Clinical and economic impact of pharmacist interventions on sampled outpatient prescriptions in a Chinese teaching hospital

This study was performed in a Chinese teaching hospital according to an order of the NHFPC [17‐19]. The period of current study lasted from January 1, 2011, to December 31, 2016. The outpatient department of this hospital is open 6 days per week and provides services to approximately 10,000 outpatients per day. Outpatient prescriptions were generated via a computerized physician order entry (CPOE) system.

The seventh prescription of every prescriber was automatic sampled by the hospital information system (HIS) every day and sent to the pharmacist workstation. Four experienced pharmacists reviewed these prescriptions. This task was the routine work of the Department of Pharmacy. If a prescription was considered to be potentially IP by pharmacists, the issue was communicated to the prescriber via telephone. If the prescriber accepted the intervention and modified the prescription, the previous prescription was judged as an IP. The information about IPs was recorded on a designated form, based on the standards of the NHFPC [17‐19]. The pharmacists completed the forms daily as per the specific categories. All of these jobs took pharmacists four hours per working day. An appointed pharmacist summarized and analyzed the information monthly and sent it to HPAC and the prescribers by office automation system, which required another 16 h per month. The same appointed pharmacist presented and educated about the IP monthly to the medical staff, which required another 2 h per month.

With reference to ASHP guidelines on a standardized method for pharmaceutical care [20], inappropriate issues were divided into 13 categories (Table 1).

Cost-benefit analysis was performed to estimate the economic benefits of the pharmacist intervention. Cost was defined as the expenses of pharmacist time [15], which included time for prescription reviewing, intervention, summarizing, analyzing and education. Salary of individual pharmacists is different due to their titles of the professional positions. So we used the average hourly salary of pharmacists, which was calculated based on the annual salary and working hours of a regular pharmacist in our hospital. Other factors such as salary promotion could potentially affect the results. The average hourly salary of a pharmacist was calculated based on the annual salary and working hours of a regular pharmacist in our hospital [15]. Benefit was estimated through cost savings, which was defined as the potentially avoidable cost of inappropriate issues intervened in by pharmacists before the medication was dispensed to the outpatient [15]. Benefit was equal to the sum of the expenses of inappropriate issues. The price of drugs that we used was the price when the prescription was made because it changed three times during our study. All of the costs were recorded in RMB and then converted to US dollars [21]. The final values are reported in US dollars. The cost-benefit analysis is expressed as benefit-to-cost ratios, which were calculated by dividing the cost of IPs by the cost of pharmacist time.

Categorical variables are presented as numbers and percentages, and continuous variables are presented as the means and standard deviations (SDs). Continuous variables were tested for normal distribution using the Kolmogorov-Smirnov test and Q-Q plots. Categorical variables were compared using the chi-square test or Fisher’s exact test and continuous variables by the independent sample t-test or the Mann-Whitney test. A two-tailed p < 0.05 was considered statistically significant. Analyses were performed with SPSS software, version 21 (IBM Corporation, Armonk, New York, USA).

The seventh prescription of every prescriber was automatic sampled by the hospital information system (HIS) every day and sent to the pharmacist workstation. Hence, prescriptions from every prescriber in different departments could be sampled. A total of 101,271 outpatient prescriptions were reviewed in this study. The number of reviewed prescriptions increased from 14,646 to 19,567 from 2011to 2016. The percentage of sampling varied from 0.67 to 0.78%. General data of the reviewed prescriptions are shown in Table 2. Approximately one-half of the prescriptions were paid at the patient’s own expense. The number of prescriptions paid by medical insurance increased slowly from 2011 to 2015 but remained less than 50% of the total reviewed prescriptions in 2015.There were no significant changes in age, sex, and expenditure type from 2011 to 2015 (P > 0.05).

With the intervention of pharmacists, the number and percentage of IPs decreased year by year. The number of IPs decreased from 1845 to 238, while the inappropriate percentage decreased from 12.60 to 1.22% from 2011 to 2016 (Fig. 1). Pharmacist interventions can optimize outpatient prescriptions gradually and continuously.

With the intervention of pharmacists, the inappropriate number and percentage in different departments were decreased annually from 2011 to 2016 (Fig. 2). The inappropriate number in the Department of Medicine was decreased from 856 to 92, while that in Department of Surgery decreased from 542 to 89. In 2016, the numbers of IPs in other departments were very few. The inappropriate rate in the Department of Surgery was decreased from 16.34 to 1.70%, while that in Department of Psychology decreased from 12.94 to 1.50%. In 2016, the inappropriate rate in the Department of Surgery and Department of Psychology was higher than that of other departments. The inappropriate percentage in the Department of Oncology was lower than in other departments.

With the intervention of pharmacists, numbers and percentages of IPs of different types decreased annually in 2011–2016 (Fig. 3). Category 1 (medications with no medical indication) was always the majority inappropriate type, although its percentage decreased annually from 2011 to 2016. Category 3 (inappropriate choice of medication) was the main type of IP in 2011–2013 but was a minority inappropriate type in 2014–2016. Prescriptions of Category 4 (inappropriate medication use) decreased annually in 2011–2016, but constituted 22.69% of IPs in 2016. The proportions of Category 5 (therapeutic duplication) were always low. With the intervention of pharmacists, inappropriate prescriptions of Category 2, 6 and 7 disappeared in 2016. Inappropriate prescriptions of Category 8–13 have not been found in 2011–2016.

From 2011 to 2016, the cost, which defined as the expenses of pharmacist time, grew from $4983.62 to $7867.52. The benefit, which expressed as total cost of all inappropriate issues, decreased from $43,500.30 to $8978.16. The benefit was always higher than the cost, and the benefit-to-cost ratio was always more than 1 (Table 3). This cost-benefit analysis showed positive economic effect for the pharmacist interventions.

The benefit-to-cost ratio was more than 1 in every payment type in 2011–2016. However, the benefit-to-cost ratio with public payments was higher than those with insurance payment and self-payment in the same year (Table 4). In 2016, the benefit-to-cost with public payments was 1.63, while that that with insurance payment and self-payment was 1.02 and 1.19.