A new study in Frontiers in Pediatrics has used machine learning to look at a large medical claims database to determine the prevalence, gender demographics, and costs for individuals living with ME/CFS. Since ME/CFS is such a heterogeneous disease, there is a lack of diagnostic testing and confusion about how to code the disease, making it difficult for an individual to be diagnosed. This also makes it problematic to estimate other aspects of the disease, such as prevalence, but this study offers a new approach to this problem by using machine learning to look at characteristics of patients who have been given a diagnosis code for ME or CFS to makes these estimates.

Based on the number of diagnoses of the disease in the medical claims database, the authors calculated the number of patients who are undiagnosed, estimating the total prevalence of the disease to be 0.519-1.038% of people, indicating that ME/CFS is actually a relatively common disease. Interestingly, the authors also claim that more men have the disease than was previously thought, with gender demographics data showing that an average of 60-65% of individuals diagnosed are female, and the remaining 35-40% are male. The authors were also able to estimate medical costs to the ME/CFS patient at an average of 4 times higher than someone in the general population. Costs to ME/CFS patients were also ~50% higher than for patients with lupus or MS.

Knowing this information about prevalence, demographics, and costs associated with ME/CFS has important implications for how the medical and research fields think about the disease. The authors say in the article that, based on their results, “ME/CFS should get more attention in research and provider communities, and warrants more education to providers (primary care, specialties, and allied health sciences) to improve the quality of health care and quality of life for affected individuals.”

To read the article in Frontiers, click here.