The objective of the Managed Care Outcomes Project was to examine the relationship of various health maintenance organization cost-containment strategies with the utilization of ambulatory care visits, hospital admissions, and prescription drugs.

In this longitudinal, prospective study, we compared utilization of ambulatory services for patients having at least one of five diseases (arthritis, asthma, epigastric pain/ulcer, hypertension, and otitis media) in HMO settings with various levels of cost-control measures, including visit copayment and limitations on drugs to treat these diseases. We used multivariate regression to control for other variables, including severity of illness, provider count, age, gender, and time in the study.

The study setting was six health maintenance organizations in six states (three is the eastern US; three in the western US). Between 1,309 and 3,938 patients were available for analysis for each disease studied. A total of 12,997 patients were assessed for severity of illness at every encounter for one year, accounting for more than 99,000 office visits, 480 emergency department visits, 1,000 hospitalizations, and 240,000 prescriptions. This was a prospective observational study. Outcome measures were prescription count, prescription cost, office visit count, emergency department visit count, and hospital admissions.

The study found healthcare utilization to be strongly associated with severity of illness; all analyses took this finding into account. For all conditions except otitis media, formulary limitations on drug availability were significantly positively related to higher rates of emergency department visits and hospital admissions, and positively, but not always significantly, related to drug cost, drug count, and office visits. Ratios of utilization between the site with the unrestricted formulary and the one with the most restricted formulary for a disease ranged up to more than twice as great. Use of multisource drugs was strongly positively related with drug use for all conditions, but was less consistently related to visit frequency, emergency department visits, or hospital admissions.

Stricter second-opinion requirements, salaried physicians, and strictness of case management had mixed, mostly nonsignificant, associations with drug and other resource utilization. Visit copayment always was associated with lower drug utilization, but was mixed in its association with office visits, emergency department visits, and hospitalizations. As expected, "strictness of gatekeeper" was associated with lower drug utilization, emergency department visits, and hospitalizations and had a mixed effect on office visits.

In this study, we found some expected and some unexpected associations of common health maintenance organization cost-containment practices with utilization of healthcare services. In the case of limited formularies, we found the unintended consequence of increased utilization of healthcare resources. It is important to assess combinations of cost-containment strategies, because the individual strategies do not function independently, and the results of changing one component may not be easily predicted. A systems approach to cost containment, rather than individual component management techniques, is needed.