Presentation (CTI)

OBJECTIVES: Cost-effectiveness thresholds in Sweden appear to have remained broadly unchanged despite substantial inflation and rising healthcare costs. Over the past decade, the healthcare price index used by Swedish regions (price index with quality-adjusted wages for regions, LPIK) increased by 37%, while the estimated value of a statistical life used by the Swedish Transport Administration has also risen markedly. This study aimed to assess whether the accepted cost per QALY in TLV reimbursement decisions changed between 2015 and September 2025, in both nominal terms and after adjustment for LPIK.
METHODS: We extracted all public TLV reimbursement decisions (general and restricted), published between 2015 and September 2025 that reported an accepted cost per QALY and disease severity assessment. Disease severity was classified according to TLV definitions as low, moderate, high, or very high. An additional category captured cases with both very high severity and rarity designation. Two linear regression models were estimated: one using nominal ICERs and one using ICERs adjusted for LPIK. Explanatory variables included disease severity, rarity, year, and presence of confidential price agreements.
RESULTS: A total of 125 cases were included. In the model using nominal ICERs, accepted cost per QALY showed a small, non-statistically significant increase over time. In contrast, in the LPIK-adjusted model, accepted cost per QALY declined significantly by SEK 20,457 per year (p < 0.05).
CONCLUSIONS: The informal reimbursement threshold in Sweden, measured by accepted cost per QALY in TLV decisions, appears stable in nominal terms. However, when the accepted cost per QALY is adjusted with LPIK, the cost per QALY accepted by TLV have decreased. This suggests that new technologies are being assessed against a lower real willingness-to-pay threshold over time, with potential implications for access, consistency, and the interpretation of value-based reimbursement decisions in Sweden.