Clinical Research by Developmental and Behavioral Pediatrics Specialists: Which Models Work Best?

    loading  Checking for direct PDF access through Ovid

Excerpt

T. Berry Brazelton and William Carey conducted pioneering developmental and behavioral pediatrics (DBP) research in their own primary care practices. However, small patient populations and scarce resources limit single-site research. In addition, studying patients with specific clinical conditions necessitates the pooling of patients and resources, often across multiple subspecialty care sites. In this issue of JDBP, Sices and colleagues describe a feasibility test of a “lean” multisite translational research model designed to collect biomarker data in patients with autism spectrum disorder (ASD). Briefly, research data were collected during regular clinic visits to capitalize on scheduled encounters and maximize convenience for families. Medical record documentation of clinical diagnosis, confirmed by a Diagnostic and Statistical Manual checklist at the time of visit, substituted for extensive diagnostic ascertainment. DBP researchers from 5 sites who care for ASD patients used their clinic infrastructures to collect data from patients and families, aiming at minimal disruption to clinic workflow. Although all sites had at least 20% support for a research assistant, their very part-time status meant that clinicians and staff were relied upon to do much of the research work.
In light of the calls for involvement of more DBP subspecialists in research and development of efficacious and scalable research practices, we should praise the authors for systematically assessing this lean approach, which, in spite of its virtues, proved lean in too many ways. Among the meanings given for the adjective “lean” at m-w.com, we find not only positive senses, such as “characterized by economy,” but also negative ones: “lacking productiveness,” and “deficient in an essential or important quality or ingredient.”1 In harmony with these latter definitions, this lean model was insufficiently productive, likely due to a shortage of important ingredients. The lack of productivity was shown by the fact that only 7% of potential participants (11% of those approved for recruitment by their physician) ultimately were recruited. Thus, leanness did not translate into economic efficiency. From a methods standpoint, such low participation rates preclude any generalization of study results.
In regard to deficient ingredients, what was missing from this effort? In clinical research, subject recruitment is the foundation on which all is built. If there are too few subjects, there is no study. We can only speculate, but it seems that resources devoted to and experience in subject recruitment were both lacking. Over half of potential subjects were lost at the lead site because caregivers were not reached after 3 phone calls. Loss of potential subjects at this critical juncture was even more pronounced at the 4 other sites. The authors do not reveal the extent to which they piloted their recruitment protocol, but I suspect that insufficient resources may have limited piloting across sites and precluded closer monitoring of study recruitment. A brief period of pilot testing should have quickly revealed a problem and prompted an earlier shift in study protocol and resources. In the 21st century, multimodal recruitment strategies are needed. Three phone calls is a fraction of the number typically required to get through to families in an era when contact mainly occurs in the evenings, after parents return home from work. Also, parents are highly suspicious of phone calls from unknown numbers, so letters and postcards (and even texts) warning families in advance of upcoming recruitment phone calls often must be used. Other resources may have been missing as well, but the voltage drop between the number of potential subjects and those actually reached was the key. In addition, study-specific blood draws dissuaded many of the parents who were reached. A different protocol might have capitalized on blood needed for clinical purposes.
    loading  Loading Related Articles