Optimizing recruitment in rare disease research: a cross-sectional online study evaluating sampling strategies for hard-to-reach populations

Researchers in the field of rare diseases are often confronted with difficulties in achieving sufficiently large sample sizes, given the small case numbers and geographical dispersion of patients. We aimed to evaluate three different sampling strategies that have proven effective for other hard-to-reach populations and compare their effectiveness in the context of rare diseases. Within a cross-sectional online study, we compared three sampling strategies (respondent-driven sampling, online-based sampling, and location-based sampling) for their effectiveness in recruiting patients with three rare diseases. Additionally, we compared characteristics of recruited patients. All participants completed an online questionnaire using REDCap. Our primary outcome was the number of patients recruited by each sampling strategy. We further assessed study perception, sociodemographic and clinical characteristics, as well as validated measures to assess depression severity (PHQ-9), anxiety severity (GAD-7), illness cognitions (ICQ), health-related quality of life (SF-12) and psychological burden through somatic symptoms (SSD-12). A total of N = 254 individuals accessed the survey website and N = 225 completed the sociodemographic characteristics and were included in the analysis. Mean age was 42.53 years (SD = 13.06) and N = 156 (69%) participants were female. Online-based sampling yielded the highest number of participants (N = 184, 82% (95% CI [79%, 85%])), followed by location-based sampling (N = 22, 10% (95% CI [4%, 16%])) and respondent-driven sampling (N = 19, 8% (95% CI [2%, 14%])). Patient characteristics differed significantly regarding gender and satisfaction with medical care, with online-sampling having the highest share of female participants and patients recruited via location-based sampling reporting a higher satisfaction with their overall care. Across all three sampling strategies, participants showed typical features of populations affected by rare diseases such as high rates of depression and anxiety symptoms and reduced quality of life. Our study identified online-based sampling as the most effective recruitment strategy for patients with rare diseases. It may be the most promising approach, especially with limited recruitment periods. Potential biases such as gender imbalances should be considered. We encountered substantial challenges with respondent-driven and location-based sampling. Addressing these challenges in future studies may help to make better use of the potential that lies in these sampling strategies.