A pre- and post-intervention study was undertaken at a 790-bed public health service in Melbourne, Australia over two time periods (December 2017 to July 2018 and January 2020 to July 2020). The intervention was introduced during a one-month run-in period (December 2019) prior to commencement of the post-intervention period. The methodology for the pre-intervention evaluation of usual care has been detailed elsewhere . The methodology below focuses on the post-intervention period, but briefly outlines key details of the pre-intervention study.
All patients who had undergone a TKA and were supplied with an opioid upon discharge, were included in this study. Consecutive patients were invited to participate by the clinical pharmacist and were required to provide verbal consent. Patients’ caregivers were also required to provide consent if they were responsible for the management of the patient’s medications, as they would need to be interviewed. Patients were excluded if they required an interpreter, were readmitted to hospital within 3-weeks of hospital discharge or were unable to be contacted after three attempts.
Hospital Discharge Process – usual care
The hospital discharge process was identical in both the pre- and post-intervention periods. When patients were confirmed for discharge, a hospital prescription was electronically generated by the hospital medical officer (HMO) using Cerner Millennium. The prescription was given to the clinical pharmacist who undertook medication reconciliation and discussed the prescription with the patient, reviewing their pain control and opioid requirements. Any prescription changes proposed by the pharmacist were discussed with the HMO. The medications were then dispensed by the hospital pharmacy and provided to the patient along with medication education prior to discharge.
Post-discharge Pharmacist Review - intervention
The post-discharge pharmacist review was undertaken by one of three HOMR pharmacists employed by the health service. The HOMR pharmacists had between 15 to 30 years of experience. Intervention patients were contacted via telephone by the HOMR pharmacist approximately three to five days after hospital discharge. The pharmacist ascertained the current analgesics and how they were being used by the patient, and explored the use of any non-pharmacological management strategies. They provided education and advice to the patient about optimizing their analgesics (e.g. ensuring non-opioids such as paracetamol were used at maximal doses), incorporation of non-pharmacological strategies, expectations about the duration of opioid use and when (if required) to review their analgesics with their GP. In addition, pharmacists could refer patients to the hospital’s Peri-Operative Pain Management, Education and De-escalation (POPPMED) clinic if they felt that specialist review was required. The POPPMED clinic was serviced by an anesthetist and two specialist pain consultant nurses. Patients could also be referred to the clinic during their inpatient stay by HMOs and pharmacists if they were deemed to be at high risk for developing chronic opioid use. A written report outlining the patient’s analgesic usage and recommendations made by the pharmacist was provided to the patient’s GP (via facsimile) and scanned into the hospital medical record.
During the pre- and post-intervention periods, patients or their carers underwent a telephone interview 3-weeks after hospital discharge. Interviews were conducted by a pharmacist not involved in the patient’s care. The same questionnaire was used during the pre- and post-intervention periods, to allow comparison between the patient groups . Information obtained from the interview included i) the name, dose and frequency of analgesics currently used, ii) whether refills had been obtained for opioids, iii) whether the patient felt they had an adequate supply of hospital provided opioids, and iv) the number of unused opioid pills remaining from the initial hospital supply. In addition, post-intervention patients were asked whether they recalled receiving a review from a pharmacist after hospital discharge and their level of satisfaction with the service on a scale of 1 to 5 (with 1 being extremely dissatisfied and 5 being extremely satisfied).
Opioid naïve patients who commenced on opioids in hospital and were still taking them at the 3-week post-discharge interview, were telephoned at 3-months after discharge to determine whether they were still taking opioids (and had therefore transitioned to prolonged use).
Patient demographics, length of hospital stay and documented past medical history (according to the admission notes) were obtained from the hospital’s electronic medical record. The number and type of pre-admission medications were obtained from the hospital’s Medication History on Admission Form completed by a clinical pharmacist. Patients were deemed to be opioid naïve if they were not taking any opioids immediately prior to their hospital admission (regular or prn [“as required”]). Analgesics and quantities supplied on discharge were obtained from the hospital pharmacy dispensing system. An analgesic was defined as any medication used to treat or prevent pain . Between the pre- and post-intervention periods, ascorbic acid (to prevent complex regional pain syndrome) was removed from the orthopedic unit’s pain management protocol and therefore data from the pre-intervention period was amended to exclude ascorbic acid use from these results. Analgesics with multiple forms were considered individually (e.g. short and long acting agents), whereas multiple strengths of the same medication in the same dose-form were considered as one. Opioids were converted to an analgesic equivalent dose of oxycodone and 5mg oxycodone was deemed equivalent to one opioid pill .
Reports prepared by the pharmacist were retrospectively reviewed by a study investigator (TT). A descriptive analysis was undertaken of the types of recommendations and advice provided by the pharmacist.
The primary endpoint of this study was the percentage of patients who were still taking opioids 3-weeks after hospital discharge. Secondary endpoints were:
Number of hospital-supplied opioid pills remaining 3-weeks after discharge;
Percentage of patients needing to obtain an opioid refill prescription from their GP;
Percentage of patients who felt that they had adequate supplies of opioids upon hospital discharge;
Percentage of opioid naïve patients who became chronic opioid users (continued usage beyond 3-months post-discharge);
Percentage of patients who could recall having their medications reviewed by the pharmacist;
Patient satisfaction with the post-discharge pharmacist review;
Recommendations and advice provided by the pharmacist.
Sample Size Calculation
During the pre-intervention period, 63 patients underwent a TKA and were supplied with an opioid on discharge; 73.4% were still taking an opioid at 3 weeks post discharge.9 If a clinically significant reduction in the percentage of TKA patients taking an opioid at 3 weeks is a one-third reduction (to 49.2%), then at least 62 patients were needed in each group (80% power, 2-sided alpha 0.05). We therefore aimed to recruit the same number of patients (n = 63) to the post-intervention group. This study was terminated early after only 44 patients were recruited, because elective surgery at the study hospitals was cancelled for a prolonged period due to the COVID-19 pandemic. A post-hoc power calculation whereby 29.5% of post-intervention patients were still taking an opioid at 3 weeks revealed that the study had 99.7% power (alpha 0.05) to detect a difference between the groups if one existed.
Statistical analysis was undertaken using SPSS version 26 (IBM Corporation, Armonk, NY). Median and interquartile range (IQR) were used to describe non-normally distributed data, whereas the mean and standard deviation (SD) were used to describe normally distributed data. Chi-square test was used to compare differences in percentages. Student’s t test was used to compare means, while Mann-Whitney U test was used to compare distributions of non-normally distributed data. A p value of less than 0.05 was considered to be statistically significant for all comparisons.