Kielland et al. (2025)
- Authors: Anne Kielland, Jing Liu, Guri Tyldum, Leonard A. Jason
- Institutes: Fafo Institute for Labour and Social Research, Oslo, Norway, Center for Community Research, DePaul University, Chicago, IL, USA
- Publisher: Journal of Health Psychology
- Link: DOI
Summary
This study demonstrates that traditional medical registry data for ME/CFS can be significantly biased by sociodemographic factors, potentially excluding underserved populations. By implementing an online peer-referral sampling method (RDS), researchers can achieve a more representative sample of the patient community. This approach is vital for generating accurate prevalence data and ensuring that healthcare policies address the needs of all affected individuals, regardless of their socioeconomic status.
What was researched?
The research investigated the effectiveness of online Respondent-Driven Sampling (RDS) as a method to reach a more diverse and representative ME/CFS population than what is found in official medical registers. It specifically aimed to address selection biases in the G93.3 (ICD-10) registry data in Norway.
Why was it researched?
Official medical registers often suffer from bias because patients with higher socioeconomic status or better access to specialized healthcare are more likely to receive a formal diagnosis. There is also a concern that registry data may include individuals who do not meet strict research case definitions like the Canadian Consensus Criteria (CCC).
How was it researched?
Researchers utilized a peer-referral recruitment process (RDS) to gather a sample of 660 respondents and applied validated DePaul University algorithms to verify if they met the CCC. The sociodemographic and medical characteristics of this sample were then compared to official Norwegian registry data using regression analysis to identify diagnostic biases.
What has been found?
The study found that the likelihood of receiving an official G93.3 diagnosis was influenced by sociodemographic factors even when symptom severity was controlled for. The RDS method successfully reached a broader spectrum of the patient community, including those potentially missed by clinical registries. Using standardized algorithms provided a more homogenous and verified patient sample for research purposes.
Discussion
A primary strength of the study is its ability to reduce the socioeconomic bias inherent in many clinical databases. However, a limitation remains that even peer-referral networks may struggle to reach the most socially isolated or severely ill patients who are not active online.
Conclusion & Future Work
The authors conclude that online Respondent-Driven Sampling is a feasible and superior methodology for identifying representative ME/CFS cohorts. They recommend its use in future epidemiological studies to complement register data and improve the scientific validity of patient samples.