BACKGROUND: While life expectancy quantifies average length of life, health expectancy represents the average lifetime in different health states and offers the possibility to evaluate quality of life with respect to health. The purpose of the study was to estimate changes in health expectancy in Denmark from 1987 to 2000 and to assess theories about the relation between increased total lifetime and lifetime in various health states. METHODS: Data on health status derived from the Danish Health Interview Surveys carried out in 1987, 1991, 1994 and 2000 were combined with life-table data. Expected lifetime in selfrated good health, life expectancy without longstanding illness and disabilityfree life expectancy were estimated by Sullivan's method. RESULTS: In 1987, the life expectancy of a 65-year-old man was 14.1 years, 8.9 years of which were expected to be disabilityfree. In 2000, life expectancy had increased to 15.0 years, 11.3 years of which were disabilityfree. Thus, life expectancy had increased by 0.9 years, whereas disabilityfree life expectancy had increased by 2.4 years. Among 65-year-old women, life expectancy had increased by 0.2 years and disabilityfree life expectancy by 1.1 years. Expected lifetime in selfrated good health had also improved, but the trend in life expectancy without longstanding illness went in the opposite direction, and expected lifetime with longstanding illness had increased. CONCLUSION: The recent rise in life expectancy in Denmark after many years of stagnation appears to be accompanied by generally improved health status among the elderly, but health expectancy trends depend on the health indicator chosen. KEY POINTS: Health expectancy expresses average lifetime in various states of health. The study examines changes in expected lifetime in selfrated good health, lifetime without longstanding illness and lifetime without longterm disability. Among 65-year-olds the percentage of disabilityfree life expectancy increased from 63.4% to 74.9% for men and from 55.6% to 61.0% for women between 1987 and 2000. Health status among elderly Danes has apparently improved, but secular trends in health expectancy depend on the choice of health indicator.