No abstract

No abstract

Background Intermittent fasting (IF) is an increasingly popular approach to dietary control that focuses on the timing of eating rather than the quantity and content of caloric intake. IF practitioners typically seek to improve their weight and other health factors. Millions of practitioners have turned to purpose-built mobile apps to help them track and adhere to their fasts and monitor changes in their weight and other biometrics. Objective This study aimed to quantify user retention, fasting patterns, and weight loss by users of 2 IF mobile apps. We also sought to describe and model starting BMI, amount of fasting, frequency of weight tracking, and other demographics as correlates of retention and weight change. Methods We assembled height, weight, fasting, and demographic data of adult users (ages 18-100 years) of the LIFE Fasting Tracker and LIFE Extend apps from 2018 to 2020. Retention for up to 52 weeks was quantified based on recorded fasts and correlated with user demographics. Users who provided height and at least 2 readings of weight and whose first fast and weight records were contemporaneous were included in the weight loss analysis. Fasting was quantified as extended fasting hours (EFH; hours beyond 12 in a fast) averaged per day (EFH per day). Retention was modeled using a Cox proportional hazards regression. Weight loss was analyzed using linear regression. Results A total of 792,692 users were followed for retention based on 26 million recorded fasts. Of these, 132,775 (16.7%) users were retained at 13 weeks, 54,881 (6.9%) at 26 weeks, and 16,478 (2.1%) at 52 weeks, allowing 4 consecutive weeks of inactivity. The survival analysis using Cox regression indicated that retention was positively associated with age and exercise and negatively associated with stress and smoking. Weight loss in the qualifying cohort (n=161,346) was strongly correlated with starting BMI and EFH per day, which displayed a positive interaction. Users with a BMI ≥40 kg/m2 lost 13.9% of their starting weight by 52 weeks versus a slight weight gain on average for users with starting BMI <23 kg/m2. EFH per day was an approximately linear predictor of weight loss. By week 26, users lost over 1% of their starting weight per EFH per day on average. The regression analysis using all variables was highly predictive of weight change at 26 weeks (R2=0.334) with starting BMI and EFH per day as the most significant predictors. Conclusions IF with LIFE mobile apps appears to be a sustainable approach to weight reduction in the overweight and obese population. Healthy weight and underweight individuals do not lose much weight on average, even with extensive fasting. Users who are obese lose substantial weight over time, with more weight loss in those who fast more.

Two-phase sampling is a cost-effective method of data collection using outcomedependent sampling for the second-phase sample. In order to make efficient use of auxiliary information and to improve domain estimation, mass imputation can be used in two-phase sampling. Rao and Sitter (1995) introduce mass imputation for two-phase sampling and its variance estimation under simple random sampling in both phases. In this paper, we extend the Rao-Sitter method to general sampling design. The proposed method is further extended to mass imputation for categorical data. A limited simulation study is performed to examine the performance of the proposed methods.

Multilevel analysis is an appropriate and powerful tool for analyzing hierarchical structure data widely applied from public health to genomic data. In practice, however, we may lose the information on multiple nesting levels in the multilevel analysis since data may fail to capture all levels of hierarchy, or the top or intermediate levels of hierarchy are ignored in the analysis. In this study, we consider a multilevel linear mixed effect model (LMM) with single imputation that can involve all data hierarchy levels in the presence of missing top or intermediate-level clusters. We evaluate and compare the performance of a multilevel LMM with single imputation with other models ignoring the data hierarchy or missing intermediate-level clusters. To this end, we applied a multilevel LMM with single imputation and other models to hierarchically structured cohort data with some intermediate levels missing and to simulated data with various cluster sizes and missing rates of intermediate-level clusters. A thorough simulation study demonstrated that an LMM with single imputation estimates fixed coefficients and variance components of a multilevel model more accurately than other models ignoring data hierarchy or missing clusters in terms of mean squared error and coverage probability. In particular, when models ignoring data hierarchy or missing clusters were applied, the variance components of random effects were overestimated. We observed similar results from the analysis of hierarchically structured cohort data.