Summary: In a diabetes and prediabetes exercise trial, combining electronic health record screening with persistent telephone follow-up filled the study, but men, Hispanic individuals, and Spanish speakers were significantly less likely to participate at every recruitment step, while neighbourhood deprivation did not predict dropout.
PICO Summary
| Element | Detail |
|---|---|
| Population | 11,152 adults aged 18–77 with prediabetes or non-insulin-treated diabetes and low physical activity (recruitment analysis). |
| Intervention | EHR-based eligibility screening plus telephone follow-up (up to five contact attempts). |
| Comparison | Mail or email notification without subsequent telephone outreach. |
| Outcome | 4,033 reached by phone, 721 consented, 451 randomised. Men, Hispanics, and Spanish speakers significantly less likely to advance (all p<0.001); women and those with prediabetes over-represented; neighbourhood deprivation did not affect randomisation. |
Expert Commentary
This is a recruitment-science paper rather than a treatment trial, and it is worth reading precisely because that plumbing is usually invisible. The headline craft lesson is unsurprising but well quantified: passive mail and portal invitations barely fill a study, and active multi-attempt phone follow-up is what converts eligibility into consent. The more uncomfortable and more valuable finding is who falls away, men, Hispanic individuals, and Spanish speakers dropping out at every step, which is exactly the selective attrition that quietly undermines the external validity of the exercise evidence I later rely on. I take the deprivation result as a useful corrective too: once people were reached, area-level poverty did not predict dropout, so the barrier was language and engagement, not postcode. Can I use this with my patients? Not as a therapy. Its real use is to make me read behavioural-trial results with a sharper eye for who was actually enrolled, and to support bilingual recruiters and Spanish materials when trials are designed. The obvious limit is that this is one health system’s pipeline, descriptive, and cannot explain why the gaps arise.
References
Young DR, Preciado M, Soto V, et al. Sociodemographic characteristics of patients throughout the recruitment process into a randomized, controlled behavioral trial. Trials. 2025;26(1):519. doi:10.1186/s13063-025-09225-7