Background/Purpose: Participant recruitment is a critical component of randomized controlled trials, particularly large ones with long term outcomes like work disability. Recruitment needs to be cost-effective; but few guidelines are available to estimate recruitment costs.  The objective of this study is to evaluate the recruitment methods to date of the “Work-It Study”, a large randomized controlled trial examining the effects of a work barrier problem-solving intervention delivered by physical and occupational therapists on preventing work disability among persons with arthritis or rheumatic conditions.   

Methods: Recruitment approaches included: 1) rheumatology and medical registries (displaying brochures in practices, mailing physician letters to patients, and mailing letters to persons in a medical registry), 2) social media (Facebook, LinkedIn, and Patientslikeme) and 3) community advertising (Craigslist, flyers, newspapers, direct marketing, and other (support groups, professional associations, and arthritis foundation events)). Recruitment costs were calculated by summing printing, mailing, and personnel costs. Percentages were calculated for screened eligible, enrolled, and yield (# enrolled/#participants contacted), as available. Cost per enrolled participant in each recruitment approach was calculated. The different approaches were compared in terms of yield and cost-effectiveness. A 20-month recruitment time period was used for all approaches.   

Results: 440 people were screened, with 74% eligible; 190 were enrolled to date (58% of screened eligible; 54% of targeted sample); average cost per subject enrolled was $115 (see Table 1 for details). Letters to patients and community approaches resulted in the highest numbers of enrolled participants. Social media generated zero participants enrolled. Yield for mailings ranged from 1-5%; the medical registry had the highest yield. Costs per-participant enrolled were lowest for medical registry, physician letters, flyers, and ‘other’ approaches; displaying brochures, newspapers, and direct marketing had the highest cost per-participant enrolled. Though cost-effective, physician letters and patient registries would require large numbers of persons to be contacted to meet the desired sample size of 350 (n=8750 and n=7000 respectively). With the exception of direct marketing, community advertising approaches can generate participants but longer time periods would be needed to recruit (e.g., approximately 4 years would be needed to recruit 350 participants).  

Conclusion: Direct mailings to patients from rheumatology practices and medical registries are the most cost-effective approaches but may not be feasible given the low yield. Physician willingness to support these methods is essential. Community advertising is feasible but may take longer time periods if large sample sizes are needed.