Measures of Engagement with mHealth Interventions in Patients with Heart Failure: A Scoping Review (Preprint)

Abstract

BACKGROUND The computing and communication features of mobile devices are increasingly leveraged in mHealth interventions to provide comprehensive and tailored support that may have positive outcomes in patients with heart failure (HF). However, examination of mHealth intervention effectiveness has provided mixed findings. Considering that patient engagement is a prerequisite for the effectiveness of interventions, understanding how patients engage with mHealth interventions, and the effects of patient engagement on HF outcomes may explain the mixed findings. OBJECTIVE Our aim was to synthesize current evidence on measures of patient engagement with mHealth interventions, and the effects of engagement on HF outcomes METHODS A comprehensive search of the literature was conducted in 7 databases for relevant studies published in the English Language from 2009 to September 2021. Descriptive characteristics of studies were reported. Content analysis was conducted to identify themes that described patient engagement with mHealth in the qualitative studies included in the review. RESULTS We synthesized 32 studies that operationalized engagement with mHealth interventions in 4771 patients with HF (67.9% male), ranging from a sample of 7 to 1571, with a median of 53.3 patients. Patient engagement with mHealth interventions was measured only quantitatively based on system usage data (71.8%, 23/32), only qualitatively based on data from semi-structured interviews and focus groups (6.3%, 2/32), and by a combination of both quantitative and qualitative data (21.9%, 7/32). System usage data were evaluated using 6 metrics of engagement: (1) number of physiological parameters transmitted (63.3%, 19/30); (2) number of HF questionnaires completed (6.7%, 2/30); (3) numbers of logins (13.3%, 4/30); (4) number of short message service (SMS) responses (3.3%, 1/30); (5) time spent (16.7%, 5/30); (6) number of features accessed/screen viewed (9.5%, 4/30). There was a lack of consistency in how system usage metrics were reported across the studies. Eighty percent of the studies reported only the descriptive characteristics of the system usage data. Emotional, cognitive, and behavioral domains of patient engagement were identified in qualitative studies. Patient engagement levels ranged from 45% to 100% and decreased over time. The effects of engagement on HF knowledge, self-care, exercise adherence, and HF hospitalizations were inconclusive. CONCLUSIONS The operational definitions of patient engagement with mHealth interventions are underreported and lack consistency. The application of inferential analytical methods to engagement data is extremely limited. More research focused on developing optimal and standardized measures of patient engagement that may be applied across different study designs is warranted.