This study was a single-site, retrospective cohort study. A retrospective review of medical records was conducted for all patients who attended the PREVENT clinic between 1 January 2018 and 31 December 2019 at an Australian metropolitan hospital. The group of case-matched control patients did not attend the clinic and received usual care at the same hospital.

The PREVENT clinic commenced in September 2016 to facilitate timely post-discharge follow-up by a pharmacist targeting patients at risk of medication misadventure, [23, 24] meaning those at risk of medication misuse or medication harm and early readmission to hospital. The PREVENT clinic service was modelled on the Home Medicines Review (HMR) service in Australia, which aims to support the quality use of medicines, reduce MRPs or healthcare use and help patients better understand their medicines [25,26,27].

Patients were referred to the PREVENT clinic by a member of the inpatient treating team who were identified to be at risk of medication misadventure. A member of the treating team included a doctor, pharmacist, nurse, or other health professional. The key patient criteria for referral to the PREVENT clinic included taking five or more regular medications, taking a high-risk medication, one or more medication changes during admission, identified to be non-adherent with medication therapy or having difficulty managing their medicines, and/or requiring further education and follow-up. High risk medications were defined as medicines that have an ‘increased risk of causing significant patient harm if they are misused or used in error’ as defined by the Australian Commission on Safety and Quality in Health Care [28].

At the clinic appointment, a senior hospital pharmacist undertook a comprehensive medication review including, confirmation of medication history and medication understanding, reconciliation of current medicines with the recent hospital discharge medication record (DMR), review of each medication for safety and effectiveness, as well as identified and rectified MRPs. These MRPs were resolved directly with the patient through education or negotiation, implementation of dose administration aids if indicated, referral to another health practitioner (e.g. diabetes educator) or sent to the patient’s general practitioner (GP) via facsimile letter to optimise their medicines or disease state management. These processes are based on professional practice standards [22, 29]. Each interaction with the patient was individualised based on their needs.

The clinic appointment occurred in the hospital outpatient department either face to face, by phone or telehealth as close to the time of discharge as possible, ideally within 14 days. Each initial appointment consisted of a 45 to 60 min consultation with the patient, with documentation and GP follow-up completed after each consultation. Further follow-up with individual patients could be arranged as needed. For the purposes of this study, only the initial post-discharge review consultation has been analysed.

No post-discharge care was provided by the clinic pharmacist at the study hospital. All patients received standard inpatient pharmaceutical care services as outlined in professional practice standards [22, 29]. This included undertaking a best possible medication history, medication reconciliation between the medicines taken at home and current inpatient medication chart and treatment plan, medication review of all active medical problems and treatments, identification and resolution of MRPs, discharge medication reconciliation, provision of a DMR and patient education [22, 29].

The primary outcome measured was unplanned, all-cause 30-day hospital readmission, which is internationally recognised as a key outcome of interest [15, 16, 30, 31]. Secondary outcomes included unplanned all-cause 30-day Emergency Department (ED) representation, a composite of all-cause 30-day hospital readmission and/or ED representation and total hospital readmissions and/or ED presentations 12 months after the index admission. The term representation is used to describe a non-admitted, unplanned hospital presentation (e.g. ED representation without admission). Day zero was measured as the date of discharge of the index admission and readmission and/or representation dates were collected from The Viewer, a statewide database of public hospital admissions and ED presentations available to hospital clinicians and GPs in Queensland. The dates taken from The Viewer accounted for an entire episode of care where patients may have transferred to or from another facility for specialist care not available at the facility of patient presentation.

The intervention group included admitted patients who were 18 years or older, lived independently and were seen in the PREVENT clinic within 30-days of discharge from hospital. Case-matched control patients were selected on a 1-to-1 basis with each patient in the intervention group based on gender, age (± 2–5 years), unit of admission, index admission discharge date (± 30–60 days), and LOS or discharge related group (DRG). These criteria were chosen as patient demographics that can contribute to the risk of readmission and were able to be defined by the hospital administration database [12, 32, 33]. Index admission is the term used to designate the admission to hospital immediately preceding the PREVENT clinic appointment or the prior admission if evidence of referral was identified in an earlier admission in the medical record. The term DRG refers to the primary diagnosis as the cause of admission.

Patients who lived in a nursing home, had previously attended the PREVENT clinic, failed to attend the PREVENT clinic appointment or readmitted to hospital prior to their PREVENT clinic appointment (and their case matched control) have been excluded. Patients identified to be a non-PREVENT clinic review or patients seen by another outpatient pharmacist clinic from the clinical notes were also excluded.

A sample size calculation for the primary outcome of unplanned, all-cause 30-day hospital readmission, was undertaken based on results from a pre-study audit of preliminary data comparing PREVENT clinic review patients with a second group of patients who had been referred to the clinic but failed to attend an appointment [34]. This produced a sample size of 185 patients per group, based on a 28-day readmission rate of 21.75% in the ‘not seen’ group vs 10.75% in the ‘PREVENT clinic’ group, using a two-sided Z-test with pooled variance with alpha of 5% and power of 80%. The difference in readmission rate observed in this audit data was similar to the 30-day readmission reduction reported in a study by Budlong et al. for their elevated, and higher risk groups [30].

Data were collected by the principal investigator (JC) between 1st July 2022 and 30th June 2023, who reviewed hospital databases and individual patient medical records. Data was entered into REDCap [35, 36], an online database system stored on the hospitals secure server. The full list of data collected and their corresponding sources is outlined in Appendix A.

Demographic data were summarised using descriptive statistics, including frequencies and percentages for categorical variables. Continuous variables have been summarised as mean and standard deviation or median and interquartile range. Independent t-test or Mann Whitney U test (if data was not normally distributed) has been used to compare means/medians between groups, respectively. Pearson Chi-squared with Fisher’s exact test has been used to compare categorical outcome measures between the two groups. P values less than 0.05 are considered significant. Missing data or missing values were excluded from the analysis.

The disease states presented are commonly compared in the literature or represent the chronic diseases targeted for referral to the PREVENT clinic in the concept brief. The pharmaceutical care received by all patients was collected from the medical record and compared to test homogeneity of usual care received. All data analysis was conducted using SPSS Statistics® software version 29.0.1.0.