Variations in healthcare costs by body mass index and obesity-related complications in a UK population

The global prevalence of obesity (BMI > 30) has risen over recent decades, and this trend is set to continue. According to estimates in the 2023 World Obesity Atlas, 38% of the global population over 5 years old was either overweight or obese in 2020. This figure is expected to rise to 51% by 2035, and 24% of individuals are projected to be obese by that year.

This study aimed to estimate healthcare resource utilisation (HCRU) and healthcare costs BMI in a UK cohort and to explore how this varied by defined BMI strata.

This retrospective open cohort study used data from individuals identified in the Discover dataset.

The study period began on 1 January 2004 and ended on 31 December 2019, to exclude healthcare data from the period during which healthcare systems were disrupted by the COVID-19 pandemic.

Adults were stratified by BMI as: overweight (25–<30), obesity class I (30–<35 kg), obesity class II (35–<40 kg) or obesity class III (≥40 ). Cost data, comprising primary care, secondary care (inpatient admissions, outpatient appointments and emergency room visits) and prescriptions, were reported for 2015–2019.

Overall, 1,008,101 individuals were overweight, 278,782 had obesity class I, 80,621 had obesity class II, and 42,642 had obesity class III.

Healthcare costs and HCRU events per person per year increased over time (2015: £851–£1321 and 10.6–13.4 events; 2019: £1143–£1871 and 11.4–14.9 events), and were higher for each successive BMI group.

Groups with Chronic Kidney Disease or Cardiovascular Disease incurred particularly high costs. In 270,493 individuals with obesity in 2019, more than 72% of total healthcare costs were incurred by the highest cost quintile, which had a higher mean age and more obesity-related complications (ORCs) than lower cost quintiles.

The economic impact of obesity could be alleviated by weight management support based on unmet need, to limit the effects of BMI progression and ORC development.

You can read the abstract in full here.