Article Text

Download PDFPDF

Medication beliefs and self-reported adherence–results of a pharmacist's consultation: a pilot study
  1. Malin Johansson Östbring1,
  2. Tommy Eriksson2,
  3. Göran Petersson1,
  4. Lina Hellström1
  1. 1Department of Medicine and Optometry, eHealth Institute, Linnaeus University, Kalmar, Sweden
  2. 2Department of Clinical Pharmacology, Lund University, Lund, Sweden
  1. Correspondence to Malin Johansson Östbring, Department of Medicine and Optometry, eHealth Institute, Linnaeus University, Kalmar SE-391 82, Sweden; malin.ostbringcarlsson{at}lnu.sem


Objectives Clinical outcomes in the secondary prevention of cardiovascular disease depend on the patients’ adherence to prescribed medicines. Motivational interviewing (MI) is a patient-centred approach used to change different health behaviours. The objective of this pilot study was to explore the impact of a clinical pharmacist's consultation on beliefs about medicines and self-reported medication adherence among patients with coronary heart disease (CHD).

Methods CHD-patients participating in a prevention programme at the Kalmar County Hospital were randomised to control or intervention. The intervention consisted of a medication review focused on cardiovascular drugs, and a semistructured interview based on MI-approach, with a follow-up phone call 2 weeks later. The intervention was conducted by a clinical pharmacist at the cardiology unit 3 months postdischarge. Primary outcome measures were the results from the Beliefs about Medicines-Specific (BMQ-S) and the 8-Item Morisky Medication Adherence Scale (MMAS-8) 2 weeks after intervention.

Results 21 enrolled patients (11 intervention) all completed to follow-up. MMAS-8 was very similar in the intervention and control groups. In BMQ-S the intervention group had a mean (SD) necessity score of 21 (4) and a concern score of 12 (6), corresponding results in the control group were 21 (3) and 10 (5). However, since there was a difference in BMQ-S at baseline, seven intervention patients shifted towards more positive beliefs compared with two control patients.

Conclusions No difference was found in adherence and beliefs at follow-up. However, after consultation, a larger proportion of patients changed towards more positive beliefs compared with control.

  • Beliefs about medicines
  • Medication review
  • Coronary disease
  • Medication adherence
  • Motivational interviewing
View Full Text

Statistics from



Low adherence to pharmacological secondary prevention treatment of cardiovascular disease increases mortality.1 ,2 To improve treatment adherence in patients with hypertension, dyslipidaemia and diabetes mellitus, different complex interventions have been developed, but few have shown improvements in adherence and treatment outcomes.3 Adherence behaviour has many dimensions and an intervention to improve adherence needs to include technical, behavioural, cognitive and emotional tools; and the intervention always has to depend on the character of the specific problem.4 ,5 Reminders, simplified drug regimes, increased patient-centredness, and emphasis on the patient's perspective, all show improved adherence to lipid-lowering drugs.6

Motivational interviewing (MI) is a patient-centred approach that has been proven effective in changing different health behaviours, such as medication adherence.7

Nurses and doctors involved in the secondary prevention programme at the Kalmar County Hospital perceive that inquiries and concerns about medicines are common among their patients, and that their time and knowledge are insufficient to meet the demand. Supplementing the team with a pharmacist focusing on patients’ drug-related problems and attitudes, and evaluating the appropriateness of prescribing, could be a model for improvement.

The objective of this pilot study was to explore if a clinical pharmacist's consultation, based on MI-approach, has any impact on beliefs about medicines and self-reported medication adherence among patients with coronary heart disease (CHD).


Setting and population

A randomised controlled pilot study was set up at the cardiology unit at the Kalmar County Hospital. All patients with CHD admitted to the secondary prevention programme, visiting a cardiology nurse during November to December 2011, were asked to participate. Twenty-one patients were included in this pilot study.

The secondary prevention programme at the cardiology unit consists of a nurse consultation 2 weeks after hospital discharge, a revisit to a doctor 4–6 weeks later, and an individualised physiotherapy programme of 12 weeks. The nurse provided written and oral information about the study and obtained informed consent from each patient. Randomisation to intervention or control arm (standard care) was stratified according to gender, and performed by the pharmacist with sealed, opaque, envelopes. Patients were blinded to group allocation, and the difference between the two treatment arms was that control patients received the intervention later (figure 1).

Figure 1

Study design. The grey boxes show the steps of inclusion, randomisation and when questionnaires were sent at baseline and follow- up-1 and -2. Black boxes indicate when the intervention took place. White boxes show the steps of the standard care process.


The intervention consisted of (1) a medication review by a clinical pharmacist collaborating with the cardiologist; the review focused on cardiovascular drugs with emphasis on treatment guidelines, therapeutic goals and adverse drug reactions, (2) a semistructured interview by a clinical pharmacist based on MI-approach, with a follow-up phone call 2 weeks later. The intervention activities were performed also in the control group but after follow-up-1 (figure 1).

Intervention development

Standard operating procedures for data collection, the semi-structured interview and the follow-up phone call were produced by the pharmacist. The guide for the semistructured interview had been prepared by translation and adaption of the Drug Adherence Work-Up Tool (DRAW).8 DRAW includes questions about intentional and unintentional adherence, problems with regimen, adverse events among others, and it is followed by a guide for the pharmacist how to respond to different problems. A patient-information leaflet containing CHD-drug information was written by the pharmacist and reviewed by the chief cardiologist. A protocol stating which drug-related problems (DRP) require a discussion with the cardiologist was also prepared by the pharmacist and reviewed by the chief cardiologist.

An experienced clinical pharmacist (MJÖ) with 60 ECTS postgraduate education in clinical pharmacy, specialised in cardiology, conducted the intervention. The pharmacist had no formal education or training in MI, but applied the theory of MI to the instruments and the interviews.

Intervention procedures

The patients visited the pharmacist at the cardiology outpatient clinic 3 months after the hospital discharge. In close proximity to the visit, the pharmacist prepared the interview and conducted the medication review, using patient data from the electronic health record shared between hospital and primary care. When seeing the patient, the pharmacist used the adapted DRAW and MI-principles. The pharmacist also used the patient's baseline BMQ-questionnaire as a guide for the interview. At the end of the visit, the patient was given the information leaflet. Any problems found during the process of review and interview that could not be solved by the pharmacist and patient together, were discussed with the cardiologist after the visit. The follow-up phone call 2 weeks after the visit suited as a second chance for the patient to ask questions and the pharmacist to strengthen the message from the interview. Patients who had their drug regimen changed as a result of the medication review were informed about the changes in the follow-up phone call, and further phone calls were made if considered necessary. The pharmacist documented the consultations in the electronic health record. See online supplementary appendix A, for a table showing the activities of the intervention.

Patient characteristics

A total of 38 patients were asked to participate, and among these, 23 (60%), 16 men and 7 women, gave their consent. Recruited patients were randomised to intervention (n=11) and control (n=12) group. In the control group, two patients were lost during the study, one who did not have the possibility to visit the hospital and one participant withdrew without giving any reason.

Demographic information and medical history of the intervention and control groups are presented in table 1.

Table 1

Patient characteristics

Data collection and outcome measures

The Morisky Medication Adherence Scale (MMAS-8-Item) was used to measure self-reported adherence; this scale has been validated as an adherence measure in hypertension.9 The BMQ-S was used to assess beliefs of necessity and concern about medicines, and has been validated in cardiac patients.10 The patients received the BMQ-Specific and MMAS-8 questionnaires by mail according to the outline in figure 1. They were requested to answer and return the questionnaires within 10 days.

The DRPs were categorised according to the seven categories described by Cipolle et al,11 and complemented with ‘Insufficient knowledge of purpose and/or function of drug’ and ‘Other’. The time needed for each activity in the intervention was recorded by the pharmacist.

Primary outcome measures were the results from the BMQ-S and MMAS-8 in the intervention group compared with the control group at first follow-up. Secondary outcome measures were the persistence of the effect 4 weeks later in the intervention group (figure 1).

Statistical analysis

The results from MMAS-8 were reported as the proportion of patients having low (<6 p), medium (6–8 p), or high (8) adherence, respectively. The results for BMQ-S were reported as stated by the originator: as the change in sum of the necessity and concerns scales, respectively, as well as the difference between perceived necessity and concern, that is, the necessity-concerns differential.

BMQ-S has ordered verbal categories to which numerals are assigned in order to create the global results mentioned above. However, the global score arising from the sum of scores has no interpretable meaning, since ordered categories do not represent mathematical values. Therefore, the results for BMQ-S in this trial were also interpreted according to a statistical model by Bondessonet al,12 that take into account the ordered categorical structure of data.13 In this model, a global score that can be studied graphically for a group of patients is constructed from the combination of scores on the subscales, necessity and concern, for each patient. The parameter that is thought to influence the patients’ medicine-taking behaviour the most is chosen as the score, for instance the median score on each scale, or the most negative or most positive belief. Figure 2 shows the distribution of global scores across the necessity and concern scales. For this paper, the most negative belief global score was chosen; the patients were categorised according to their highest agreement on the concern scale, and their lowest agreement on the necessity scale. Individual change between time points could then be presented using the five global scores.

Figure 2

The five global scores distribution. The global score ‘A’ represents the group of patients with the most strongly perceived concern, and with the least perceived necessity. With the global score ‘E’, patients have little perceived concern and a strong belief in the necessity of their drugs.

Missing data was identified in one of the BMQ-S questionnaires and four of the MMAS-8 questionnaires, all from different participants. Missing data was handled according to the principle of last entry carried forward, or backward.

Ethical considerations

The study was carried out as part of the Master of Science Programme in Clinical Pharmacy at the Uppsala University in Uppsala, and a quality assurance project at the cardiology unit, Kalmar County Hospital; therefore, no ethical approval was needed, but the study was undertaken according to the Helsinki declaration.


Main analysis

Self-reported adherence

At follow-up, the results from the MMAS-8 were very similar among control and intervention patients (table 2); no statistical test was applied due to the low number of patients.

Table 2

Self-reported adherence, MMAS-8, and beliefs about medicines, BMQ-S

However, at baseline, there were six patients in the intervention group with low or medium adherence compared with two in the control group. After the intervention, only one patient stated low adherence. In the control group, one patient shifted from high to medium adherence during the control period. Adherence measures are summarised in table 2.

Beliefs about medicines

At follow-up, the intervention group had a mean (SD) necessity score of 21 (4), a concern score of 12 (6) and the differential 9 (8); no patient had higher perceived concern than necessity (table 2). In the control group, the corresponding results were 21 (3) for necessity and 10 (5) for concern, with the mean differential being 11 (5).

The groups differed at baseline, the intervention group having lower necessity scores, more concern and a lower necessity-concerns differential than the control group. Three patients in the intervention group showed more concern than necessity at baseline, compared with none in the control group.

Figure 3A shows the global scores based on the most negative belief stated by the patient on the necessity and concerns subscales; individual change between baseline and follow-up-1 is also presented. In the intervention group, seven patients shifted towards more positive beliefs, whereas in the control group two patents shifted towards more positive beliefs and three patients towards more negative beliefs.

Figure 3

(A) Individual change at follow-up-1. Graphical presentation of global scores (number of patients) at baseline and follow-up-1 for intervention group and control group, respectively. The colour within the circles represents the global score at baseline (individual patient). The global score ‘A’ represents the group of patients with the most negative attitudes, and ‘E’ the most positive. (B) Individual change at follow-up-2. Shows the persistence of the effect 4 weeks after the intervention. The colour within the circles represents the global score at follow up-1 (individual patient). The global score ‘A’ represents the group of patients with the most negative attitudes, and ‘E’ the most positive.

Secondary analyses


MMAS-8 scores 4 weeks after the intervention are shown in table 2. For BMQ-Specific, one patient in the intervention group again perceived more concern than necessity and the results for the group shifted slightly towards a lower necessity-concerns differential (table 2).

Individual change with regards to the most negative belief is shown in figure 3B. Four weeks after the intervention one patient had shifted from positive towards more negative beliefs, whereas two patients had shifted towards more positive beliefs during this period.

Process measures

The mean time for preparation, interview and documentation was 18, 50 and 24 min, respectively, not including discussion with the cardiologist. Discussion with the cardiologist about drug treatment and possible DRPs took place for 11 of the patients, evenly distributed between the groups. All but one patient were found to have at least one DRP.

In the intervention group, a total of 27 DRPs, mean (SD) 2.5 (1.2) were found, and in the control group, 18 DRPs, mean (SD) 1.8 (0.9). A summary of the DRPs and the actions taken upon them is shown in table 3.

Table 3

Drug-related problems and actions taken upon them

Nine patients had their drug regimen changed after discussion with the cardiologist (5) or primary care physician (2); the most changes concerned change of drug because of adverse drug reactions (3) or need for more effective drugs (4). Adverse drug reactions (13) were reported by, or identified in, eight patients; the pharmacist acted upon them with information, MI, and individual risk-benefit assessments together with the patient, and in some cases change of drug or dosing time.

The mean time to follow-up by phone was 13 days. Five patients received more than one follow-up call due to changes in drug treatment, one received three calls.


The patients’ self-reported adherence, as well as their beliefs about medicines, were similar among intervention and usual care patients at follow-up. However, after consultation by a clinical pharmacist using MI-approach and a semistructured interview, a larger proportion of patients changed towards more positive beliefs compared with control.

In this study, no change was detected in the MMAS-8 in the control group, but the beliefs of three patients changed in a negative direction and for two patients in a positive direction. Changes in beliefs about medicines have previously been investigated in a larger group of patients (n=812) with CHD, and a trend towards more negative beliefs was shown 12 months postdischarge.14

Calvert et al15 studied an adherence intervention in CHD-patients (n=108 at follow-up) with longer follow-up (6 months) and more robust measures of adherence. No effect was detected in the primary outcome, self-reported adherence, but the study showed trends towards better adherence to statins and a significant effect in adherence to β-blockers according to prescription filling records. Their failure to detect any effect in the self-report might be attributable to self-report instruments’ general weakness; their overestimation of adherence due to recall bias and social desirability. A further limitation was that they used an older version of the Morisky-scale, which is more prone to overestimating the adherence.9 The intervention involved pharmacists from the hospital setting and the community pharmacy in the healthcare team, and also the primary care physician was involved.15 This is in line with the conclusions of reviews of interventions aiming at improving adherence to cardiovascular medications. The interventions showing most promising results use in-person interventions with a behavioural (motivational) effort and possibly in the setting of a multiprofessional team.16 ,17 In this study, the intervention had a behavioural effort and was conducted in person; the pharmacist worked within the multiprofessional team of the cardiology unit, but without established contact with primary care.


This pilot study has several limitations, of which the small number of patients is one. However, the size of the study was decided in order to test the feasibility of the intervention, for example, in terms of acceptability of the pharmacist at the clinic and by the patients, the recruitment process, and retention of patients in the study. The size was also thought to be sufficient to find clues about effect sizes as to make future sample-size calculations.18 No conclusions can be drawn about the patients’ true adherence, since the only measures were self-report and beliefs about medicines. The study also suffers from selection bias.

The short time interval for follow-up and the group differences at baseline makes it impossible to draw conclusions about actual changes of self-reported adherence and beliefs.

Statistical analysis

The results from BMQ-Specific were presented according to the originators’ method, but also according to the model by Bondesson et al, which can be used for repeated qualitative assessments. The method can be used for descriptive purposes and for evaluation of change in scale assessments. There are alternative ways to evaluate the change14 where the proportions of patients whose necessity or concern score fell from higher than the midpoint to lower than the midpoint, or the opposite. However, this too defines that a change around the midpoint is discriminating, but we do not know if this bears any more meaning than changing from 10 to 5 on the concern scale. With the method used in this paper, we are taking account of the ordered verbal categories of the questionnaire, and changes anywhere on the scale can be detected. However, since this method was first presented, more research has been conducted using the attitudinal analysis.19 In that analysis, patients are categorised according to the necessity-concerns framework into four groups; the groups termed accepting ambivalent, neutral and sceptical, have been found to have different adherence behaviours. Patients categorised as accepting report higher adherence than patients in the three other categories. Ambivalent patients report lower adherence than accepting patients, but higher adherence than neutral and sceptical patients, no difference has been found between neutral and sceptical patients.19 ,20 Future studies might want to use the attitudinal analysis to enable comparisons; however, further development of a statistical method for the BMQ-S would also be welcomed.

Implications for future studies

In this study, the use of a cardiology-focused medication review and the DRAW-instrument together with MI-approach, and the ability to easily discuss problems with the cardiologist meant that potential or actual DRPs were identified and handled. Thus, the process described in this study is in accordance with the model of pharmaceutical care,15 and has the potential to change patients’ beliefs about medicines in a positive direction; no patient in the intervention group changed towards negative beliefs. A larger study with longer follow-up and more robust measures of adherence, as well as surrogate measures of clinical outcomes, such as blood-pressure or effect on lipids, is needed to draw any conclusions of the effect of the intervention. Since a convincing majority of patients had at least one drug-related problem, and the adherence to long-term treatment has been found to be suboptimal in large patient groups,3 we suggest that the intervention described herein should be aimed at all patients with secondary prevention for CHD. However, patients with low self-reported adherence and/or negative beliefs about their medicines presumably need a more extensive intervention with visits or phone calls regularly during a longer follow-up period.7


No difference between groups was found in adherence and beliefs at follow-up. However, after consultation by a clinical pharmacist using MI-approach and a semistructured interview, a larger proportion of patients changed towards more positive beliefs compared with the control group.

This and other information can be used as a base for performing a larger trial to study the impact on relevant process and outcomes variables.

Key messages

  • What is already known on this subject

  • Patient adherence to secondary prevention of ischaemic heart disease is low, resulting in unnecessary morbidity and mortality.

  • Adherence interventions often fail to improve adherence, and are not grounded on modern theory about the adherence behaviour, such as medication beliefs and health behaviour change.

  • What this study adds

  • In this pilot study of an intervention using medication review and Motivational interviewing-approach, no difference was found between groups in adherence and beliefs.

  • Intervention patients with more negative beliefs about medicines were affected in a positive direction.


We thank the chief cardiologist, professor Jörg Carlsson, MD, PhD, and the cardiology nurses, at the Kalmar County Hospital for the contribution in planning and conduction of the study. We thank the County Council of Kalmar, for giving us the opportunity to perform the study.


View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Contributors All four authors have contributed to the design or analysis of this work, as well as the draft and the final approval.

  • Funding Malin Johansson Östbring has received funding from the Kamprad Family Foundation for Entrepreneurship, Research and Charity – grant number FKS 2013-0065.

  • Competing interests None.

  • Ethics approval Regional Ethic Committee Linköping, Sweden.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.