Approximately 10% to 13% of women of reproductive age worldwide have polycystic ovary syndrome (PCOS) [1]. The prevalence of PCOS is increasing globally because of changes in environmental variables and lifestyle, as well as increased disease awareness and detection rates. PCOS is characterized by a range of heterogeneous symptoms with varying severities, including abnormal menstruation, hyperandrogenism (hirsutism or hyperandrogenemia), and ovarian dysfunction (oligo-ovulation or polycystic ovary morphology) [2]. One of the most prevalent endocrine disorders among patients with PCOS is insulin resistance, with variable degrees of insulin resistance present in approximately 75% of patients with PCOS [3]. For such patients, insulin resistance reduction is essential in the management of various endocrine and metabolic disorders [4].

The first-line treatment for PCOS is lifestyle intervention, and studies have demonstrated that positive lifestyle changes can improve hyperandrogenemia, menstrual cycle, hirsutism, and insulin resistance in the affected patients [5]. However, lifestyle intervention is associated with poor compliance and low sustainability. Nevertheless, there are no indicators to determine whether patients are making enough progress in improving their lifestyle. Studies have demonstrated that digital therapy accessed on mobile phones, which are convenient and accessible, can effectively improve behavioral changes in diet, exercise, and medication adherence while predicting disease progression, reducing the frequency of disease-related symptoms, and promoting effective disease management [6,7]. Regarding PCOS, which is a chronic disease, there are limited digital treatment methods. A previous 3-component lifestyle intervention consisting of diet, exercise, and cognitive behavioral therapy, facilitated using texting to track patients’ food intake, physical activity, and mood, can effectively improve eating disorder and mood in patients with PCOS and contribute to weight loss [8,9]. This study focused on the effect of digital intervention compared with metformin in improving homeostatic model assessment for insulin resistance (HOMA-IR) levels and other metabolic and reproductive indicators.

This was a single-center, prospective, randomized controlled clinical trial designed to determine whether digital intervention was effective in patients with PCOS after 12 weeks of treatment. Evaluations were conducted on improvements in PCOS-related clinical parameters and side effects for both groups. Simultaneously, a patient satisfaction survey was conducted for the digital intervention. This study aimed to confirm the efficacy and safety of digital intervention in the treatment of PCOS.

A total of 80 women with PCOS (including 73 newly diagnosed and 7 previously diagnosed women with PCOS) from the endocrinology clinic of a tertiary affiliated hospital of Tongji University were enrolled between June 21, 2022, and August 12, 2023, in this single-center, unblinded, 2-arm parallel study with 1:1 allocation. PCOS was diagnosed in patients attending PCOS outpatient clinics, and eligible patients were invited to participate in this study. The CONSORT eHEALTH (Consolidated Standards of Reporting Trials of Electronic and Mobile Health Applications and Online Telehealth) checklist is presented in Multimedia Appendix 1.

The inclusion criteria were being aged 18 to 45 years, having a clinical diagnosis of PCOS according to the 2003 Rotterdam diagnostic criteria [10], and having a HOMA-IR score ≥1.8 (in Asian countries, a HOMA-IR score ≥1.8 can diagnose insulin resistance and predict the occurrence of metabolic syndrome [11]). The exclusion criteria included the inability to use WeChat; hyperthyroidism or hypothyroidism; severe abnormal liver function (liver enzyme levels >3 times the standard limit); severe abnormal renal function (serum creatinine level >123.8 μmol/L or estimated glomerular filtration rate <45 mL/min/1.73 m²); congenital adrenal hyperplasia, hyperprolactinemia, or adrenal tumor; severe infection, severe anemia, neutropenia, and other blood system diseases; type 1 diabetes, monogenic mutation diabetes, or diabetes due to pancreatic damage or other secondary diabetes; existence of mental illness or dementia; history of using contraceptives, metformin, glucagon-like peptide-1 analogs, pioglitazone, all classes of antidepressant medications, and other drugs in the past 3 months; pregnancy or intention to become pregnant within 6 months; participation in another clinical trial within 3 months; and declining to participate. Moreover, perimenopausal or menopausal women were excluded from this study. As this last exclusion criterion was not reported in the ClinicalTrials.gov registration (NCT05386706), this exclusion was a protocol deviation.

The study protocol was approved by the ethics committee of the Shanghai Tenth People’s Hospital (SHSY-IEC-5.0/21K191/P02); the study was registered in ClinicalTrials.gov (NCT05386706); and the name of the trial registry is Digital Intervention PCOS. No changes were made to this clinical trial after registration. All participants provided a signed written informed consent. Every participant was informed that their participation in the study was completely voluntary, that the information gathered for the study would be kept private, and that they could revoke their agreement at any moment without it having an adverse effect on their care. Each participant was given a unique study ID for data entry, maintenance, and analysis, and their personal information was anonymized.

Following baseline evaluation, based on a predefined, computer-generated number in a 1:1 assignment, the participants were randomly assigned to receive either digital intervention or metformin. Allocation concealment was achieved by using opaque, sealed envelopes with sequential numbers. One researcher created the random number sequence for this investigation, and another researcher enrolled and assigned participants. During intervention, both groups were aware of their allocation. After intervention, the percentage of adherence to the intervention was calculated as adherence days divided by intervention days.

The PCOS guidelines recommend lifestyle interventions as the first-line treatment for PCOS (diet, exercise, sleep, and mood) [12,13]. Cognitive behavioral therapy is a type of treatment that combines behavioral and cognitive methods [14]. Prior studies evaluated the effectiveness of 20 cognitive behavioral therapy lifestyle courses with physical therapy, healthy eating, and 9 months of extra feedback via a mobile SMS text messaging service to conventional treatment [8]. In this regard, we designed a digital intervention method based on the WeChat mini program. The mini program was pretested with 10 users for usability and face validity before being used in the intervention. The WeChat mini program was designed by the Endocrinology and Metabolism Department of Shanghai Tenth People’s Hospital and developed by Shanghai Zoey Information Technology Co, Ltd. Patients with PCOS assigned to the digital intervention group can register a new account for free through WeChat, and they need to use the WeChat mini program for 12 weeks. As presented in Figure 1, the program consists of 3 modules. In module 1, patients can watch 2 videos per week, with each video being about 5 minutes long; a total of 24 videos can be watched over a 12-week period. These videos describe the 4 elements of a healthy lifestyle that should be prioritized by patients with PCOS: healthy diet, regular exercise, good sleep, and stress reduction. The video specifically explains how the 4 elements affect the occurrence and development of PCOS, what can be effectively done to achieve a good lifestyle, and how to adhere to these lifestyles. Multimedia Appendix 2 presents the title and main content of each video. In module 2, patients need to record their menstruation during their menstrual period (duration of menstruation, amount of bleeding, body temperature, and mood) and record their daily weight (weight change value, BMI, and weight status appearing after weight submission); daily exercise time (according to different exercise intensities, low-, moderate-, and high-intensity exercises were recorded, and the type of exercise needed to be selected); daily sleep time (bedtime and wake-up time); and daily meditation time (also provides 24 pieces of meditation music). Simultaneously, the list of recommended food and food to avoid is provided in the food recommendation part. After recording, the patients can see a statistical chart of their weight and exercise in module 2. To allow patients to sufficiently understand the important role of lifestyle improvement in PCOS treatment, we have set up module 3, which they can selectively read by themselves, including an overall introduction to PCOS; healthy lifestyles suitable for PCOS; different dietary patterns (including low-carbohydrate, high-protein, and restricted energy-balanced diets); choices of different exercises (the intensity of each exercise and the amount of energy that can be consumed); and importance of sleep hygiene (the dangers of circadian disruption). Meanwhile, contact between the patient and doctor through WeChat’s electronic message was synchronized, and communication was scheduled every Monday. The doctor acts as the patient’s coach, downloading the patient’s record data every week; analyzing the patient’s performance that week; summarizing the weekly exercise time, weight, sleep, and menstrual period and sending it to the patient; and suggesting possible improvements. In addition, the doctor asks the patients if they encountered problems that cannot be solved and provides support and encouragement. While using the digital intervention, 5 patients experienced temporary system problems (4 patients had video playback issues and 1 patient had data entry issues). After communicating with the programmer, we optimized and improved the system in time and solved the problems encountered by the patients, but the content has not changed.

Metformin, which improves ovulation and enhances insulin sensitivity, is the first-line insulin-sensitizing medication used to treat PCOS; it was used as a treatment drug in the control group in this study [15]. Metformin should be used for at least 3 months after completion depending on the patient’s ovulation recovery, androgen status, and metabolic clinical indicators [16,17]. During the 12-week intervention, the metformin group was administered 1000 mg of metformin (Glucophage, Merck KGaA) per day. In previous clinical trials, the dose of metformin ranged from 850 mg to 2000 mg per day [18]. To reduce the side effects of drugs, the dose of 1000 mg per day was selected for this study, and a doctor educated the patients regarding healthy lifestyles during the first visit.

The process of recruiting ran from June 2022 to May 2023, and the follow-up assessments were completed in August 2023. Of the 137 outpatients who had PCOS from June 2022 to August 2023, a total of 80 met the admission criteria and volunteered to participate in the study (Figure 2). They were randomly divided into the digital intervention group (40/80, 50%) and metformin group (40/80, 50%). The dropout rates were 13% (5/40) and 8% (3/40) for the digital intervention and metformin groups, respectively. The digital intervention group used a WeChat mini program for 12 weeks, whereas the metformin group received daily doses of 1000 mg of metformin also for 12 weeks. As shown in Multimedia Appendix 4, the adherence rate of the digital intervention group was 89% (95% CI 83%-95%), which was calculated by the participants’ recorded data on the WeChat mini program. On the basis of the number of days participants took metformin within 12 weeks, the adherence rate of the metformin group was 98% (95% CI 96%-99%). The adherence rate was higher in the metformin group than in the digital intervention group (P=.003).

Table 1 presents the clinical characteristics measured at baseline. No differences were observed between the two groups in terms of age, body measurements, menstrual cycles, sex hormones, glucose and lipid metabolism, uric acid, body fat distribution, hepatic steatosis, and hepatic fibrosis levels. However, the digital intervention group had a higher mean AST of 25.66 U/L (95% CI 20.78-30.53) compared with the metformin group, which had a mean AST of 19.23 U/L (95% CI 17.00-21.45). Furthermore, there were no appreciable variations in terms of sleep quality, anxiety, depression, eating habits, weekly physical activity, and daily sitting time between the two groups.

Our results provide evidence that WeChat-based digital therapies significantly improve self-management and the metabolic and reproductive aspects of the disease among patients with PCOS. While the WeChat-based digital intervention was not statistically different from metformin treatment in terms of lowering HOMA-IR levels, it was more effective in reducing waist circumference, waist-to-hip ratio, total fat mass, and DHEAS. In terms of safety, the main adverse events in the digital intervention group were not as high as those in the metformin group. With a randomized controlled trial design and a 12-week follow-up, the study’s findings offer the first convincing proof of the efficacy and safety of self-supervised, WeChat-based digital intervention to improve PCOS.

Using gold standard procedures, it was estimated that approximately 75% of patients with PCOS had insulin resistance and that the endocrine and reproductive characteristics of PCOS can be brought on by insulin resistance [27]. In clinical studies, HOMA-IR is frequently used as a surrogate marker to detect insulin resistance. Multiple studies have shown that alleviating insulin resistance is closely associated with improving menstrual cycles, ovulation, and in vitro fertilization results in PCOS [28-30]. Consequently, HOMA-IR was selected in this study as the major end point as it plays a significant role in the onset and progression of PCOS. This study demonstrated that, in patients with PCOS with insulin resistance, HOMA-IR can be significantly decreased by digital intervention. Prior research has shown that when several therapy methods (such as metformin, rosiglitazone, and acupuncture) were given for approximately 12 weeks, patients with PCOS experienced improvements in their hyperandrogenemia, weight loss, and metabolic and biochemical parameters [17,25,26].

Furthermore, the study showed that, in the same population, both digital intervention and metformin therapy reduced body weight, BMI, waist circumference, hip circumference, and waist-to-hip ratio; however, the former was associated with a greater reduction in waist circumference and waist-to-hip ratio levels than the latter. Similarly, previous research showed that metformin did not significantly reduce waist circumference as much as lifestyle change did [31,32]. In a recent study, anthropometric indicators, including body weight, waist circumference, and fat mass, significantly improved in the web-based lifestyle intervention group compared with the control group [33]. These results may be attributed to an increase in patients’ physical activities, as only the digital intervention group experienced an increase in weekly total physical activity after the intervention in this study.

Furthermore, the menstrual frequency, HbA_1c, fasting blood glucose, and fasting insulin of both groups significantly improved; however, there were no significant differences between the groups. Consistent with our results, no significant differences were observed in menstrual patterns between the groups in a previous randomized controlled study of metformin versus lifestyle changes [32]. Another study demonstrated that metformin therapy improved anthropometric parameters, menstrual cycles, and glucose metabolism [34]. Alternatively, the digital intervention significantly reduced the HbA_1c levels in patients with prediabetes and diabetes [35,36]. In a study of digital intervention in patients with obesity, the intervention group exhibited significant improvements in HbA_1c, fasting blood glucose, and HOMA-IR levels while retaining muscle mass and reducing body fat and obtained equivalent outcomes to an intensive obesity management program [37].

The digital intervention group exhibited improved liver function (ALT and AST), liver steatosis, and liver fibrosis but did not show a statistically significant difference from the metformin group. Previous studies have shown that nonalcoholic fatty liver disease and liver function in patients with PCOS could be improved by diet control and exercise [38], and patients in the digital intervention group probably changed their diet and exercise routines because of these benefits in our study. In this study, the creatinine levels also decreased in the metformin group but not in the digital intervention group. Consistently, metformin decreased the serum creatinine levels and renal oxidative damage and fibrosis in a study on diabetic mice [39].

In terms of sex hormones, LH was elevated in the digital intervention group compared with the metformin group, but there was no difference in the LH levels before and after treatment in either group. This may be because not all reproductive hormones in this study were measured during menstruation. As the DHEAS levels increased after metformin treatment, the DHEAS levels decreased in the comparison of differences between the groups (digital intervention–metformin). Metformin therapy has been previously reported to increase DHEAS levels in women with infertility due to PCOS [40]. As metformin can improve insulin sensitivity and adrenal cells have insulin receptors [41], insulin may have an impact on peripheral or adrenal tissue DHEAS metabolism. Meanwhile, the free testosterone levels decreased and the SHBG levels increased in the metformin group. In previous studies, both insulin resistance status improvement and circulating insulin reduction have been shown to increase serum SHBG levels, thereby reducing the blood level of free testosterone in patients with PCOS [34,42].

Because central obesity is frequently associated with patients with PCOS [43], we examined the distribution of body fat in these patients. The percentage of total body fat, percentage of total body lean mass, total fat mass, subcutaneous adipose tissue mass, and visceral adipose tissue mass were improved significantly in both groups, but the decline of the total fat mass was more pronounced in the digital intervention group. This may be explained by the metformin group receiving oral lifestyle guidance. Similar to our results, a meta-analysis suggests that lifestyle interventions are the best way to improve body composition and cardiorespiratory health in women with PCOS [44].

Using the questionnaire, we determined changes in the patients’ sleep quality, mood (anxiety and depression), eating behaviors, and exercise condition. Both groups exhibited significantly improved depression and eating behaviors. The digital intervention improved patients’ anxiety and level of physical activity. In a recent study of digital intervention for other diseases, it effectively improved physical activity, depression, anxiety, and eating behavior [45-47]. Metformin, alternatively, significantly improved sleep quality in our study. Studies have demonstrated that in people with type 2 diabetes, metformin therapy improved sleep quality [48]. In addition, it reduced social stress–related anxiety-like behaviors [49].

In terms of safety, the main adverse events in the digital intervention groups were sensations of hunger (5%) and dizziness (3%). The main adverse events in the metformin groups were gastrointestinal treatment–emergent adverse events (30%). Overall, the digital intervention was found to have fewer side effects, which is advantageous. In the satisfaction questionnaire, most of the patients who completed the follow-up of the digital intervention were satisfied with this treatment and would recommend it to other patients with PCOS. On the basis of the questionnaire’s results, the patients believed that the digital intervention could provide a more thorough understanding of PCOS, solve PCOS-related issues, and save more time. This suggests that the digital intervention is feasible and beneficial while saving doctors’ time and being beneficial for more patients with PCOS. Regarding adherence, while the digital intervention helped patients follow through on lifestyle modifications more successfully, metformin adherence outperformed the digital intervention.

In the digital intervention group, 81% of patients felt that the coaching was most effective. In previous studies, for women with overweight or obesity and PCOS, lifestyle therapies with added SMS text messaging assistance appear to aid in enhancing physical fitness, activity, and aerobic capacity as well as minimize sedentary behavior. It shows that supervision and communication play an important role in maintaining a healthy lifestyle.

However, this study has several limitations. First, the selected population had insulin-resistant PCOS and thus does not represent the entire population. Second, sex hormone indicators were not monitored during menstruation, which affected the judgment of sex hormone results. Third, the participants in this study were not blind, and multiple evaluations of the results increased the chances of type 1 errors. Fourth, no formal evaluation of the cost of the intervention was conducted. Finally, there is a lack of long-term follow-up data to assess whether these changes will persist over time. Thus, multicenter, large-sample studies are warranted to further elucidate the efficacy of digital intervention.

This study summarizes and confirms the feasibility and application of digital intervention for women with PCOS. The results of the study indicated that the intervention is promising in improving self-management behavior and the metabolic and reproductive features of PCOS. Furthermore, in terms of reducing waist circumference, waist-to-hip ratio, total fat mass, and DHEAS, the digital intervention was more beneficial than metformin. Regarding safety, there were fewer adverse events in the digital intervention group than the metformin group.