Abstract
BACKGROUND: A prescribing monitoring policy (PMP) was implemented in November 2015 in Anhui province, China, the first province to pilot this policy to manage the use and costs of select drugs based on their large prescription volumes and/ or costs in hospitals. This study evaluated the impact of PMP on the use and expenditures of different drugs in three tertiary hospitals in Anhui. METHODS: We obtained monthly drug use and expenditures data from three tertiary hospitals in Anhui (November 2014 through September 2017). An interrupted time series (ITS) design was used to estimate changes in defined daily doses (DDDs per month) and drug expenditures (dollars per month) of policy-targeted and non-targeted drugs after PMP implementation. Drugs were grouped based on whether they were recommended (recommended drugs) by any clinical guidelines or not (non-recommended drugs), or if they were potentially over-used (proton pump inhibitors, PPIs). RESULTS: After PMP, DDDs and costs of the targeted PPIs (omeprazole) declined while use of non-targeted PPIs increased correspondingly with overall sustained declines in total PPIs. The policy impact on recommended drugs varied based on whether the targeted drugs have appropriate alternatives. The DDDs and costs of recommended drugs that have readily accessible appropriate alternatives (atorvastatin) declined, which offset increases in its alternative non-target drugs (rosuvastatin), while there was no significant change in those recommended drugs that did not have appropriate alternative drugs (clopidogrel and ticagrelor). Finally, the DDDs and costs of non-recommended drugs decreased significantly. CONCLUSION: PMP policy impact was not the same across different drug groups. PMP did help contain the use and costs of potentially over-used drugs and non-recommended drugs. PMP did not seem to reduce the use of first-line therapeutic drugs recommended by clinical treatment guidelines, especially those lacking alternatives; such drugs are unlikely appropriate candidates for PMP.