Application of H-ARF led to successful recruitment, as demonstrated by STORY. The findings highlight three areas that researchers should consider when devising recruitment plans: absolute numbers versus recruitment rate, cost, and efficiency with HIPAA waiver. However, the model remains a work in progress. Additional clarity is needed in the operational definitions, and outcome measures should be refined. The work is an early effort to develop the model and describes its use and evaluation in one study.
When considering findings in – collectively, recruitment rates were the lowest for the institutions that granted a HIPAA waiver; however, the authors noted that they actually recruited the largest number of participants from these sources. This highlights the need for researchers to consider not only recruitment rates for their various recruitment plans, but also the absolute numbers each source is expected to yield. The researcher must weigh the benefits and drawbacks for the various plans and make an informed decision that will yield the most efficient and cost-effective method. Unfortunately, this is not always clear-cut because interplay often exists between the various methods. For example, social marketing techniques increase study visibility and name recognition, which yield lower absolute numbers but could ultimately impact recruitment rates for any location. In addition, the authors would have preferred a waiver at all sites, but unwillingness from physicians, institutions, or institutional review boards precluded using this approach at all locations.
Comparison of Sources for Recruited Patients
On initial evaluation, the least expensive approach with the greatest yield is the use of HIPAA waiver. However, in a research-naïve environment, the authors cannot comfortably recommend this as a sole approach and would expect recruitment efforts without social marketing to be abysmally low. The general population has been poorly educated about clinical trials and behavioral studies and is potentially more distrustful, particularly African Americans. In addition, the population’s known low literacy rate increased the importance of word-of-mouth marketing and presentations to increase knowledge about the project. Analysis of the impact of health fairs, posters, and other social marketing activities was beyond the scope of the current project. No data exist on motivation for social marketing responses or population baseline knowledge of research. Market researchers encountered similar issues when trying to determine the impact of an ad campaign. In seven waves, the authors obtained a total of eight patients through social marketing techniques. This represented 3% of the total planned sample (n = 240) for the study. Of the social marketing techniques, self-addressed return postcards attached to brochures or in a pocket with a poster yielded good results (25 of 39 responses). Printing of the posters and cost of the acrylic stands was less than $5 each, but the staff time and gas and mileage to place them across the state was expensive. If this method were used again, the authors would develop a coding system to determine exactly where a woman received the brochure or picked up a return card from the poster.
Print advertisements were very expensive. Ad placement in African American–focused publications yielded only one response, and that occurred a year after the ad was placed. In contrast, articles in local newspapers were free but involved staff time to meet with reporters, provide information, and provide consent for patients to be interviewed. However, the articles built good will with community partners. At a minimum, social marketing techniques for this kind of study should include an information brochure, a poster, and newspaper articles. The authors do not recommend paid media advertising; instead, placement of posters in locations where patients are likely to see them (e.g., cancer boutiques, treatment areas, diagnostic centers) was much more effective.
Results point to the need for additional relationship building with hospitals and physicians to obtain a waiver. Clear differences are present in recruitment rates when comparing HIPAA waiver to release of information or social marketing. This suggests that, when patients are approached by the treatment team, they may be too overwhelmed to agree to be in the study. However, if the communication to the patients occurs in the privacy of their homes and at their convenience, it may lead to higher recruitment rates and a sample that is more representative of the population of patients who would participate. This may be particularly true if patients need more information or time to decide about the study. In addition, the release of information may be a burden on staff and may lead to some selection bias about who is approached for the study.
In summary, H-ARF provides a testable model for researchers developing or implementing a recruitment plan. More research is needed on the social marketing component, particularly regarding cost. Additional ways to reduce barriers toward granting HIPAA waivers should be explored.