The influences of Taiwan's generic grouping price policy on drug prices and expenditures: Evidence from analysing the consumption of the three most-used classes of cardiovascular drugs

Background Controlling the growth of pharmaceutical expenditures is a major global challenge. Promotion of generic drug prescriptions or use is gaining increased support. There are substantial contextual differences in international experiences of implementing pharmaceutical policies related to generic drugs. Reporting these experiences from varied perspectives can inform future policy making. This study describes an experience of Taiwan, where patients with chronic (long-term) conditions are usually managed in hospitals and drugs are provided in this setting with costs reimbursed through the National Health Insurance (NHI). It investigates the effects of Taiwan's reimbursement rate adjustment based on chemical generic grouping in 2001. This research also demonstrates the use of micro-level longitudinal data to generate policy-relevant information. The research can be used to improve efficiency of health care resource use. Methods We chose the three most-used classes of cardiovascular drugs for this investigation: beta blocking agents, calcium channel blockers mainly with vascular effects, and plain ACE inhibitors. For each drug class, we investigated changes in daily expense, consumption volume, and total expenditures from a pre-action period to a corresponding post-action period. We compared an exposure or "intervention" group of patients targeted by the action with a comparisonor "control" group of patients not targeted by the action. The data sources are a longitudinal database for 200,000 NHI enrolees, corresponding NHI registration data of health care facilities, and an archive recording all historical data on the reimbursement rates of drugs covered by the NHI. We adopted a fixed effects linear regression model to control for unobserved heterogeneity among patient-hospital groups. Additional descriptive statistics were applied to examine whether any inappropriate consumption of drugs in the three classes existed. Results The daily drug expense significantly decreased from the pre-action period to the post-action period for the exposure group. The average magnitudes of the decreases for the three classes of drugs mentioned above were 14.8%, 5.8% and 5.8%, respectively. In contrast, there was no reduction for the comparison group. The number of days of the prescription increased significantly from the pre- to the post-action period for both exposure and comparison groups. The total expense also significantly increased for both patient groups. For the exposure group, the average magnitudes of the growth in the total expenditure for the three classes of drugs were 47.7%, 60.0% and 55.3%, respectively. For the comparison group, they were 91.6%, 91.6% and 63.2%, respectively. After the action, approximately 50% of patients obtained more than 180 days of prescription drugs for a six-month period. Conclusion The 2001 price adjustment action, based on generic grouping, significantly reduced the daily expense of each of the three classes of cardiovascular drugs. However, in response to this policy change, hospitals in Taiwan tended to greatly expand the volume of drugs prescribed for their regular patients. Consequently, the total expenditures for the three classes of drugs grew substantially after the action. These knock-on effects weakened the capability of the price adjustment action to control total pharmaceutical expenditures. This means that no saved resources were available for other health care uses. Such expansion of pharmaceutical consumption might also lead to inefficient use of the three drug classes: a large proportion of patients obtained more than one day of drugs per day in the post-action period, suggesting manipulation to increase reimbursement and offset price controls. We recommend that Taiwan's government use the NHI data to establish a monitoring system to detect inappropriate prescription patterns before implementing future policy changes. Such a monitoring system could then be used to deter hospitals from abusing their prescription volumes, making it possible to more effectively save health care resources by reducing drug reimbursement rates.