Executive Summary

1. Background
Over the past decade there have been moves to extend the traditional role of the community pharmacist from the supplier of medicines to pharmaceutical carer. With pharmaceutical care, pharmacists take responsibility for the management of the patient's medicines and their associated drug-related needs. If community pharmacists can contribute to improved health outcomes in coronary heart disease (CHD) through medicines management, this might provide a cost-effective method of addressing a national priority.

2. Aims and Objectives
The aim of this study was to evaluate the introduction of the Medicines Management Service by community pharmacists for patients with CHD.

3. Method

3.1 Study design
The study was a randomised controlled trial. The primary subjects for the study were patients with CHD identified from general practice computer systems for example, history of coronary artery-bypass graph (CABG) or angioplasty as well as diagnosis of myocardial infarction (MI) and angina after cardiovascular diagnoses. Secondary subjects were health care professionals caring for recruited patients (i.e. GPs and their staff (Practice Pharmacists, Practice Managers and Practice Nurses). The main data source was routinely held information in the patients' medical records. Data for both intervention and control patients was obtained from these records by audit clerks. All participants were sent postal questionnaires at baseline and follow-up. Qualitative interviews were undertaken with a sample of patients, community pharmacists, and General Practitioners (GPs), and focus groups were undertaken with a sample of community pharmacists.

3.2 The Medicines Management Intervention
The intervention comprised an initial consultation with the pharmacist who determined the number of subsequent consultations that would be appropriate based on the needs of each patient. It was anticipated that there might be up to three subsequent consultations within the 12-month follow-up period. Appropriateness of therapy, compliance and concordance, lifestyle and social and support issues were to be addressed. Training for community pharmacists prior to the delivery of the intervention was designed and provided Centre for Pharmacy Postgraduate Education (CPPE).
The participating community pharmacists were informed of their patients who had been allocated to the intervention group and received a core set of information about each patient. Patients in the control arm of the study continued to receive normal care.

3.3 Primary and Secondary Outcome Measures
The effectiveness and efficiency of the intervention was assessed using primary and secondary outcome measures derived from data in the medical records, information on the intervention form, and patient and health care professional questionnaires at baseline and follow-up. Interview and focus group data were also used.
The primary outcome measures were appropriate treatment (derived from the National Service Framework (NSF), health status, and an economic evaluation. The secondary outcome measures were patient risk of death, role of over the counter medicines (OTCs), and satisfaction, experiences and attitudes of patients and health care professionals. All measures were assessed at baseline and 12 months.

3.4 Analysis
Basic descriptive statistics and regression modelling were conducted. The models compared the intervention with the control group to produce adjusted odds ratios (OR) or adjusted mean differences and corresponding 95% confidence intervals (CI). Qualitative data were analysed using standard methods, which included the identification of themes emerging from the interviews.

4. Results

4.1 Recruitment
The project involved 1493 patients (980 intervention, 513 control), 62 pharmacists and 164 GPs.

4.2 Primary Outcomes
4.2.1 Drug therapy
There were no statistically significant differences between intervention and control groups at follow-up for any of the NSF-related outcome measures. The main results are reported below.

At baseline, over 90% of patients were receiving aspirin, anti-coagulant or anti-platelet treatment (referred to as aspirin-related management). At follow-up 94% of both intervention and control groups were receiving aspirin-related management; there was no significant difference between the groups.

At baseline, 59% of intervention and 57% of control patients were on target total cholesterol levels. At follow-up, these proportions were 58% and 55%, respectively. For both groups, the proportion of patients with a recorded serum cholesterol measurement during the 12 months prior to baseline was high (>90%). At follow-up this figure decreased to less than 80% with both groups. There were slight increases in the proportion of patients on statin therapy in both groups

The majority (>90%) of patients in both groups had a blood pressure measurement recorded during the 12 months preceding both baseline and follow-up. At baseline, 47% of intervention and 43% of control patients had achieved the blood pressure management target (i.e. systolic 140mmHg and diastolic 85mmHg). At follow-up, these proportions were 49% and 47%, respectively.

At baseline, approximately half the post-Myocardial Infarction (MI) patients in both groups were receiving ß-blockers. No statistically significant difference was found between groups at follow-up. A global score for appropriate management was calculated which ranged from 0-7, with the higher the score reflecting the most appropriate management. At baseline, this was similar in both groups (intervention 4.3; control 4.2), increasing to 4.6 in both groups at follow-up.

4.2.2 Lifestyle
Patients were also assessed for the lifestyle targets of the NSF; smoking status, alcohol consumption, exercise and diet. No difference was demonstrated for any of the measures between intervention and control groups at baseline or follow-up.

4.2.3 Quality of life
Quality of life was measured using the SF-36 and EQ-%D generic measures. There were no differences between the groups in adjusted mean levels of the SF-36 domains nor in the overall EQ-5D at follow-up.

4.2.4 Economic evaluation
The total NHS cost increased between baseline and follow-up in both groups but to a greater extent in the intervention group (£738 to £835 in control, an increase of £97; £852 to £971 in intervention, an increase of £119 per patient). The greater cost in the intervention group largely reflects the cost of the training. The difference in cost was statistically significant (p=0.001).

4.3 Secondary Outcomes

4.3.1 Risk of death
No difference was shown between the groups in adjusted mean risk of death score at follow-up.

4.3.2 Role of OTC Medicines
Eighty-eight OTC medicines were recorded for 63 patients in the intervention group at baseline. The most commonly used OTC medicines was aspirin, used by half the OTC users.

4.3.3 The medicines management process
Sixty community pharmacists consulted with patients, and saw between one and sixty two patients each. They identified at least one medicines management issue, or made recommendations, for 89% of patients. For 738 patients, there were 2337 issues noted or recommendations made. The median issues noted and recommendations made was three per patient. The greatest proportion of issues identified was related to monitoring (32%) (mostly with statins), followed by the need for treatment (i.e. the introduction of statins). The need for lifestyle changes accounted for approximately 10% of issues.

As an addition to original protocol, an in-depth study of issues noted and recommendations made was undertaken by research reviewers on a systematic sample of patients seen (169/739, 23%). This indicated that the majority of issues noted and recommendations made were appropriate. For the three categories that were the main focus of the intervention, 42% of possible issues were documented by study pharmacists. There was considerable variation across community pharmacists in the proportion of possible issues documented. Fifteen percent of community pharmacists identified 50% or more possible issues.

4.3.4 Patients' perspectives: satisfaction, experience and attitudes ¹
Response rates to the patient questionnaire were 95% at baseline and 85% at follow-up. The majority of respondents was over 65 years (76%); male (69%), and white (99.6%). The median consultation time of 30 minutes was reported to be "About right" by the vast majority (98%). The consultation was often held in a separate room in the pharmacy (66%) which permitted private discussion (>90%). The in depth interview data however, indicated there may be some concerns about privacy. Twenty-eight patients reported a formal follow-up consultation.

Most patients were satisfied with the service from the community pharmacist, and satisfaction with the most recent visit to the community pharmacist improved for intervention compared with controls. At baseline, both self reported and pharmacist assessed compliance was high. At follow-up, intervention patients compared with controls were: more likely to agree that knew more about their medicines (73% vs 65%), less likely to prefer to see their GP (76% vs 85%), and more likely to find it easier to talk to the community pharmacist (32% vs 18%). The interview data showed that patients liked the reassurance, check on medicines, and lifestyle advice during the consultation, although a minority felt it had been a waste of time and would have preferred to see their GP.

4.3.5 Community Pharmacists experiences of and attitudes to the Medicines Management Service
The response rate to the community pharmacist questionnaire was 77% at baseline; and 53% at follow-up. The majority of respondents was female (63%). Of these, 30% had a postgraduate qualification and 9% had worked in a medical practice. At baseline 64% were satisfied with their job. This increased to 69% at follow-up with indications that this was due to: more freedom; more responsibility; more variety; and recognition for good work. Taking part in the intervention increased their belief that a holistic approach is part of medicines management.

The majority of community pharmacists believed that confidential discussions and access to medical records were required for medicines management to be provided, and that medication review includes all medicines. The majority delivered the service in their own time (83%) and from their usual pharmacy (87%). Although 61% of community pharmacists believed they had received sufficient information to provide the service, many (67%) had contacted the GP for more information. At the end of the project almost all pharmacists felt that being involved had been a positive experience. Taking part in the project increased their perceived knowledge of CHD and clinical pharmacy, and made them more likely to discuss prescribed medicines with patients. They felt that changes were required to their current working practice if the service was to be sustainable. In formal evaluation, basic knowledge of CHD was high at baseline although only a small number felt confident in their own knowledge and skills. The training was rated of 'medium to high' usefulness by the majority; and perceived to be sufficient to allow them to deliver the service. After training there was an increase in perceived knowledge and skills and after conducting medicines management interventions, there was an increase in comfort with delivering the service.

4.3.6 General Practitioners' experiences of and attitudes to the medicines management service
The response rate to the GP questionnaire was 70% at baseline and 52% at follow-up. Sixty one percent of GPs had a pharmacist working in their practice. At baseline, most GPs (77%) were positive about a community pharmacy-led Medicines Management Service. This largely persisted at follow-up, although there were some small detailed changes in attitudes. GPs believed that medicines management would lead to more appropriate medicine and fewer problems with multiple therapy. The qualitative data demonstrate more ambivalent findings.

5. Discussion
Important strengths of the study included: the design and quality assurance procedures, the use of standard validated instruments, good participant response rates, and 'blind' analysis of trial results. Limitations included choosing a high profile condition, limited information on whether recommendations were implemented by the GP, potential bias from the use of self-reported questionnaires and the self selected nature of the participants.

The primary objective was to assess the proportion of patients with CHD who received appropriate secondary prevention treatments, or who had beneficial changes in CHD-related lifestyle, quality of life, or future risk of death. There was no difference between the intervention and control group for any of these measures. The absence of effect might have been due to ceiling effects in some of the main outcome measures (e.g. aspirin use) in a population that was already receiving a high level of care for example statin use was very high at baseline or because of inconsistencies in the application of the intervention. In-depth analysis of the issues identified by community pharmacists suggests these were appropriate, although many issues were missed.

Community pharmacists were very positive about the service. They believed it had improved their skills and knowledge, and that they had made a positive difference to their patients. They recognised that many of their recommendations had been relatively minor and felt that access to patients' medical records was necessary to deliver the service to full effect. This was also agreed by the GPs although they had concerns about the practicality and desirability of this for reasons of patient confidentiality.

Patients reported they were highly satisfied with the Medicines Management Service, although the detailed interviews revealed concerns which may need to be addressed.

On the basis of the questionnaire and interview data, a range of barriers to the introduction of a community pharmacy-led Medicines Management Service were identified. These have generally been supported by findings in other studies and the data suggest these need to be addressed before effective collaboration can be achieved. Barriers within the pharmacy setting include: time and support, remuneration, the provision of privacy to the patient, limited access to medical records and a reluctance on the part of community pharmacists to document their activities. Patient barriers include a prevailing preference for medical care, and a general lack of understanding of the skills and roles of community pharmacists. GP barriers included professional boundary concerns, lack of complete trust in community pharmacists and concerns over the privacy of patient medical records.

6. Conclusions and Recommendations
The results of this study, according to the agreed a priori objectives and outcome measures are as follows:

Primary Objectives
There was no change in the proportion of patients receiving appropriate medication as defined by the NSF. There was no change in health as defined by the SF-36 and EQ 5-D. The Medicines Management Service was more expensive than standard care.

Secondary Objectives
Patients were satisfied with the service and satisfaction with their most recent pharmacy visit was statistically significantly higher in the intervention compared with the control group. Patients in the intervention group also received significantly more information from the community pharmacist about side effects of their medication, their health and had statistically significantly better understanding of how to take their medicines compared with controls.

GPs remained broadly supportive of the service but after the intervention almost all of them (98%) believed the community pharmacists needed access to medical records to deliver a medicines management service.

Community Pharmacists felt more knowledgeable about CHD and clinical pharmacy in general and reported being more likely to discuss prescribed medicines with patients. Fifty percent of community pharmacists also reported that their relationship with GPs had improved.

The overall levels of pharmacist activity were similar to those reported in other studies, and individual pharmacists in this study recommended optimal treatment. There is a clear need to understand the reasons why an effective service was not universally delivered and to learn from the examples of good practice.

Our recommendations are:

• The model of community pharmacy-led medicines management needs to be developed and refined, and the exact process and support mechanisms reviewed.
• Any service needs to have local ownership, particularly by GPs. There has to be local development, with models varying between different community pharmacies in recognition of differing skills, settings and approach to practice.
• Current barriers in terms of the organisational environment, infrastructure, skill mix, access to records and remuneration need to be addressed.
• Where patients are already on appropriate treatment and optimum therapy, the objectives of a community pharmacy led Medicines Management Service of improving appropriateness of therapy may need to be reconsidered. Better informed, empowered patients may be a more realistic objective