Table 4. Summary of the risks of bias in the MAVIE cohort and strategies for its minimization.

	Bias risk or limitation	Strategies
Sample size and participation bias	Low response rate	Diversifying recruitment forms: mailing from mutual insurance, posters, media, and social networks Analysis conducted excluding injuries reported on the consent date Using a reference person and in particular caregivers for the elderly Incidence rates standardized according to sociodemographic characteristics
	Lack of representativeness
	Particular interest in injury prevention or intellectual motivation
	Risk of exclusion of non-users or infrequent users of the internet
	Participation due to recent HLI for oneself or for a relative
Dropout	Forgetting or losing interest in the study	Diversifying retention strategies: quarterly reminders, e-mails, monthly newsletter, online forum, smartphone application Using a reference person and in particular caregivers for the elderly
Response	Events under-reporting	Quarterly email reminders One-click event reporting. Reporting possible from the smartphone application
	Data are self-reported	Confidentiality and anonymity are guaranteed by the study investigators Events are collected prospectively
	Reliability of responses	Measuring exposures before the events report (prospective approach) Involving reference persons to communicate data on the elderly Assessing consistency of information
	Missing data	Questionnaires can be completed in several goes Imputation in data analysis