Evaluating the impact of certificate-of-need regulation using measures of ultimate outcome: some cautions from experience in Massachusetts.

Prior evaluations of certificate of need (CON) have concluded that such regulation did little to reduce the level of hospital capital investment. These studies, however, failed to examine the underlying patterns of hospital investment behavior onto which CON was imposed and further neglected to determine the lag period that separates the introduction of regulation and the appearance of outcomes actually attributable to the presence of CON. This article addresses these two issues by using rate-setting data to examine the details of hospital capital investment across the whole voluntary hospital industry in one state--Massachusetts--both before and after the introduction of certificate of need. Massachusetts voluntary hospitals were found to devote most of their capital expenditures to the construction of major new inpatient facilities built in cycles of 14 years average duration throughout the post-World War II period. The date of completion of such a facility proved to be the major determinant of a hospital's capital expenditure pattern over time, and hospitals of similar teaching status and geographic location showed considerable synchrony in the construction of such facilities. At least in Massachusetts, the introduction of certificate-of-need regulation coincided with but was not responsible for the conclusion of a major construction peak among nonteaching hospitals and the beginning of such a peak among teaching hospitals, greatly complicating the evaluation of actual CON effects. Further, capital expenditures actually approved by CON did not appear until the third year of program operation, and even five years post-CON, the majority of hospital capital expenditures reflected projects approved in the first year of regulation, a period during which the program had neither the expert staff nor extensive review capacity that characterized CON function in later years. These findings hold implications for the evaluation of CON effectiveness and, indeed, for the future existence of certificate of need.