The study was undertaken to assess the impact of health sector reform from 1993 to 1997 in Zambia in respect of health care service utilization and the shift of caseload from hospitals to health centres. Four key indicators were chosen: general attendance, measles vaccinations, general admissions, and deliveries. Complete sets of district data were analyzed, covering 4.5 million people out of the total population in 1997 of 9.7 million. The results show, on the one hand, a dramatic decrease of about one-third in general attendance for both hospitals and health centres over a 2-year period, followed by a period with a continued but slower decrease. On the other hand, the results also show increases at health centres in measles vaccinations (up 40%), in admissions (up 25%) and in deliveries (up 60%). The study further documents a shift of caseload from hospitals to health centres for some key services. The health centre share increased from 72.2% to 79.8% for measles vaccinations, from 23.9% to 31.0% for general admissions, and from 22.9% to 32.4% for deliveries. However, the intended overall shift in outpatient caseload from hospitals to health centres did not materialize. The main lessons are: utilization patterns can be influenced by policies such as user-payment and decentralization; user payment in poor populations leads to dramatic declines in utilization of services; and decentralization with local control of resources could be an alternative to the traditional vertical disease programme approach for priority interventions.