Quantifying the causal impact of biological risk factors on healthcare costs

Background A critical step in evaluating healthcare interventions is to understand their impact on healthcare costs. However, there is a limited understanding of the causal impact that biomarkers and risk factors for disease have on healthcare-related costs. Previous studies based on observational data have major limitations including residual confounding and reverse causation. Here, we used a genetically-informed design, Mendelian Randomization (MR), to infer the causal impact of 15 routinely measured and clinically relevant risk factors on annual total healthcare costs.

Methods We considered 373,160 participants from the FinnGen Study, which were linked to detailed healthcare costs covering inpatient, outpatient, and medication costs. Several MR approaches were used to assess the causal effects of 15 risk factors (e.g., waist circumference (WC), HDL cholesterol, vitamin D), with strong genetic bases on annual total healthcare costs, as well as stratified by service type, age, and sex. We further assessed the generalizability and robustness of our results by accounting for selection bias and by leveraging additional data from 323,774 individuals from the United Kingdom and Netherlands.

Results Robust causal effects were observed for waist circumference (WC), adult body mass index, and systolic blood pressure, in which a one standard deviation increase in the risk factors corresponded to 22.78% [95% CI: 18.75, 26.95], 13.64% [10.26, 17.12], and 13.08% [8.84, 17.48] increased annual total healthcare costs, respectively. The relative effect of WC on annual total healthcare costs was consistent across age and sex and was not attenuated when accounting for increased risk of five major diseases: back pain, chronic ischemic heart disease, type 2 diabetes, chronic obstructive pulmonary disease, and stroke. A lack of causal effects was observed for some clinically relevant biomarkers, such as albumin, C-reactive protein, and vitamin D.

Conclusion Our results indicated that increased WC is a major contributor to annual total healthcare costs and more attention should be given to WC screening, surveillance, and mitigation. On the contrary, several biomarkers relevant in clinical settings did not have a direct impact on annual total healthcare costs.