Response to Recent Commentary
We write in response to the recent commentary on our article “Feasibility of conducting autism biomarker research in the clinical setting.”1 We respectfully disagree with the author's assertion that our study was not successful in collecting biomarker data on patients with autism spectrum disorder (ASD), as collection rates ranged from 87% to 94% across individual biomarkers. Although Dr. Wasserman makes important points about current gaps in clinically based Developmental Behavioral Pediatric research, we take issues with some statements regarding our study methodology.
First, Dr. Wasserman comments that the recruitment rate of 11% was “insufficiently productive.” Although we agree, as noted in our limitations, that this recruitment rate is low, this rate is in line with other multisite ASD studies. For example, in a larger national study (the Study to Explore Early Development [SEED]) conducted by the Centers for Autism and Developmental Disabilities Research and Epidemiology (CADDRE) network, the final recruitment rate was also 11%.2 These authors contacted 26% of potential participants by phone; of those reached, 34% declined and 43% consented. In our study, we reached 27% by phone; 32% declined and 43% enrolled. In addition, many published ASD studies use convenience samples and do not provide the total participant pool precluding assessment of recruitment rate.
Dr. Wasserman states his suspicion that our study lacked piloting and close monitoring of the recruitment process. However, as noted in our Methods and Discussion sections, the purpose of this study was in fact to pilot the methodology, and we conducted continuous monitoring and iterative modification of the recruitment process (pp 486, 489).3 We agree that increasing the number of calls might have improved recruitment, but note that this strategy was used in the SEED study and did not result in a higher recruitment rate. Furthermore, although we were limited by the number of calls because of regulatory constraints, calls were made during the day and evenings to maximize the potential for contact. In addition, the recommended use of letters before phone contact was in fact our method of recruitment (p485).3 The suggestion to develop a methodology which might capitalize on clinically required blood draws is well taken. One challenge, however, is that only a small percentage of clinical visits for ASD result in recommendation for blood draw.
Nevertheless, additional methods of recruiting will need to be considered in future. We believe that low recruitment rates in studies on ASD reflect unique participation barriers for this patient population. These barriers include family stress related to the child's disability, child behaviors that make new situations and long visits challenging, and perhaps also a lack of perceived direct personal benefit. In contrast to fields such as oncology where participation in research provides access to cutting-edge and life-saving therapy, the ASD field has not yet reached this stage. We agree that clinically based networks can capture rich data, allowing for studies with higher power and comparison of outcomes across sites and settings. Clinical networks can also expand the socioeconomic and clinical diversity of participants, and this will be critical in ASD and other complex heterogenous neurodevelopmental disorders.