Presentation (CTI)

OBJECTIVES: Pay-for-performance (P4P) programs are expected to improve formally reported quality indicators, but they may also generate process spillover in untargeted testing and medication management. This study examined whether participation in Taiwan’s diabetes P4P program was associated with selective process spillover beyond required reporting.
METHODS: We conducted a nationwide longitudinal claims-based study of patients with type 2 diabetes. Patients newly enrolled in the diabetes P4P program were compared with matched non-participants using propensity score matching. Difference-in-differences models were used to estimate changes in annual care processes after P4P participation. Required reporting indicators were examined as benchmark measures of program response. Additional outcomes included potentially low-value or discretionary untargeted examinations outside the P4P-recommended annual diabetes examination set and medication-related behaviors, including antihyperglycemic medication adherence and prescribing of statins, ACEI/ARB, and beta-blockers. Because these cardiovascular medications are subject to indication-based reimbursement or clinical eligibility criteria under Taiwan’s National Health Insurance Administration, prescribing changes were interpreted as observed prescribing patterns rather than simple underuse or overuse.
RESULTS: Required reporting indicators increased significantly after P4P participation, consistent with the expected response to formally monitored program requirements. Responses in potentially low-value or discretionary untargeted examinations were heterogeneous rather than uniformly expansive: ECG, sodium, and uric acid testing declined; calcium and potassium showed no significant increase; and BUN testing increased. Medication-related responses also varied. Antihyperglycemic medication adherence improved modestly, while statin and ACEI/ARB prescribing declined and beta-blocker use showed no significant change.
CONCLUSIONS: Taiwan’s diabetes P4P program was associated with significant improvement in required reporting indicators, while spillover responses in untargeted testing and medication-related domains were selective rather than uniformly expansive. These findings suggest that disease-management P4P may strengthen formally monitored care without broadly increasing potentially low-value testing or pharmacologic prescribing.