Skip to main navigation Skip to search Skip to main content

Selection Incentives in the Dutch Basic Health Insurance: To What Extent Does End-of-Life Spending Contribute to Predictable Profits and Losses for Selective Groups?

Research output: Contribution to journalArticleAcademicpeer-review

129 Downloads (Pure)

Abstract

Existing risk-equalization models in individual health insurance markets with premium-rate restrictions do not completely compensate insurers for predictable profits/losses, confronting insurers with risk selection incentives. To guide further improvement of risk-equalization models, it is important to obtain insight into the drivers of remaining predictable profits/losses. This article studies a specific potential driver: end-of-life spending (defined here as spending in the last 1–5 years of life). Using administrative (N = 16.9 m) and health survey (N = 384 k) data from the Netherlands, we examine the extent to which end-of-life spending contributes to predictable profits/losses for selective groups. We do so by simulating the predictable profits/losses for these groups with and without end-of-life spending while correcting for the overall spending difference between these two situations. Our main finding is that—even under a sophisticated risk-equalization model—end-of-life spending can contribute to predictable losses for specific chronic conditions.

Original languageEnglish
Pages (from-to)819-833
Number of pages15
JournalMedical Care Research and Review
Volume79
Issue number6
DOIs
Publication statusPublished - Dec 2022

Bibliographical note

Publisher Copyright:
© The Author(s) 2022.

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

  1. SDG 3 - Good Health and Well-being
    SDG 3 Good Health and Well-being

Fingerprint

Dive into the research topics of 'Selection Incentives in the Dutch Basic Health Insurance: To What Extent Does End-of-Life Spending Contribute to Predictable Profits and Losses for Selective Groups?'. Together they form a unique fingerprint.

Cite this