JMIR Publications

ABSTRACT

Background:

Social isolation and reduced engagement carry well-documented health risks for older adults, yet staff in senior living communities rarely have a systematic signal for which residents need attention on a given day. We developed a connection score, an automated measure that flags residents who may need additional staff connection based on signals such as reduced activity participation and upcoming significant dates. Higher scores indicate greater need, and the score is surfaced to staff to prioritize outreach.

Objective:

We examined how quickly flagged residents return to baseline, and whether residents who attend more community events resolve faster.

Methods:

We retrospectively analyzed connection score histories across 144 senior living communities (5,771 residents). A flag was defined as a score of 100 or higher and resolution as the first subsequent score below 100. Time to resolution was estimated using the Kaplan-Meier method to account for flags still open at the end of follow-up. In one community (97 residents) with 35-day follow-up and individual event-attendance records linked by resident identifier, we compared resolution between residents above and below the median number of events attended, and we examined documented 1:1 staff visits within 48 hours of a flag.

Results:

We identified 849 flag episodes; 51% resolved within the initial 14-day window, the estimated median time to resolution was 6.6 days, and 92% of resolutions returned fully to baseline. In the community with longer follow-up, 80% of 166 flags resolved (estimated median 7.0 days). Residents who attended more events resolved faster (estimated median 6.0 vs 8.0 days) and more completely (100% vs 86% resolved by 35 days), and spent more of their observed time at baseline (Spearman rho = 0.57). Documented 1:1 Connects within 48 hours of a flag were infrequent (4% and 12% across two communities).

Conclusions:

An automated connection score surfaces at-risk residents, most flags return to baseline within one to two weeks, and resolution is typically complete rather than partial. Higher event attendance was associated with faster resolution. Because reduced activity is itself an input to the score, this association is partly definitional and not causal, and findings rest on one or two communities. Prospective, multi-community study linking documented staff outreach to resident outcomes, including social isolation and antipsychotic use, is warranted. Clinical Trial: Not applicable (this is a retrospective observational analysis, not a clinical trial).