rebound of cardiometabolic risk factors.^10,11

This challenge is further compounded by structural capacity bottlenecks. Even prior to recent pharmaceutical breakthroughs, specialized obesity centers operated at maximum capacity. In France, for example, the reimbursement framework for Wegovy and Mounjaro in severe obesity requires initial prescription and regular coordination by specialist physicians. While clinically justified to ensure proper diagnosis and safety, this framework channels hundreds of thousands of eligible patients through an already overloaded specialist gateway, without adding a single new physician. Without continuous remote assessment tools, specialist teams face overwhelming administrative and triage burdens.

2. Why consumer apps and unguided digital tools fall short

Could consumer apps, commercial wellness platforms, or calorie counters fill this void? In practice, they cannot. Calorie counting applications, diet trackers, and step counters were never engineered for clinical rigor. Their metrics are unstandardized, disconnected from healthcare teams, rarely validated against clinical outcomes, and suffer from precipitous engagement decay (the well-documented "law of attrition").¹² Furthermore, obsessive daily caloric entry frequently aggravates psychological burden and disordered eating patterns in vulnerable individuals.

Even regulated digital therapeutics must prove their specific design hypothesis. In the recent multicenter randomized DEMETRA trial, a standalone digital behavioral therapeutic failed to achieve its primary weight-loss endpoint across the intent-to-treat population, showing benefit only in the subgroup with the highest adherence.¹³ Such findings do not negate the value of digital technologies. Rather, they demonstrate the peril of deploying software before rigorously defining what needs to be measured, when, and why. As highlighted in recent reviews by Plessis and colleagues, digital tools in obesity must transition toward validated psychological and behavioral phenotyping.¹⁴

When interviewing dozens of practicing endocrinologists and obesity specialists, a uniform reality emerged: there is currently no clinical consensus on what data should be gathered between visits. Care teams improvise with crude tools: sporadic bathroom scale logs, retrospective patient recall during hurried 15-minute consultations, and unstructured paper questionnaires.

3. Three foundational pillars waiting to be connected

Remarkably, the necessary scientific foundation already exists. High-quality scientific advances have emerged across distinct domains, yet they have remained isolated in separate silos:

• Pillar 1: The ICHOM Standard Outcome Set (What matters). In 2025, the International Consortium for Health Outcomes Measurement (ICHOM) published an international consensus outcome set for adults living with obesity (Mekonnen et al., eClinicalMedicine).⁴ Developed by 29 international experts and patient advocates across 21 countries, this set identifies 20 essential outcome measures spanning physical health, psychosocial functioning, health behaviors, and adverse events. It provides an international consensus on what to measure, but is intended for scheduled periodic assessment (every 6 to 12 months) rather than week-to-week clinical decision support.
• Pillar 2: Actionable Obesity Phenotypes (Why it happens). Pioneered by Andres Acosta, Michael Camilleri, and their Mayo Clinic colleagues, pathophysiological phenotyping has demonstrated that obesity is not a homogenous condition. In a landmark pragmatic clinical trial, tailoring anti-obesity medications to four distinct phenotypes (abnormal satiation / "hungry brain"; abnormal postprandial satiety / "hungry gut"; hedonic or emotional eating / "emotional hunger"; and reduced resting energy expenditure / "slow burn") nearly doubled 12-month total weight loss: 15.9% versus 9.0%.⁵ However, Mayo phenotyping historically relies on costly, invasive physiological tests performed once at baseline, going dark just as patient physiology shifts under treatment.
• Pillar 3: Proven Feasibility of Weekly cPROs (How to deliver at scale). The oncology community has proven that continuous, weekly patient-reported outcome (cPRO) monitoring is not only achievable but transformative. Pioneered by Ethan Basch and colleagues, weekly digital symptom reporting with automated clinical alerts significantly reduced emergency visits, enhanced quality of life, and lengthened overall survival.⁶ In their 52-community-practice PRO-TECT randomized trial, an unprecedented 91.5% of expected weekly surveys were completed over a full 52-week period by cancer patients undergoing active chemotherapy.⁷ If patients undergoing complex cytotoxic regimens can sustain weekly reporting when the interface is respectful, brief, and clinically actionable, patients living with obesity certainly can as well.

4. The proposed architecture: Minimal common core plus phenotype modules

In our open-access preprint published on Zenodo (doi:10.5281/zenodo.22283761), we describe how these pillars coalesce into an actionable clinical framework.¹ A weekly digital data collection in obesity must fulfill three simultaneous functions:

1. Understand: Characterize each patient's longitudinal behavioral phenotype as it evolves over time (appetite signals, emotional eating, sleep, vitality).
2. Evaluate: Track whether therapeutic interventions are improving patient-centered outcomes at the cadence of disease progression rather than annual retrospectives.
3. Manage: Support ongoing pharmacotherapy and lifestyle adaptation by tracking dose escalations, detecting adverse events early, and preempting silent disengagement.

Why a weekly cadence? Decisive neurobehavioral processes (craving surges, mood shifts, tolerability adaptations) oscillate over days and weeks: too dynamic for quarterly consultations, yet not requiring daily interrogation. Crucially, a 7-day recall window is the validated standard across established psychometric instruments (such as the Control of Eating Questionnaire, CoEQ, and PRO-CTCAE).^15,16 Furthermore, weekly cadence deliberately decouples measurement from drug posology: while today's leading GLP-1/GIP receptor agonists are weekly subcutaneous injections, daily oral non-peptide small molecules and monthly long-acting therapies are arriving. Tying data architecture to injection schedules would render any standard obsolete upon the next drug launch.

The 8-item minimal common core: Completed in under two minutes, the common core applies to every patient across eight standardized dimensions:

• Mealtime fullness (satiation / "hungry brain")
• Post-meal hunger recurrence (satiety / "hungry gut")
• Frequency and strength of food cravings (Control of Eating)
• Emotional or stress-induced eating ("emotional hunger")
• Subjective sleep quality and restorative rest
• General emotional state and psychological distress
• Perceived eating self-efficacy and control
• Physical energy and daily vitality

Standardized, versioned modules: Layered upon the common core, targeted modules activate based on explicit clinical rules:

• Initiation and Titration Module: Activated during the first 16 to 20 weeks of pharmacotherapy or following dose increases, capturing dose taken, injection site issues, and graded gastrointestinal symptoms using validated PRO-CTCAE items (nausea, constipation, vomiting, reflux).¹⁶
• Phenotype-Specific Exploration Modules: Deeper behavioral evaluations triggered when the core identifies acute signals (e.g., intensive cravings or severe emotional stress).
• Maintenance Phase: Once clinical stability is achieved, modules deactivate, and the collection cadence can taper, respecting patient autonomy.

Does a modular architecture compromise data comparability for clinical research? Not if disciplined engineering principles are enforced. The 8-item core remains identical across all participants. Modules are standardized, version-controlled instruments, and module assignment criteria are recorded explicitly as structured variables within the dataset. Asking every patient every question is what drives attrition; intelligent modularity ensures long-term longitudinal completion.

5. Designed for reality: Weight stigma, equity, and multi-channel delivery

Patients living with obesity are frequently mischaracterized as "difficult to engage." As documented in the landmark international consensus statement on ending weight stigma (Rubino et al., Nature Medicine 2020), individuals with obesity experience systemic bias throughout society and healthcare settings.⁷ Disengagement from digital tools or clinical appointments is not a character flaw: it is a learned protective response to stigmatizing encounters. Moreover, biological features of obesity (fatigue, executive fatigue, internalised weight bias) directly impair self-monitoring capacity.¹⁷ In our framework, a missed weekly entry is treated as a meaningful clinical signal, not as "patient non-compliance."

Furthermore, health equity cannot be treated as an afterthought. Obesity disproportionately impacts communities with lower socio-economic status, lower health literacy, and limited digital connectivity. A continuous data architecture that relies exclusively on premium smartphones and high-speed data plans exacerbates existing healthcare disparities. Our proposed implementation mandates multi-channel administration: native mobile applications, responsive web portals, interactive SMS text messaging, automated voice systems (IVR), and plain-language interfaces. Data completion metrics must be transparently reported across socioeconomic strata rather than obscured in clinical trial appendices.

6. What this framework produces, and what it does not

Clarity regarding clinical boundaries is essential for patient safety and professional trust:

• It is structured, asynchronous data collection, NOT emergency telemonitoring: Patients must be explicitly informed that weekly reports are reviewed asynchronously in conjunction with scheduled clinical consultations. The system is not an emergency response service. Acute clinical symptoms (such as signs of pancreatitis or severe dehydration) must continue through standard urgent care pathways.
• It is an empirical measurement proposal, NOT an unvalidated therapeutic intervention: Whether regular weekly reporting and clinician dashboard visualization directly improve clinical outcomes in obesity is an empirical hypothesis that this standard enables us to formally test. Completion rates, physician time efficiency, and clinical outcomes must be pre-registered and published openly.

7. An open invitation to clinicians and researchers

No single company, academic center, or clinical department can or should establish an international standard in isolation. Building a consensus-driven framework requires diverse multidisciplinary competencies: deep clinical expertise in obesity medicine, psychometric validation rigor, and the technological infrastructure to deploy compliant, secure software under medical device regulations (such as EU MDR 2017/745 Class I standards).

We invite endocrinologists, primary care physicians, clinical psychologists, patient advocates, and health economists to review the full preprint on Zenodo, challenge its hypotheses, and collaborate on validating its candidate items.

References

1. Belletante PL. What should we ask patients living with obesity every week? Towards a phenotype-informed framework for continuous digital patient-reported data in obesity care. Zenodo Preprint 2026. doi:10.5281/zenodo.22283761.
2. Rubino F, Cummings DE, Eckel RH, et al. Definition and diagnostic criteria of clinical obesity. Lancet Diabetes Endocrinol 2025;13(3):221-262. doi:10.1016/S2213-8587(24)00316-4.
3. Rodriguez PJ, Zhang V, Gratzl S, et al. Discontinuation and reinitiation of dual-labeled GLP-1 receptor agonists among US adults with overweight or obesity. JAMA Netw Open 2025;8(1):e2457349. doi:10.1001/jamanetworkopen.2024.57349. PMID 39888616.
4. Mekonnen T, Staniford L, Connell S, et al. Development of a core patient-centred outcome set for adults living with obesity: a modified Delphi-based international consensus. eClinicalMedicine 2025;87:103422. doi:10.1016/j.eclinm.2025.103422. PMID 39803154.
5. Acosta A, Camilleri M, Abu Dayyeh B, et al. Selection of antiobesity medications based on phenotypes enhances weight loss: a pragmatic trial in an obesity clinic. Obesity (Silver Spring) 2021;29(4):662-671. doi:10.1002/oby.23120. PMID 33759389.
6. Basch E, Deal AM, Kris MG, et al. Symptom monitoring with patient-reported outcomes during routine cancer treatment: a randomized controlled trial. J Clin Oncol 2016;34(6):557-565. doi:10.1200/JCO.2015.63.0830. PMID 26644531.
7. Rubino F, Puhl RM, Cummings DE, et al. Joint international consensus statement for ending stigma of obesity. Nat Med 2020;26(4):485-497. doi:10.1038/s41591-020-0803-x. PMID 32132690.
8. Wilding JPH, Batterham RL, Calanna S, et al. Once-weekly semaglutide in adults with overweight or obesity. N Engl J Med 2021;384(11):989-1002. doi:10.1056/NEJMoa2032183. PMID 33567185.
9. Jastreboff AM, Aronne LJ, Ahmad NN, et al. Tirzepatide once weekly for the treatment of obesity. N Engl J Med 2022;387(3):205-216. doi:10.1056/NEJMoa2206038. PMID 35658024.
10. Wilding JPH, Batterham RL, Davies M, et al. Weight regain and cardiometabolic effects after withdrawal of semaglutide: the STEP 1 trial extension. Diabetes Obes Metab 2022;24(8):1553-1564. doi:10.1111/dom.14725. PMID 35441470.
11. Aronne LJ, Sattar N, Horn DB, et al. Continued treatment with tirzepatide for maintenance of weight reduction in adults with obesity: the SURMOUNT-4 randomized clinical trial. JAMA 2024;331(1):38-48. doi:10.1001/jama.2023.24945. PMID 38079144.
12. Eysenbach G. The law of attrition. J Med Internet Res 2005;7(1):e11. doi:10.2196/jmir.7.1.e11.
13. Bertoli S, Capodaglio P, Colosimo S, et al. Effectiveness of a digital therapy on 6-month weight loss in people with obesity: the DEMETRA randomized clinical trial. J Med Internet Res 2025;27:e72054. doi:10.2196/72054.
14. Plessis C, Galmiche M, Déchelotte P, Haro A. Review of the use of digital therapeutics for obesity treatment: toward a psychological phenotyping. Adv Nutr 2026;17(4):100613. doi:10.1016/j.advnut.2026.100613.
15. Dalton M, Finlayson G, Hill A, Blundell J. Preliminary validation and principal components analysis of the Control of Eating Questionnaire (CoEQ) for the experience of food craving. Eur J Clin Nutr 2015;69(12):1313-1317. doi:10.1038/ejcn.2015.57. PMID 25920704.
16. Dueck AC, Mendoza TR, Mitchell SA, et al. Validity and reliability of the US National Cancer Institute's Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE). JAMA Oncol 2015;1(8):1051-1059. doi:10.1001/jamaoncol.2015.2639. PMID 26291664.
17. Durso LE, Latner JD. Understanding self-directed stigma: development of the weight bias internalization scale. Obesity (Silver Spring) 2008;16 Suppl 2:S80-S86. doi:10.1038/oby.2008.448. PMID 18978768.