Purpose and methods. In Markov models of healthcare interventions, cohort methods are often used to calculate population QALYs and costs. However, analysts may also wish to determine equilibrium population levels before and after an intervention. We review methods from the stochastic processes literature that can be used for this purpose. Results. We consider several real and hypothetical examples. One is a simple model of individuals at risk for a hypothetical disease with annual incidence 12 per thousand, disease mortality 10 times the healthy annual mortality of 4 per thousand, and quality of life 0.20 in the disease state. Consider a seemingly beneficial intervention that halves disease mortality. Cohort methods yield an increase of 3.75 QALYs on average for an initially healthy individual. Equilibrium methods yield an increase in disease prevalence from 23.1% to 37.5%, a decrease of 0.115 yr in population average QALYs accrued per year, and a decrease of 4.6 yr in population average lifetime QALYs. The problem with these equilibrium measures is that they do not account for population size increase due to better survival. The corresponding total population measures behave reasonably: Population total QALYs accrued per year increases by 5.7% and population total lifetime QALYs increases by 12.2%. In a second example, we show, however, that there are interventions that increase initially-healthy individual QALYs and nevertheless decrease population total lifetime QALYs. It is only population total QALYs accrued per year that behaves consistently with initially-healthy individual QALYs. We also apply these equilibrium methods to a published analysis of tamoxifen use for breast cancer prevention (Col et al 2002), where we find no comparable anomalies. Conclusions. For Markov models of healthcare interventions, equilibrium methods from the stochastic processes literature may be a useful adjunct to cohort-based QALY calculations. However, caution is advised in choosing measures of effectiveness, as seemingly intuitive equilibrium measures of population health may conflict with each other and with individual-level measures.