Aim.  This paper is a report of a cohort study of healthcare workers’ work attendance, and its long-term consequences’ on health, burnout, work ability and performance. Background.  Concepts and measures of work attendance have varied in the scientific literature. Attending work in spite of being sick can have serious consequences on health. There is little knowledge on which individual and work-related conditions that increase work attendance and the long-term impact on health and performance. Method.  Prospective analyses of three measures of work attendance i.e. sickness attendance, uninterrupted long-term attendance and balanced attendance (≤7 days of sick leave per year and no sickness attendance) were done using questionnaire data from a 2-year cohort study (2004–2006) of randomly selected healthcare workers (n = 2624). Incentives (e.g. effort-reward balance, social support, meaningfulness) and requirements (e.g. time-pressure, dutifulness, high responsibility) to attend work as well as general health, burnout, sick leave, work ability and performance were assessed. Results.  There was a positive relation between balanced work attendance and incentives, whereas high sickness attendance was associated with requirements. Follow up after 2 years showed that balanced attendance was associated with sustained health and performance while sickness attendance was associated with poor health, burnout, sick-leave and decreased performance. Conclusion.  It is important to distinguish between measures of work attendance as they differ in relation to incentives, and health- and performance-related consequences. Sickness attendance seems to be an important risk indicator. A balanced work attendance should be promoted for sustained health and performance in healthcare organisations.