Recruitment Strategies in Registry-Based Observational Research: A Single-Center Experience in the United States from the HeartShare Study.

BACKGROUND

Recruiting a representative sample of US adults is challenging, and strategies to improve recruitment have not been well evaluated. This study examined differences in consent rates across 3 recruitment approaches for the HeartShare registry at a single academic center.

METHODS

The HeartShare registry is a multicenter observational cohort enrolling adults age ≥30 years with or without heart failure. We included participants contacted at a single academic center between August 1, 2023 and August 1, 2024, comparing consent rates across 3 recruitment methods: (1) telephone; (2) electronic health record (EHR) messaging; or (3) in-person. Secondary analyses evaluated differences in consent rates within recruitment strategies by sociodemographic factors.

RESULTS

Among 9712 patients contacted (mean age, 72.0; standard deviation [SD] 13.1 years; 46.5% female), 4.2% consented. Of the 1439 patients contacted by telephone, 175 (12.1%) consented. Of the 8080 patients contacted through EHR messaging, 144 (1.8%) consented. Of the 193 contacted in-person, 79 (41%) consented. Compared with phone recruitment, EHR had the lowest consent rate (adjusted odds ratio [aOR], 0.13; 95% confidence interval [CI], 0.10-0.17) while in-person contact had the highest consent rate (aOR, 6.49; 95% CI, 4.42-9.53). In-person recruitment showed no significant differences in consent rates between Black and White participants (aOR, 0.69; 95% CI, 0.30-1.56). Among those approached via EHR or phone, Black individuals had significantly lower odds of consenting compared with White individuals (EHR: aOR, 0.31; 95% CI, 0.12-0.67; Phone: aOR, 0.36; 95% CI, 0.18-0.67).

CONCLUSIONS

In-person recruitment achieved the highest consent rates across groups in this observational registry-based cohort. Although less time intensive, automated EHR recruitment was less effective in engaging underrepresented groups.