A novel approach to medicines optimisation post-discharge from hospital: pharmacist-led medicines optimisation clinic

One strategic target in health care systems is to reduce the frequency of unplanned re-hospitalisations, which have progressively increased in rate and cost burden worldwide. Since optimising medicines use is a key process in effective disease management, pharmacists have a significant role to play in patient care after discharge from hospital.

In the United Kingdom £17.4 billion was spent on medicines during 2016/17 [ 1]. Importantly, however, a high percentage (33–50%) of medicines prescribed for long term conditions are not taken as recommended, with only 16% of newly prescribed medicines taken correctly [ 2– 4]. Further to this, one in eight patients has a medicine related problem [ 5]. Certain patients are at high risk of having difficulties in managing their medicines when discharged from hospital which can often result in the patient being readmitted [ 6]. Although care is taken to support patients managing their medicines and medical conditions prior to discharge [ 7– 9], they often face significant challenges when they return home.

It is recognised that performing medication reconciliation alone (10) and isolated discharge planning efforts cannot resolve all problems (11) nor can isolated discharge support (12). Evidence has shown that positive effects are mainly observed when interventions from the discharge planning and post-discharge support perspectives are combined across the hospital-home interface (10, 13). However, further well-designed and carefully conducted studies to improve the effectiveness of healthcare delivery for recently discharged patients are required in this area (10, 14).

In order to perform the present research, a collaborative team developed an outpatient medicines optimisation clinic (MOC), led by clinical pharmacists, to provide support post-discharge to patients deemed to be at risk of medicine related problems. The hypothesis was that patients at risk of medicine related problems, could benefit from outpatient follow-up by hospital based clinical pharmacists, building on the Integrated Medicines Management services [ 7, 9] provided to hospital inpatients at the study site hospital.

The aim was to assess the impact of a pharmacist-led MOC on both patient and health care system outcomes, when delivered to patients at high risk of medicine related problems after a period of hospitalisation.

Ethics approval for the study was obtained from the Office for Research Ethics Committees Northern Ireland (IRAS 11/NI/0127)/Clinical Trials.Gov registration no: NCT01534559.

The study was designed as a parallel group, randomised, controlled trial with 1:1 allocation ratio. Patient recruitment was conducted in the respiratory and cardiology wards of Antrim Area Hospital, a 426-bed district general hospital in Northern Ireland.

Adult patients (≥ 18 years old) who were admitted into the study wards, as acute/unscheduled medical admissions, who met at least one of the inclusion criteria were invited to participate and provide written informed consent (Box 1).

During their hospital stay, recruited patients within both groups completed three self-reported questionnaires: Medication Adherence Report Scale (MARS), Patient Beliefs about Medicines Questionnaire (BMQ) and health-related quality of life (HRQOL) questionnaire (EQ-5D-3L). These questionnaires were mailed to participating patients for self-completion at 30, 90, 180 and 365 days post-discharge. Complete questionnaires were mailed to the researcher for analysis.

This time bound (12 month) pilot study evaluated the impact of the new clinic service. Patients were screened for eligibility in participating wards before being invited to participate. During patient enrolment, concealed allocation to a control or intervention group was achieved through a closed envelope system [ 10] prepared by an independent investigator. Block randomisation with random block sizes, ensured allocation balance and avoided selection bias by preventing allocation prediction [ 11]. A sample size calculation was not carried out for this pilot study. Researchers and pharmacists were blind to the block size sequence and randomisation envelopes were unopened until after patient recruitment. Patients were blinded to the primary outcome measure.

The MOC was patient centred, focused on the patients’ needs and was conducted at the study-site hospital within 2 weeks of discharge, in a semi-structured format using a bespoke Medicine Review (clinical; level 3) Questionnaire. The clinical pharmacist ensured that each patient received the following services at the clinic: medicines reconciliation; lab test review; medicines review; general patient education (including medicine purpose, dose, timing, potential side effects); lifestyle advice, as appropriate (e.g. physical activity, diet, weight management, smoking cessation, secondary prevention advice); assessment of medication adherence, disease management advice and self-management advice as required. At the end of the appointment, the clinical pharmacist agreed next steps with the patient, and documented key points on a tailored Take Home Action leaflet for the patient to take away. After this first outpatient clinic visit, based on the individual patient’s needs, he/she was invited to a further follow-up (either at the clinic or by telephone) within 6-8 weeks of discharge. A summary of each review was forwarded to the patient’s GP and appropriate hospital medical consultant, which detailed a list of recommendations from the pharmacist. Control patients received the normal post-discharge care, which did not involve any follow-up by a hospital based clinical pharmacist.

The primary outcome measure was 30-day unplanned readmission to hospital. Secondary outcomes included: readmission within 7, 14, 90, 180 and 365 days, time to hospital readmission and length of hospital stay during first readmission. Unplanned GP consultations (including out-of-hours) and Emergency Department (ED) visits over the study follow-up period were recorded. Information regarding healthcare resource usage was obtained from the hospital and corporate information systems by the researcher. Adherence, beliefs about medicines and HRQOL were assessed at baseline and at 30, 90, 180 and 365 days. Patient satisfaction for the intervention group was assessed after the delivery of the service. Finally, a cost–benefit analysis was performed [ 12, 13]. Analysis was carried out from hospital perspective over a 1 year follow-up period. All relevant costs were obtained from the corporate department of the hospital.

Data were entered and analysed using SPSS version 24 for intention-to-treat (ITT) and per-protocol (PP) analyses. Alternative software (MedCalc ^®) was utilised when SPSS was not appropriate and standard statistical procedures were applied. Survival analysis using Kaplan–Meier curves [ 14, 15] was used to examine time to readmission. The difference in frequencies of multiple readmissions, unplanned GP consultations and ED visits were analysed using the Mann–Whitney U Test [ 16], which was also used to test differences in length of hospital stay during the first readmission. Relative risk and number needed to treat were calculated when appropriate [ 17]. Prediction models for number of multiple readmissions, ED visits and unplanned GP consultations were carried out through the development of a generalised linear model, Poisson regression [ 18, 19]. All missing data was appropriately recorded in the SPSS database prior to analysis.

The recommended standard method for analysis of data collected in a RCT is the ITT approach as emphasised by the CONSORT guidelines [ 20, 21]. Nonetheless, this analysis has a number of drawbacks [ 22]. In the present study due to a fixed recruitment period, the work should be considered a pilot study rather than a definitive study. Since the intention was to determine the efficacy of the MOC intervention, PP analysis was carried out alongside an ITT analysis.

In the present study, a reduction in the 30-day readmission rate was noted (16.1% reduction in the PP assessment and a 9.6% reduction in the ITT assessment). Since this is the first evaluation of a new model of care delivery, there were no published data with which to compare the results. These data, however, compare favourably with a study in which pharmacist-led face-to-face discharge and telephone follow-up resulted in a 30-day readmission rate reduction of 14.2%; P = 0.01 [ 23]. In another study, pharmacist participation in follow-up appointments had a positive impact and led to a significant reduction (10.2%; P = 0.023) in 30-day readmission rates [ 24], where the control group involved matched patients from the previous year.

This study showed that clinic attendance decreased both ED visits and unplanned GP consultations. In a recent systematic review, which included 19 randomised controlled trials [ 25] it was estimated that in-hospital clinical pharmacist-led services (e.g. medicine review, medicine reconciliation, adherence assessment, therapeutic education, discharge counselling) resulted in a significant 30% reduction (risk ratio was 0.70, 95% CI 0.59–0.85, P = 0.001) in emergency department visits post-discharge. It should be noted, however, that many of these latter components are already provided in the Integrated Medicines Management (IMM) programme [ 7– 9] that is delivered as routine practice to patients in the study site hospital by clinical pharmacy staff, i.e. the present results demonstrate additional authenticity to the IMM program. Variance in rates will exist due a greater throughput of older more comorbid patients who may not all have received the IMM service.

The positive impact of the provision of pharmacist-led interventions on HRQOL has been demonstrated by other researchers in a range of studies, utilising both generic and specific instruments [ 26– 29]. A number of studies employed the EQ-5D tool. Some of the later studies confirmed the positive impact of pharmacist interventions [ 30, 31], while a number of studies reported no significant impact [ 32, 33]. In a recent systematic review [ 34] the authors concluded that pharmacist care can significantly improve at least one dimension of HRQOL, usually the pain related dimension, i.e. similar to the present findings.

Improvements in beliefs about medicines as a result of pharmacist interventions are consistent with previously published studies [ 35, 36]. However, the present research showed that the core improvement was realised via the concern scale rather than the necessity scale. The maximum impact was seen within 90 days but continued impact was observed at 12-months post-discharge. The results for patient scores in the MARS indicated that self-reported adherence levels were high across both groups over the study period. The level of patient satisfaction with the MOC was generally high indicating the high acceptability by patients of extended post-discharge services noted by others [ 37– 39].

This is the second proposed model of post-discharge care put forward by this research team; the other being a post-discharge telephone follow-up [ 40]. The need for both services exist because some patient/problems need a face-to-face interaction in order that an appropriate solution can be achieved. More work is required in this area to determine the criteria that would direct patients to the most appropriate service. Further work could also incorporate a comparison of the hospital-based model versus the community pharmacy based Ensing model to determine which would be more cost effective [ 41– 45].

The implementation of a pharmacist-led MOC post-discharge had a positive impact on unplanned readmissions, multiple readmissions, utilisation of ED visits and GP consultations. Positive cost–benefit results and patient centered humanistic outcomes including beliefs about medication, HRQOL and satisfaction resulted from the intervention. This novel approach to post-discharge care should be examined in a large, multi-centre trial to evaluate its impact on patient care.