Table 1

Findings from the feasibility study and corresponding modifications in the SMART-India study

Theme	Findings	Modification
Sampling strategy	Recruiting an equal distribution of participants across the different age–sex stratum was difficult Approaching households in a random fashion to recruit participants resulted in a bias towards selection of middle-aged female and older participants who were more likely to be present at home and available for screening	Use of census data to recruit an age-representative and sex-representative sample through stratified sampling to mitigate potential selection bias
Motion artefact	Participants found it difficult to maintain steady hands while placing their fingers on the AliveCor case without additional support Participants found 2 min screening period to be too long to maintain steady hands	Use of clipboards to stabilise participants’ hands Shorter recording time: 60 s (AliveCor) and 90 s (ANAND)
Screening protocol	A single time-point screening underestimated true AF prevalence (owing to paroxysmal AF) Five screenings in a week were time consuming and led to participant attrition	Three screenings over the course of 5 days to account for paroxysmal events and prevent study attrition
Data management	Research staff had difficulties uploading ECG tracings for adjudication in a timely manner Questionnaires in the field could not be entered using smartphones and tablets due to poor internet connectivity	Development of a web platform to assist with sharing of de-identified ECG recordings Use of an off-line smartphone-based application

AF, atrial fibrillation; ANAND, Automated Novel Atrial fibrillation Noninvasive Detection; SMART-India, Smartphone Monitoring for Atrial fibrillation in Real-Time—India.