Agreement in Qualitative and Quantitative Assessments of Disease Severity: Evidence from Pharmaceutical Reimbursement in Sweden

Abstract

Background and Objectives

Publicly funded healthcare systems that consider a trade-off between efficiency and equity by allowing a higher cost per patient benefit in patients with more severe conditions must somehow assess disease severity. Some countries employ quantitative measures of shortfall, whereas others rely on qualitative assessments. Despite its importance in pharmaceutical reimbursement and pricing, the operationalisation of disease severity in real-world decision making has rarely been scrutinized. The aim of this study was to investigate the relationship and agreement between qualitative disease severity assessments and quantitative measures of disease severity in Swedish pharmaceutical reimbursement.

Methods

Information from 36 pharmaceutical reimbursement decisions made by the Dental and Pharmaceutical Benefits Agency (TLV) in Sweden from 2018 to 2023 was extracted, including the qualitative assessment of disease severity (moderate, high, or very high). Based on publicly available decision documents from the agency, we calculated absolute QALY shortfall (AS) and proportional QALY shortfall (PS). Linear regression was used to describe the mean shortfall across severity classifications. Ordinal logistic regression was used to analyse the role of AS and PS as predictors of TLV’s qualitative disease-severity assessments and the predictive ability of both measures was compared using the coefficient of discrimination (D’).

Results

The mean AS and PS was 12.2 and 0.796, respectively, in the very high disease severity category, which was approximately twice the mean shortfall observed in the moderate and high severity categories (Moderate: AS = 6.0, PS = 0.340; High: AS = 6.2, PS = 0.405). When the quantitative measures of severity were used as predictors of the qualitative assessments, PS was better able to discriminate between TLV’s severity classifications than was AS (D’ = 34.6% vs 22.3%). However, both measures frequently predicted low probabilities of the qualitative assessments that were observed and there was both substantial variation in shortfall for diseases with the same qualitative assessment (AS, R² = 35.8%; PS, R² = 61.0%) and overlaps in observed shortfall across different severity classifications.

Conclusion

Proportional QALY shortfall agrees more closely than AS with qualitatively assessed disease severity applied in the Swedish reimbursement system but there are large variations in the qualitative assessments that cannot be explained by either measure. Further investigation is warranted to understand if this is an intended and desired outcome.