The use of an automatic remote weight management system to track treatment response,identified drugs supply shortage and its consequences: A pilot study

Abstract

Objective

The objective of this pilot study is to evaluate the feasibility of using an automatic weight management system to follow patients’ response to weight reduction medications and to identify early deviations from weight trajectories.

Methods

The pilot study involved 11 participants using Semaglutide for weight management, monitored over a 12-month period. A cloud-based, Wi-Fi-enabled remote weight management system collected and analyzed daily weight data from smart scales. The system's performance was evaluated during a period marked by a Semaglutide supply shortage.

Results

Participants achieved a cumulative weight loss of 85 kg until a supply shortage-induced trough in October 2022. This was followed by a 6–8 week plateau and a subsequent 13 kg cumulative weight gain. The study demonstrated the feasibility of digitally monitoring weight without attrition over 12 months and highlighted the impact of anti-obesity drug (AOD) supply constraints on weight trajectories.

Conclusions

The remote weight management system proved important for improving clinic efficacy and identifying trends impacting obesity outcomes through electronic data monitoring. The system's potential in increasing medication compliance and enhancing overall clinical outcomes warrants further research, particularly in light of the challenges posed by AOD supply fluctuations.

Introduction

The growing epidemic of obesity, worldwide, prompted the introduction of novel interventions to counteract the protracted inefficacies associated with lifestyle modifications and the restricted accessibility and complications linked to bariatric surgical methodologies. ¹ Over the past decade, medical interventions faced challenges, encompassing limited efficacy (e.g., Orlistat), post-approval recalls due to severe side effects (Lorcaserin, Sibutramine, Rimonabant) and the emergence of long-term valvular diseases (fenfluramine, dexfenfluramine), among other barriers. ²

However, the landscape has changed dramatically subsequent to the 2014 FDA approval of Liraglutide and the 2014 approval of semaglutide 2.4 mg ³ for the treatment of adult obesity. The widespread adoption and endorsement of incretin-based medications (GLP-1 agonists) in type 2 diabetes, coupled with their clinical efficacy in glycemic control, weight reduction and mitigating renal and cardiac complications, spurred an upsurge in prescriptions for these agents.

This increased utilization of new classes of anti-obesity drugs (AODs) has introduced a new hurdle to achieving and sustaining weight reduction, namely, adherence and compliance with prescribed medications. ⁴ A recent retrospective cohort study involving 1911 patients revealed that only 44% exhibited persistence with AOD at 3 months, diminishing to 33% at 6 months and a mere 19% at 1 year. ⁵ Persistence rates at 1 year were 40% for semaglutide, 17% for liraglutide, 13% for phentermine–topiramate and 10% for naltrexone–bupropion, and no patients were utilizing orlistat at 12 months. ⁵

The challenge of adhering to AOD is multifaceted and aligns with various dimensions elucidated by the World Health Organization's review on adherence to long-term therapies. ⁶ These dimensions encompass social/economic, provider–patient/health care system, condition-related, therapy-related and patient-related factors. Socially, obesity is often construed as a lifestyle-related disorder, potentially diminishing the perceived necessity for long-term medications compared to medications for other chronic conditions. The use of injections poses another impediment, though the current availability of once-weekly dosing and oral preparations is anticipated to alleviate some of the burden associated with daily injections. Beyond economic challenges and side effects, primarily gastrointestinal in nature, a fundamental misconception regarding the pharmacological effect of AODs further hinder adherence. The weight reduction trajectory following drug initiation reaches a plateau at approximately 4–6 weeks, ⁷ potentially leading inadequately informed patients to perceive this medication failure (‘stopped working for me’) or as an indication that sustained weight reduction can be achieved without the drug. However, discontinuation of AODs is associated with weight regain. For instance, patients discontinuing semaglutide regained two-thirds of their lost weight and concurrently forfeited cardiometabolic improvement one-year post discontinuation, ⁸ Comparable discontinuation rates were observed by Weis et al. ⁴ reporting that at 12 and 24 months, 64.5% and 59.2% were adherent to, while 45.2% and 64.7% discontinued, GLP-1 agonist therapy in type 2 diabetes, respectively.

The economic futility associated with AOD discontinuation spurred the development of an Obesity clinic tool aimed at addressing the growing number of patients seeking weight management through AOD therapy. The tool seeks to enhance outcomes by continuous track of trends in weight reduction.

The primary objective, addressing clinic capacity constraints by enhancing efficiency, is grounded in a telemedical approach featuring automatic triage and dashboarding of patients.

Weight measurement is seamlessly uploaded to a data management system generating a dashboard of the patients according to heuristics embedded in the program. Thus, the clinic's personnel homepage displays a simple pie chart illustrating the status of the clinic's patients, dividing them into three groups: on-track, borderline cases and out-of-bounds patients. The size of each slice is adjusted to reflect the proportion of patients in each group. By zooming in, clinic personnel can easily focus on a specific group, retrieve a list of its members along with their contact details and access a view of the patient's weight chart as well as their complete data. Hence, the system enables to speed up the identification of individuals facing challenges and in need of outreach, either through SMS with actionable recommendations or, if necessary, via a physical call. The selective scheduling of appointments and contacts based on clinical need, rather than arbitrary fixed periods, constitutes fundamental aspects of value-based medicine. Moreover, this approach may enhance patient satisfaction by reducing unnecessary visits.

From a user perspective, the system provides for surveillance and early detection of deviation from the trajectory of weight, at early stages. The underlying principle is to minimize user interaction with the system while maximizing each interaction's impact by providing actionable advice and integrating outcomes back into the system.

We had the opportunity to ‘pressure test’ the system during the supply shortage of semaglutide, ignited by social media coverage of the off-label use for weight reduction of GLP-1 agonists ⁹ causing worldwide shortage of semaglutide and compelling the manufacturer to issue statements on the limited Ozempic (semaglutide) stocks that affect the supply. ¹⁰

The study follows the weight changes from initiating semaglutide therapy for weight loss and maintenance and during the period of the drug supply shortage.

Methods

Population

Endocrine clinic's overweight and obese patients. The convenience sample consisted of participants who registered for the weight maintenance program, initiated semaglutide for weight reduction and consented to the use of their data for research purposes prior to the data analysis. Ethical approval for this analysis was obtained from the Ethics Committee of the Information Systems Faculty at the Academic College of Tel Aviv-Yaffo (Approval number 20246).

Outcomes

Weight change in kg from initiation in January 2022 till October 2022 and weight change from October to end of follow-up February 2023.

Remote weight management system

The system consists of four parts, the system's application, its database, the clinic's computers and customers’ smart scales.

The application has been built with Python and operates from the cloud. It can be accessed via desktops and smartphones. Data transfer between smart scales and the application relies on personal WiFi networks connected to the Internet. New smart scales can easily be added to the system, by following two simple steps. First, the physician creates a record for the new customer, which includes basic personal data and a unique identifier, and provides the customer with a smart scale and a one-pager first-time setup. Then, the customer unboxes this device, follows the one-pager instructions and concludes by connecting the device to his home network. The system's database is a relational database and was implemented with MySQL, an open-source database management system, also operates from a cloud.

The clinic's computers connect with the system for two main reasons. Initially, when new customers join, they are added by the physician to the systems, and then, regularly monitored by the clinic's dietician. The personnel interact with the system via their computer browsers.

Customers’ weight is being measured with battery-operated smart scales, connected to the system via its built-in Wi-Fi technology. The measurement recording process is simple and straightforward, performed whenever the customer steps on the smart scale.

The information is automatically analysed with no additional personnel, and the patients are triaged to those on track versus those deviating, who initially receive automatic actionable alerts via the mobile system. In case the deviation persists, and there are no signs of improvement, the user receives a call or an appointment to the clinic. All this is based on heuristic-based algorithm Figure 1.

Figure 1.

System's architecture.

This pilot study evaluates the feasibility of an automatic remote weight management system, to support weight-tracking activities, without any changes in the usual care. The pilot study's participants were a subset of the endocrine clinic, who opted-in to the digital weight maintenance program, purchased a smart scale, to track and record their weight and consented to use their data for clinical studies. The follow-up ranged between 14 August 2022 and 7 March 2023. Participants were asked to weigh themselves every morning at fasting, just after voiding. The transmitted data was stored on the research server in kilos, with a stated accuracy of 45 grams. Participants’ demographic was collected at the clinic's visit at baseline, according to routine clinic practice via a registration form. In the follow-up of the patients, there was no difference between the usual care of overweight and obese patients and those who participated in the program. The aim was to evaluate the feasibility of an automatic remote weight management system, to support weight tracking activities.

Statistical analyses

Analyses were conducted using Microsoft Excel for Microsoft 365, the enterprise version for Windows. Weight data were prepared for analyses by removal of duplicate entries and outliers. Whenever multiple weights occurred within a single day, the first (earliest) weight was retained while the others were removed. Whenever outliers occurred, implicitly indicating usage by a different person, the outlier data was removed accordingly. Missing observations were filled linearly, thus, enabling to display of an aggregated daily change.

Results

The pilot study included 11 participants, six males and five females. The average age was 48.3 ± 15.4 years (range 19–73), and the average weight at the onboarding stage was 98.3 ± 15.6 kg. Of the 11 participants, two males and one female were on weight maintenance therapy, while the others received treatment for weight reduction. Among the participants, nine graduated from college or university, one graduated from secondary school and one holds a PhD. Missing data in weight reduction studies is a common challenge. However, with the automatic remote weight management system, daily weight measurements were obtained for 54.05% of users. For those who missed some daily measurements, there were still enough data points overall that imputation of missing data was not necessary. The cumulative weight loss from the onboarding stage to the trough recorded point in November 2022 was 85 kg, coinciding with the period where the shortage of semaglutide was apparent. Following the trough, there was a 5–6 week plateau, followed by a 13 kg (cumulative) weight gain (Figure 2 and Figure 3).

Figure 2.

Cumulative weight change from baseline.

Figure 3.

Median and quartile change in weight.

The average proportional weight loss from the onboarding measure to the rough measure was 7.72% (SD = .048), with the lowest proportional weight loss of 0% and the highest proportional weight loss of 16.68%. The average proportional weight loss from the onboarding measure to the end of the follow-up was 4.49% (SD = .056), with the lowest proportional weight loss of −1.31% and the highest proportional weight loss of 15.06%.

Discussion

The app, initially targeted to assess the ability of remote weight management system to follow up the weight trajectories of participants, provides two important insights. Firstly, it addresses the feasibility of digitally monitoring weight and dashboarding results, showing no attrition during up to 12 months of follow-up. Secondly, it explores the impact of constraints on the supply of AOD.

By implementing a cloud-based, Wi-Fi-enabled remote weight management system, as described in the method section, the clinic received real-time updates on patients’ weight trajectories. Notably, this system's uniqueness lies in the additional data being collected without imposing extra burdens on both users and the health management team. Users are only required to perform daily weighing on a seamlessly connected scale, without any direct interaction with the application or connectivity challenges. Conversely, the health management team benefits from reduced unnecessary clinic visits and receives actionable recommendations only when necessary. The clinic data accumulates automatically, ‘dash boarded’ and notifications are triggered only in case of persistent and significant deviations from the expected values. This mechanism enhances clinic efficiency, increases capacity and potentially contributes to medication adherence, leading to improved outcomes. ¹¹

The concept of applying technological solutions and behavioural psychology to address practical health management and well-being challenges, with minimal effort by providing immediate feedback for successfully supervised self-management in Obesity was described 23 years ago by Parkka et al. ¹² This pioneering work predated the introduction of new medication classes and therefore mainly emphasizing behaviour modification. Contemporary approaches leverage technological advancements and devices of the past two decades to both, monitor medication adherence while maintaining and complementing the behavioural approach.

The second insight emerges from the early observation, three weeks into the drug shortage, demonstrating a plateau effect in weight reduction followed by subsequent weight gain as the shortage persisted. The assumption is that early patient awareness and contact allowed for mitigation measures, such as alternative medications, minimizing the negative effects of drug supply restrictions. Despite the unfortunate circumstances, this incident served as a stress test for the system, revealing its sensitivity to change and clinical value. Currently, if such shortage exists in one drug, alternative solution can be provided.

It is important to note that this observation is based on a small cohort, and its applicability to a larger cohort requires evaluation.

In summary, electronic data monitoring of chronic conditions, such as obesity, is important for increasing clinic efficacy and identifying trends that may impact outcomes. Whether these measures also increase compliance with medications and clinical efficacy and safety is to be assessed in larger randomized trials.

An automatic remote weight management system must prioritize ethical principles such as privacy, data security, informed consent and patient autonomy. Protecting patient privacy is essential, as the system will collect sensitive health data continuously. Compliance with regulations like HIPAA (Health Insurance Portability and Accountability Act) in the United States, or GDPR (General Data Protection Regulation) in the EU, is necessary to secure this data and control who accesses it. Informed consent is also critical; patients should fully understand how their data will be used, including any third-party involvement or potential use for research and development. Additionally, safeguards must be implemented to address cybersecurity risks, as the system's reliance on digital communication could make it vulnerable to unauthorized access. Addressing these ethical issues is crucial to building trust and ensuring that the system supports, rather than compromises, patient health and privacy. ¹³

However, the present study has some limitations. The small sample size of 11 participants is not a representative cohort and limits the generalizability of the findings. Additionally, there is a potential for selection bias due to the voluntary nature of participation in the program. Individuals who opted in may be more motivated or health-conscious than the general population, potentially influencing the study results. Furthermore, the study's observational nature means that causality cannot be established. The impact of external factors, such as the semaglutide supply shortage, also complicates the interpretation of the results. This pilot study does not address the association between the automatic remote weight management system and adherence. A formal RCT should be performed to address the association and selection bias discussed limitations. Future studies should address these limitations by including larger, more diverse populations and employing randomized controlled trial designs.

Conclusion

The implementation of an automatic remote weight management system has demonstrated its potential in enhancing clinic efficiency and identifying trends that impact obesity outcomes. The system's ability to provide real-time updates and actionable recommendations without imposing additional burdens on users or healthcare teams is a significant advantage. Further research is needed to evaluate the system's role in improving medication compliance and overall clinical outcomes.