Factors Predicting Completion of Daily Phone Assessments




Walters, Scott
Li, Tommy
Nandy, Rajesh


Journal Title

Journal ISSN

Volume Title



Objective: Participants in long-term studies often have varying rates of compliance with study activities, especially when those activities are required daily for extended periods of time. This study examined predictors of daily survey completion among a group of permanent supportive housing residents enrolled in m.Chat, a health coaching program. Methods: A subset of participants in the program received a prepaid cellphone with unlimited voice and text minutes, with the requirement that they complete a short automated assessment each morning. The phone recorded the number of days the participant was prompted to complete the assessment, the number of completed assessments, and the percentage completed each month. We examined age, sex, race, mental illnesses, reading ability, and alcohol use as predictors of percent completion. Results: Of the 109 participants who carried the phone for at least one month, three predictors impacted daily completion rate: a history of hallucinations, binge drinking, and baseline reading level. Of the three, having experienced hallucinations was the only statistically significant predictor. Subjects who had not experienced hallucinations in the past year had a completion rate approximately 10% higher than those who had experienced hallucinations. Conclusion: While some characteristics of a participant, namely their age, sex, race, and substance use, do not accurate predict the likelihood of completion rate of a daily assessment, other characteristics such as hallucinations may play a role. Additional research could help identify reliable predictors of compliance that would allow investigators to maximize response rates to long term assessment protocols. This may allow investigators to predict how reliable a participant will be when daily completion of an assessment is required on the phone.