Impacts Of Hospitals' 'Center of Excellence Initiative' Of National Health Insurance On Switching Of Antipsychotics In Outpatient Schizophrenia Patients

94

Background： On July 2004, a “Center of excellence initiative (CEI)” of hospitals was implemented by National Health Insurance (NHI). CEI granted hospitals a fixed amount of budgets in return for hospitals to fulfill its roles in caring severer patients and in improving quality of care. In response to the limited growth on budgets, would hospitals sacrifice the rights of patients in order to make more profits? Second generation (2G) antipsychotics (SGA) are more effectiveness but cost ten times more than traditional antipsychotics (1G or FGA), thus might be more likely to be replaced by less expensive drugs under CEI. Objectives： To explore how CEI might influence providers’ prescribing behaviors for outpatient schizophrenia patients. Research Questions: 1. were hospitals participated in CEI, more likely to switch 2G antipsychotic prescription after CEI? 2. Were the above switching behaviors the same among hospitals at different regions, with different ownership and accreditation levels? Materials and methods： This study, a natural experiment, was part of a Prof. Yue-Chune Lee’s project, entitled “the future direction of payment system reform of NHI” funded by the Department of Health (DOH) based “pretest-post-test controlled group design”. The study population were outpatient schizophrenia patients (ICD-9-CM codes 295) of hospitals. CEI was the intervention which implemented on July to Dec. 2004. We use Difference-in- differences” approach to compared the changes on the drug switching behavior of the interveintion and control groups before and after CEI. Because some sub-bureaus (regions) have already implemented similar payment scheme (self-management) before CEI, two reference period (Jan to June, 2004 vs. Jan-Dec,2004) were applied in the study. The intervention groups were hospitals which participate in CEI (n=98), those not were considered as the control (n=96). The screening criteria included patients who had regular source (hospital) of care in 2002-3 and have ever visit the same hospital in 2004. Patients who seek care from emergency room and of hospitals which only temporarily participated in CEI (11 hospitals) were excluded. Major source of data were various NHI claim databases, from 2002 to 2004, provided by DOH. The dependent variable was whether prescribed antipsychotics were switched. The drug switching behavior was defined as change on the prescriptions patterns from SGA to FGA or reduced dosage on SGA; including the following circumstances: 1. used monopharmacy (SGA or FGA) before and after CEI: change from SGA to FGA or reduced dosage on SGA. After CEI. 2. polypharmacy (SGA+FGA) before and after CEI: percent of reduced dosage on SGA > percent of reduced dosage on FGA.3. mix-type: change from monopharmacy (SGA) to polypharmacy with reduced dosage on SGA; or change from polypharmacy to monopharmacy with either reduced dosage on SGA or SGA replacedwith FGA. When applicable, prescription volumes of prescribed daily dose (PDD) were used to measure the average daily dosage regimen treated with antipsychotics (mg/day/person). The control variables included years, patients’ characteristics (age, sex, disease severity) and hospitals’ characteristics (ownership, accreditation level, region, and daily drug costs of patients in previous year). After data were merged and cleaned, we use logistic regression and Generalized Estimating Equation (GEE) to assess the impacts of CEI. To control selection bias due to no randomization was applied; we further developed a “Propensity Score” to reflect the likelihood of hospital to participate in CEI. The Score was the predicted value calculated from a logistic regression developed using the above hospitals characteristics (excluding daily drug cost) as the independent variables. The major findings of this study are summarized as follows： 1.There were 35,441 schizophrenia patients in the study. Those in CEI, although in general were more likely to switch drug, their changes on drug switching after CEI were not significant from that of the control group. However, the drug switching behaviors were different among hospitals with different accreditation level. 2. The changes in 2004 on likelihood of drug switching was significant compared with 2003, that might due to the 45:55 policy (OPD/IP ratio) introduced in 2004 which changed the ratio of outpatient vs. inpatient budgets from 50:50 in 2003 to 45:55 and cut 9% of outpatient budgets. Compared with Kao-Ping region, the hospitals at Eastern region were more whereas those at Taipei region were less likely to switch drug 3. The impacts OPD/IP ratio on drug switching of hospitals with different ownership varied among different region. Compared with private hospitals, military and veterans hospitals and non-profit proprietary hospitals in Taipei region were significantly more likelihood whereas hospitals affiliated with public medical schools in Central region and non-profit proprietary hospitals in Eastern region were less likely to switch drugs in 2004. 4. The likelihood of hospital transfer was higher for those patients whose prescriptions have been switched in 2004. However, the association between drug switching and hospital transfer in 2004 was not significant between those in CEI and those not. 5. Hospitals with higher daily drug costs in previous year were more likelihood to switch drug in 2004. In conclusion, our study found that the impacts OPD/IP ratio in 2004 on drug switching were greater than that of CEI; and the latter were only significant among hospitals with certain characteristics. Recommendations： Policy recommendation：(1) The payment Initiative should be carefully designed to protect patients from perverse effect (2) Drug switching measures developed in this study could be used to monitor the providers, and to develop related regulation to protect patients (3) Implementation of OPD/IP ratio should reflect the baseline differences among various hospitals. (4) Should develop guideline on the rationale switching of antipsychotics. (5) To enhance patients’ (or caregivers) drug literacy to empower them against provider’ perverse behavior. Future researches：(1) To assess the impact of other hospital factors (such as profitability from different drugs, proportion of primary care patients, patient dumpling) and physicians’ factors (age, education and training, compensation scheme) on the drug switching behavior. (2) To evaluate the impacts of drug switching on heath status, clinical outcome and quality of life of the patients. (3) To assess the appropriateness of prescription patterns of antipsychotics.