Impact on Medication Use and Adherence of Australian Pharmacists’ Diabetes Care Services

RESEARCH

Impact on Medication Use and Adherence of Australian Pharmacists’ Diabetes Care Services Ines Krass, Susan J. Taylor, Carlene Smith, and Carol L. Armour

Received March 17, 2004, and in revised form May 20, 2004. Accepted for publication June 4, 2004.

ABSTRACT Objective: To assess the effect of a specialized service implemented in community pharmacies for patients with type 2 diabetes on medication use and medication-related problems. Design: Parallel group, multisite, control versus intervention, repeated measures design, with three different regions in New South Wales, Australia, used as intervention regions, then matched with control regions as much as possible. Intervention: After initial training, pharmacists followed a clinical protocol for more than 9 months, with patient contact approximately monthly. Each patient received an adherence assessment at the beginning and end of the study, adherence support, and a medication review as part of the intervention. Main Outcome Measures: Risk of nonadherence using Brief Medication Questionnaire (BMQ) scores and changes to medication regimen. Results: Compared with 82 control patients, 106 intervention patients with similar demographic and clinical characteristics had significantly improved self-reported nonadherence as reflected in total BMQ scores after 9 months. The mean (± SD) number of medications prescribed at follow-up in intervention participants decreased significantly, from 8.2 ± 3.0 to 7.7 ± 2.7. No reduction was observed among the control patients (7.6 ± 2.4 and 7.3 ± 2.4). The overall prevalence of changes to the regimen was also significantly higher in the intervention group (51%) compared with controls (40%). Conclusion: Community pharmacists trained in medication review and using protocols in collaboration with providers improved adherence in patients with type 2 diabetes, reduced problems patients had in accessing their medications, and recommended medication regimen changes that improved outcomes. Keywords: Diabetes, pharmaceutical care, disease management, community pharmacy, adherence, medication-related problems. J Am Pharm Assoc. 2005;45:33–40.

Vol. 45, No. 1

January/February 2005

www.japha.org

Ines Krass, PhD, is Associate Professor; Susan J. Taylor, PhD, is Senior Lecturer; Carlene Smith, BPharm, is Practitioner/Teacher; Carol L. Armour, PhD, is Professor of Pharmacy Practice, Faculty of Pharmacy, University of Sydney, New South Wales, Australia. Correspondence: Ines Krass, Faculty of Pharmacy A15, University of Sydney, Camperdown, New South Wales 2006, Australia. Fax: 61-2-9351-4391. E-mail: [email protected] Funding: Supported by the Australian Department of Health and Aged Care through the community pharmacy practice research grants scheme. Acknowledgments: To Barbara Gates, Sophie Gosling, PhD, Fleur Hourihan, Sarah Stephenson, Ingrid Thuis, and Kristy Edmonds who helped in various stages of this project; to Abbott Diagnostics for donation of MediSense blood glucose meters and Precision Link software; and pharmacists who gave their time and expertise and gave us feedback on the service.

Journal of the American Pharmacists Association

33

RESEARCH

Australian Pharmacists’ Impact on Diabetes Care

D

iabetes mellitus (DM) is a chronic incurable disease whose prevalence is growing worldwide.1 Approximately 7.5% of Australians older than 25 years have diabetes, and more than one half of them are undiagnosed. In Australia, the direct annual health care costs for treating diabetes and its associated complications were estimated to be AU$1.4 billion in 1995 and may reach AU$2.3 billion by 2010.2 Since the publication of evidence to support the benefit of strict control of blood glucose and blood pressure in preventing the complications of diabetes,3–5 a plethora of treatment guidelines have been developed both in Australia and worldwide. Despite this, evidence suggests that many people with diabetes are not receiving optimal care.6 Implementation of guidelines in Australia varies, with research showing that 33% of patients had never had their feet examined and only 71% had glycosylated hemoglobin (A1C) concentrations within recommended targets.7 One of the barriers to the ideal management of diabetes is the patient’s adherence to a diabetes regimen. In one study, only 7% of patients were found to be fully adherent to all aspects of their antidiabetic regimen, which included adherence to medication, diet, exercise, and self-monitoring.8 With regard to adherence to antidiabetic medication, reported rates vary depending on the sample used

AT A GLANCE Synopsis: Optimized medication regimens and improved patient adherence are likely when community pharmacists provide pharmaceutical care services to patients with type 2 diabetes, according to this study from New South Wales, Australia. A further analysis of data first reported in the July/August 2004 issue of JAPhA, this article shows that pharmacists’ interventions were associated with a significant reduction in the number of medications each patient was using. Patient adherence improved with interventions, although significance was not reached in this relatively small group of 188 patients. Analysis: Medication adherence plays a critical role in the control of type 2 diabetes and its associated complications. The medication reviews and adherence assessments provided as part of this comprehensive disease management service are likely to have contributed substantially to the clinical improvement in glycosylated hemoglobin levels in intervention patients detailed in the previous article. Standardized medication reviews guided by clinical protocols helped optimize medication use for improved clinical outcomes. Monthly visits with the pharmacists solved patient medication access problems. The study suggests that providing adherence support, performing medication reviews, and helping solve medication access problems are important services for community pharmacists to provide.

34

Journal of the American Pharmacists Association

and the methods of measuring adherence. For example, two studies from the United States using retrospective review of pharmacy records and claims databases and a study in Dutch diabetic patients using medication event monitoring systems (MEMS) for measurement, reported similar adherence rates to antidiabetic medication of between 70% and 83%.9–11 More recently, a large populationbased study of prescription records in Scotland found adequate adherence (≥ 90%) in just 31% of patients prescribed sulfonylureas alone and in 34% of those prescribed metformin alone.12 Another barrier to achieving optimal control in the management of diabetes is the prevalence of medication problems related to suboptimal prescribing or drug misadventure through adverse events or drug interactions. A substantial proportion of medication-related problems within the health care system involve patients with diabetes.13,14 Medications play an important role in the control of type 2 diabetes mellitus and its complications, and regular medication review, ongoing diabetes self-management education, coaching,15–17 and monitoring by pharmacists can empower the patient and assist in the early identification and resolution of drug-related problems and improve overall disease control.18,19 Given the potential contribution of the pharmacist to the support of patients with type 2 diabetes in their disease management, the lack of service models suited to the Australian health care environment, and consequent lack of evidence on the potential impact of such services on medication use, health outcomes, and costs of care for patients with type 2 diabetes, we set out to develop, implement, and evaluate a disease management service model for type 2 diabetes in the community pharmacy setting. Key elements of the service involved an assessment of the medication regimen through adherence assessment and medication review.20 The effect of this service on clinical and humanistic outcomes has been reported in an earlier article.20

Objectives This article assesses the effects on medication use and medication-related problems of a community pharmacy–based disease management service for Australian patients with type 2 diabetes.

Methods Research Design and Sample As reported previously,20 a parallel-groups repeated measures design was used to compare adherence and medication use outcomes in two cohorts: 82 patients with type 2 diabetes who received usual care in a community pharmacy or an outpatient clinic (control); 106 patients who received specialized services in matched settings (intervention). Nine community pharmacists from three different regions in New South Wales, Australia, delivered interventions. The geographic regions selected included www.japha.org

January/February 2005

Vol. 45, No. 1

Australian Pharmacists’ Impact on Diabetes Care

rural and metropolitan community pharmacy sites and a diabetes clinic. A total of 14 pharmacists from matching regions and sites recruited usual care (control) patients. Patients recruited for this study had to have type 2 diabetes; be taking more than three medications, including at least one antidiabetic medication; be between 18 and 85 years of age; and be capable of understanding and completing the questionnaires. All eligible, invited patients consented to participate in the study. Having signed informed consent forms, patients were interviewed by the pharmacist in either the pharmacy or the clinic. Intervention patients were given instructions on monitoring blood glucose concentrations and MediSense (Abbott Diagnostics Division, Lane Cove, Sydney, Australia) blood glucose monitors during the initial visit. Patients were asked to use the MediSense monitors during the study. A detailed diabetes history, information on quality of life, well-being, adherence to medication, clinical data (A1C, blood pressure, lipid profile), and medication history for the previous 6 months were collected during the initial visit (baseline data). Following the initial assessment, a medication review was conducted for patients who had issues related to their medications and social history. Pharmacists were trained in how to conduct medication reviews and asked to complete a pro forma medication regimen review (MRR) report, documenting the findings and recommendations made for each patient. Pharmacists followed a protocol to deliver targeted interventions to patients at each monthly patient visit over the 9-month intervention period. Patient protocols were developed on the basis of problems identified in the initial assessment and addressed adherence and other medication-related issues. Intervention strategies to support patient adherence included feedback on self-monitoring of blood glucose levels, education about the disease and medications, adherence devices, reminders, and regular follow-up. The service model was evaluated by comparing clinical, humanistic, and economic data collected at the beginning of the project and again at 9 months. Control pharmacists collected baseline data and then repeated the data collection at 9 months. A more detailed description of the development and implementation of the disease management service and study sample recruitment and retention has been described in a previous article.20

Adherence Assessment Adherence was assessed using two methods: the Brief Medication Questionnaire (BMQ) scores and computerized dispensed medication histories (DMHs) for a 6-month interval. The BMQ is a validated self-report tool consisting of three screens.21 The regimen screen measures self-reported adherence behavior. In this part of the questionnaire, questions were asked about the medication taken in the past week. Patients had to recall what and how much medication they had taken and whether they missed any pills. Patients received a score of 1 if their initial or spontaneous response indicated potential nonadherence with the current regimen for the target medication and a score of 0 if their response Vol. 45, No. 1

January/February 2005

www.japha.org

RESEARCH

indicated adherence. Indicators of potential nonadherence include failing to mention the target medication or reporting any interruption or discontinuation because of a late refill or other reason. The score range for this screen was 0 to 7. The belief screen identifies patient concerns about the efficacy of a given medication and possible concerns about unwanted adverse effects and short- or long-term risks (score range, 0–2). The recall screen identifies potential problems in remembering all doses, opening the container, reading labels, or obtaining refills (score range, 0–2). The scores of these three screens are summed to a possible maximum of 11 for the total BMQ score. Any score greater than 0 for each of the screens or the total BMQ score indicates potential nonadherence. Patients were given a classification of “adherent” if they scored 0 in the BMQ regimen screen. This classification was chosen because the regimen screen is the only screen to ask directly about medications taken. In addition to these two main measures of adherence, the patient’s ability to match medications to diseases and problems accessing medications were assessed by interview questions designed to gauge difficulties in getting refills dispensed on time, reading the pharmacy’s prescription label, opening bottles or foil packs, taking so many medications at once, or self-injecting insulin. The second method involved estimation of 6-month medication use for each patient from the pharmacy’s computerized DMHs. If a patient was recruited from the diabetes clinic, or specified that they also used more than one pharmacy, then these pharmacies were contacted and asked to supply the DMHs for the study subjects. The number of prescribed doses for each drug was then calculated for this time period, and the number of dispensed doses was calculated. If the number of dispensed doses was within 80% to 115% of the prescribed doses for all drugs, then the patient was classified as adherent.

Medication Reviews Intervention pharmacists were instructed by a clinical pharmacist during their 2-day training program20 to follow a systematic process for the conduct of each MRR, guided by standard operational definitions and a review of medication history from the protocol.14 For each review, the pharmacist completed an MRR report, which comprised both review findings and any recommendations as appropriate for each of the patient’s medications. The pharmacist sent this report to the patient’s general practitioner. At the end of the project (9 months after the service began), outcomes of the pharmacist-conducted medication reviews were evaluated. Documentation of outcomes was verified by observation of the patient’s DMHs and the patient’s medication record, a document updated monthly by the pharmacists at each visit. Any discrepancies in medication between these two references were settled by contacting the patient’s pharmacist or if needed the patients themselves.

Journal of the American Pharmacists Association

35

RESEARCH

Australian Pharmacists’ Impact on Diabetes Care

Data Analysis Statistical analyses were performed using SPSS for Windows 10.0 (Chicago, Ill.). For categorical variables, the Pearson chisquare test for independent proportions was used to compare differences in demographic and other characteristics such as distributions of medication. The Yates continuity correction was used for two by two comparisons. McNemar’s test was used to examine within-group changes in proportions between baseline and followup. Student’s paired t test was used for continuous normally distributed variables. Repeated measures analysis of variance tested for differences in mean scores between the two groups. For variables that were not normally distributed, the Kruskal–Wallace H test was used to test for differences in nonparametric data across sites. The Wilcoxon signed rank test was used for within-group comparisons at baseline and follow-up and the Mann–Whitney U test for independent samples for between-group comparisons. The level of significance for all tests was set at P < .05. The MRR forms were collated and entered into an SPSS database. Closed-response variables were entered using preassigned numerical codes. Open-ended variables required the development of coding frames based on classifications developed for the review process and pharmacist’s descriptions. These variables included MRR findings, recommendations, and results. Data analysis considered the following: (1) comparisons of the overall proportion and types of changes to therapy, and (2) number and median monthly costs of medication by cohort and preand post-MRR, using a nonparametric test for paired samples.

Results Study Sample The demographics of the patients in intervention sites closely matched those of patients in the control sites (Table 1). Patients were on average in their early or mid-60s and had had diabetes for a median of 6–8 years (Table 1). While patients in the intervention group on average had had their disease longer and were taking more medications for their diabetes, these baseline differences were not statistically significant. Differences between patients for whom complete data sets were obtained and those who had missing data were also statistically insignificant.

Risk of Nonadherence—BMQ Scores There was a statistically significant decrease in mean BMQ scores (self-reported level of nonadherence) in all three screens and the total following intervention in the intervention group, but there was no change observed in the control group (Table 2). The effect of the intervention on self-reported nonadherence was reflected in a significant reduction in total BMQ scores for the intervention group when compared with scores for the control 36

Journal of the American Pharmacists Association

Table 1. Demographic and Clinical Characteristics of Patients (n = 188) at Baseline

Characteristic

Intervention (n = 106)

Control (n = 82)

Age in years, mean ± SDa

64 ± 9

65 ± 10

Female, %b

55

49

speaking background, %b

24

14

Employment status:

14

13

8 (1–38)

6 (1–50)

Oral hypoglycemics only, %

66

78

Insulin only, %

13

12

Insulin and oral

21

10

8.2

7.6

Ethnicity: non–English-

full-time, %b Years since diagnosis, median (range)c Current antidiabetic managementb

hypoglycemics, % Total no. medications, %d

Abbreviation used: SD, standard deviation. aNo

significant differences; t test for independent samples.

bNo

significant differences; Pearson chi-square test.

cNo

significant differences; Mann–Whitney U test.

dP

= .02; t test for independent samples.

group (P < .001) after 9 months. The decreases in scores were also statistically significant in the regimen screen (P = .008) and belief screens (P = .03) but not in the recall screen. Other positive effects were observed in patient knowledge and a reduction in difficulties in taking mediation. The percentage of patients who were able to match all self-reported medications with the appropriate medical conditions increased from 79% at baseline to 84% at follow-up in the intervention group and decreased from 82% to 72% in the control group. Reported problems in accessing medications decreased significantly from 41% at baseline to 15% at follow-up in participants recruited to the intervention sites (P < .001) but were changed little in the control group (Table 3).

Proportion of Nonadherent Patients As shown in Table 3, the regimen screen score indicated a statistical trend toward a decrease in the proportion of nonadherent patients within the intervention group (preintervention, 59%; postintervention, 48%; P = .15), while the control group showed little change (50% and 48%; P = .70). In contrast, no difference was detected in the proportion of nonadherent patients at follow-up in either the intervention or control groups using DMHs (Table 3).

www.japha.org

January/February 2005

Vol. 45, No. 1

Australian Pharmacists’ Impact on Diabetes Care

RESEARCH

Table 2. Mean BMQ Scores at Baseline and Follow-up No. Patients

Baseline Mean ± SD

No. Patients

Follow-up Mean ± SD

P Value

Intervention

87

1.38 ± 1.20

87

0.94 ± 0.89

.009b

Control

67

1.01 ± 0.88

67

1.16 ± 0.88

.25

Intervention

87

0.67 ± 0.77

87

0.39 ± 0.64

.01b

Control

67

0.24 ± 0.43

67

0.27 ± 0.51

0.69

Intervention

87

1.21 ± 0.46

87

1.05 ± 0.30

.005b

Control

67

1.15 ± 0.40

67

1.10 ± 0.43

.41

Intervention

87

3.89 ± 1.78

87

2.74 ± 1.39

Control

67

2.82 ± 1.15

67

3.90 ± 1.45

Regimen

Belief

screena

screenc

Recall screenc

Total screend < .001b .07

Abbreviation used, BMQ, Brief Medication Questionnaire; SD, standard deviation. aScore bPaired cScore

range, 0–7. t test. range 0–2.

dRegimen

screen + belief screen + recall screen; score range, 0–11.

Medication Use A total of 500 findings were classified using categories developed through content analysis. Problems with drug use (14%) and interactions (15%) were the most frequent findings, together accounting for approximately 30% of all findings. Adverse effects (11%) and adherence risk (10%) were the next most frequent findings. A total of 367 recommendations were made by the pharmacists; the most frequent was ordering pathology tests (22%), such as serum creatinine and potassium determinations. Referrals to general practitioner (14%), clinical monitoring of drug therapy (12%), and addition of nondrug therapy (10%) were the next most frequent recommendations. The overall distribution of prescribed medications changed little during the study in either intervention or control participants. While the proportion of medications prescribed to intervention participants decreased in the musculoskeletal (11.6% to 10.2%) and gastrointestinal classes (7.0% to 5.6%), the differences were not statistically significant (Table 4). The total number of medications used by members of both the intervention (from 775 to 735) and control (from 532 to 513) groups decreased during the study (Table 4). In participants with complete medication histories, the mean (± SD) number of medications prescribed at follow-up in intervention participants decreased significantly, from 8.2 ± 3.0 to 7.7 ± 2.7 (P = .02); the decline observed among control participants, from 7.6 ± 2.4 to 7.3 ± 2.4, was not significant (P = .27). The prevalence of changes in medication regimens was higher in the intervention group (51%) compared with the control group Vol. 45, No. 1

January/February 2005

www.japha.org

(40%) (P < .001). Ceasing of medications was the most prevalent change in both interventions (16%) and controls (12%), as noted in Table 5. The prevalence of changes to the hypoglycemic regimen was higher in the intervention participants (50%) compared with controls (38%) but this was not statistically significant (P = .10). Most

Table 3. Nonadherence Rates, Knowledge of Medications, Medication Access at Baseline and Follow-up Intervention (n = 87)

Control (n = 67)

Baseline

Follow-up

Baseline

Follow-up

79

84

82

72

41a

15a

24

25

59

48

50

48

70

67

77

77

% Informed about medications % Having difficulties accessing medications % Nonadherent, BMQ regimen screen % Nonadherent, DMH (6 months)

Abbreviations used: BMQ, Brief Medication Questionnaire; DMH, dispensed medication history. aP

< .001; McNemar’s test.

Journal of the American Pharmacists Association

37

RESEARCH

Australian Pharmacists’ Impact on Diabetes Care

Table 4. Distribution of Medications by Therapeutic Drug Class at Follow-upa Intervention Drug Classb

Baseline No. (%)

Control

Follow-up No. (%)

Baseline No. (%)

Follow-up No. (%)

Endocrine

227 (29.3)

217 (29.5)

127 (23.9)

127 (24.8)

Cardiovascular (includes

285 (36.8)

285 (38.8)

191 (35.9)

185 (36.0)

90 (11.6)

75 (10.2)

63 (11.8)

63 (12.2)

Gastrointestinal

54 (7.0)

41 (5.6)

40 (7.5)

39 (7.6)

Neuropsychiatric

42 (5.4)

44 (6.0)

37 (6.9)

34 (6.6

Respiratory

16 (2.1)

17 (2.3)

18 (3.4)

17 (3.3)

coagulant/anti-coagulant) Musculoskeletal (includes analgesic)

Other

61 (8.8)

No. total medications

aObtained bDrug

775

56 (7.6)

56 (10.5)

735

513

from dispensed medication history for the 6 months preceding commencement of study.

class coded according to Australian Medicines Handbook.22

changes in the intervention group related to increases in dose of medication (19%) followed by dose reduction (9%). In the controls, most changes in regimen (19%) were dose increases.

Table 5. Changes to Regimen at Follow-up for All Medications Intervention (n = 740a) No. (%)

Change to regimen

Discussion The implementation of a specialized disease management service by pharmacists for patients with type 2 diabetes in a community pharmacy resulted in improvements in adherence to medications within the intervention group. By targeting risk factors for nonadherence identified though the adherence assessment using the BMQ,21 pharmacists were able to provide appropriate individualized adherence support to each patient. The BMQ21 is a validated instrument to measure adherence that relies on self-reporting. The results from this questionnaire showed a significant decrease in the scores on each of the screens, suggesting that several of the risk factors for nonadherence (e.g., understanding of the medication regimen and medication-related problems) had been addressed. This is in agreement with the interventions documented by the pharmacists.20 Education about medications and discussion of health beliefs were the most common adherence interventions conducted, while use of adherence aids and reminders were the least common.20 Two methods were used to determine the proportion of nonadherent patients. The first method used computerized pharmacy DMHs, which did not show any changes during the research period in the intervention or control groups. However, the second method, BMQ regimen screen scores, showed a decrease of 10% within the intervention group. Although this decrease was not statistically significant, primarily because of our small sample size, it suggested a positive effect of the specialized disease management 38

48 (9.3)

532

Journal of the American Pharmacists Association

Control (n = 547a) No. (%)

Dose increased

70 (9)

45 (8)

Dose reducedb

45 (6)

13 (2)

Drug added

82 (11)

49 (9)

Drug ceased

121 (16)

66 (12)

Drug changed

58 (8)

49 (9)

Total changesc

376 (51)

222 (40)

aNumber

of medications.

bP

= .003; chi-square test.

cP

< .001; chi-square test.

service on adherence. The results of this study agree with earlier findings that the instrument chosen to measure adherence is of great importance.23 In this study, using the BMQ to determine the adherence of patients was probably more accurate than using the DMH. One item in the BMQ was used to determine changes in patient’s ability to identify the purpose of their medications improved at followup. While the proportion increased from 79% to 84%, this was not statistically significant. However, the ability to identify the purpose of medication decreased within the control group. It seems difficult to explain this decrease, but a change in interviewer could have affected our results; approximately one third of the control patients had a different interviewer at postassessment than they did at preassessment. However, because the responses to the other www.japha.org

January/February 2005

Vol. 45, No. 1

Australian Pharmacists’ Impact on Diabetes Care

questions did not show any differences, this seems unlikely to explain the difference. As expected, patients with problems accessing medications significantly decreased in the intervention group, from 41% at baseline to 15% at the 9-month follow-up. The monthly pharmacy visits by these patients contributed largely to this decrease, as this helped to resolve medication access issues, which included getting prescriptions filled in time, remembering to take medications in time, and problems with opening the container, or reading the labels. Within the control group hardly any change occurred. Intervention patients appeared to have more problems at baseline in accessing their medications. After 9 months, however, the percentage of patients within the intervention group having problems with medication access fallen to below that of the control patients (25%, compared with 15% among intervention patients). Despite the differences at baseline, these results indicate the effect of the service on access to medications. The medication review conducted for each intervention patient also resulted in a significant reduction in mean medications per patient. Substantial changes to therapy, especially in antidiabetic and cardiovascular medications, reflected an emphasis on achieving better blood glucose control, blood pressure control, and control of lipids, contributing to the important clinical improvements that were observed.20 In common with other studies evaluating the role of the pharmacist in optimizing pharmacotherapy for patients with type 2 diabetes through medication review and ongoing monitoring in a variety of settings, both community and hospital clinic, our findings confirmed the need for therapeutic adjustments for many patients involving both dose reductions and increases and/or the cessation or addition of other therapy to reach evidence base targets for A1C, blood pressure, and lipids.18,24–28 While the precise contributions of adherence assessment and support and medication review components in a complex and comprehensive specialized disease management service provided by community pharmacists are difficult to determine, they likely contributed substantially to the clinical improvement in A1C achieved in the intervention group. The proportion of patients with an elevated A1C (> 7%) decreased from 72% to 54% in the intervention group, compared with a rise from 53% to 61% in the control group after 9 months.20 Thus these elements of the service are likely to be important to include in community pharmacy services for patients with type 2 diabetes.

Limitations Several factors limit interpretation of our study findings. The BMQ, a self-reporting method, was administered by several different people and is quite sensitive to prompting by the interviewer. All interviewers were not given the same training, and different levels of probing may have resulted in different responses. The intervention pharmacists were aware that they needed to address adherence and had strategies to deal with any problems. Thus, Vol. 45, No. 1

January/February 2005

www.japha.org

RESEARCH

these pharmacists could have used the BMQ instrument differently than the control pharmacists. The DMH, which seemed the more objective method, also had several important drawbacks. First, because some of the common medications such as aspirin and insulin are not prescription-only products, and because patients often visit more than one pharmacy in a 6-month period, not all medications used by the patients could be recorded, and the frequency of supply could not be determined accurately. The DMH also involved the uncertainty of whether dosage discrepancies were a result of out-of-date directions on prescription refills or more a reflection of genuine misunderstandings between the physician and the patient. This problem was less evident in intervention patients because they had monthly meetings and were thus more likely to inform their pharmacists about changes in dose. For the control patients we had no mechanism for determining why a prescriber had changed medication regimens. Another problem in using the medication history as a tool for measuring adherence is caused by the safety net subsidy under the Australian Pharmaceutical Benefits Scheme.29 This safety net allows prescription medications to be provided without charge for the remainder of a calendar year in which a threshold cost of prescriptions has been reached. Several patients were found to be overusing their medications toward the end of the calendar year. Anecdotal evidence from intervention pharmacists suggests that the practice of stockpiling medications may have skewed the adherence results based on the DMH.

Conclusion Our results show that community pharmacists trained in medication review and using protocols in collaboration with providers improved adherence in patients with type 2 diabetes, reduced problems patients have in accessing their medications, and proposed medication regimen changes that improved clinical, economic, and humanistic outcomes.

References 1. Amos A, McCarthy D, Zimmet PZ. The rising global burden of diabetes and its complications: estimates and projections to the year 2010. Diabet Med. 1997;14:S1–S85. 2. McCarthy D, Zimmet PZ, Dalton A, Segal L. The rise and rise of diabetes in Australia, 1996—a review of statistics, trends and costs. Canberra, Australian Capital Territory: Diabetes Australia; 1996. 3. Stratton I, Adler A, Neil H, Matthews D, et al. Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes (UKPDS 35): prospective observational study. BMJ. 2000;321:405–12. 4. UK Prospective Diabetes Group (UKPDS) Group. Intensive blood glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet. 1998; 352:837–53. 5. The Diabetes Control and Complications Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med. 1993;329:977–86.

Journal of the American Pharmacists Association

39

RESEARCH

Australian Pharmacists’ Impact on Diabetes Care

6. Hiss RG. Barriers to care in non-insulin-dependent diabetes mellitus. The Michigan experience. Ann Intern Med. 1996;124:146–51. 7. Colagiuri R, McLennan M. Implementing clinical management guidelines for diabetes: a practical approach. In: International Health Outcomes Conference; 21–22 July 1999; Canberra, Australian Capital Territory: National Association of Diabetes Centres; 1999. 8. Cerkoney K, Hart L. The relationship between the health care model compliance of persons with diabetes mellitus. Diabetes Care. 1980;3:594–8. 9. Boccuzzi SJ, Wogen J, Fox J, et al. Utilization of oral hypoglycemic agents in a drug-insured U.S. population. Diabetes Care. 2001;24:1411–5. 10. Venturini F, Nichol MB, Sun JC, et al. Compliance with sulphonylureas in a health maintenance organization: a pharmacy record-based study. Ann Pharmacother. 1999;33:281–8. 11. Paes AH, Bakker A, Soe-Agnie CJ. Impact of dosage frequency on patient compliance. Diabetes Care. 1997;20:1512–7. 12. Donnan PT, MacDonald TM, Morris AD. Adherence to oral hypoglycaemic medication in a population of patients with diabetes: a retrospective study. Diabet Med. 2002;19:279–84. 13. Ellis SL, Carter BL, Billups SJ, et al. Types of interventions made by clinical pharmacists in the IMPROVE study. Pharmacotherapy. 2000;20:429–35. 14. Masood N, Armour C, Krass I. Identifying potential medication related interventions in type 2 diabetes: a mechanism for enhanced community pharmacy services. J Soc Admin Pharm. 2002;19:170–8. 15. Basina M, Kraemer FB. Effectiveness of diabetes management: is improvement feasible? Am J Med. 2002;112:670–2. 16. Norris SL, Lau J, Smith SJ, et al. Self-management education for adults with type 2 diabetes: a meta-analysis of the effect on glycemic control. Diabetes Care. 2002;25:1159–71. 17. Raji A, Gomes H, Beard JO, et al. A randomized trial comparing intensive and passive education in patients with diabetes mellitus. Arch Intern Med. 2002;162: 1301–4. 18. Baran RW, Crumlish K, Patterson H, et al. Improving outcomes of community-dwelling older patients with diabetes through pharmacist counseling. Am J Health Syst Pharm. 1999;56:1535–9.

19. Berringer R, Shibley MC, Cary CC, et al. Outcomes of a community pharmacy-based diabetes monitoring program J Am Pharm Assoc. 1999;39:791–7. 20. Armour CL, Taylor SJ, Hourihan F et al. Implementation and evaluation of Australian pharmacists’ diabetes care services. J Am Pharm Assoc. 2004;44:455–66. 21. Svarstad B, Chewning B, Sleath B, Claesson C. The brief medication questionnaire: a tool for screening patient adherence and barriers to adherence. Patient Educ Couns. 1999;37:113–24. 22. Australian Medicines Handbook. 2nd ed. Adelaide: Australian Medicines Handbook Pty Ltd.; 2000. 23. Vitolins MZ, Rand CS, Rapp SR, et al. Measuring adherence to behavioral and medical interventions. Control Clin Trials. 2000;21:188S–194S. 24. Clifford RM, Batty KT, Davis TME, et al. A randomised controlled trial of a pharmaceutical care programme in high-risk diabetic patients in an outpatient clinic. Int J Pharm Pract. 2002;10:85–9. 25. Wermeille J, Bennie M, Brown I, McKnight J. Pharmaceutical care model for patients with type 2 diabetes: integration of the community pharmacist into the diabetes team—a pilot study. Pharm World Sci. 2004;26:18–25. 26. Cranor CW, Christensen DB. Short-term outcomes of a community pharmacy diabetes care program: the Asheville project. J Am Pharm Assoc. 2003;43:149–59. 27. Cranor CW, Bunting BA, Christensen DB. Long-term clinical and economic outcomes in a community pharmacy diabetes care program: the Asheville project. J Am Pharm Assoc. 2003;43:173–84. 28. Posey LM. Proving that pharmaceutical care makes a difference in community pharmacy [editorial]. J Am Pharm Assoc. 2003;43:136–9. 29. Australian Government Department of Health and Ageing Web site. Schedule of pharmaceutical benefits for approved pharmacists and medical practitioners. Last updated February 1, 2004. Accessed at www1.health.gov.au/pbs/index.htm, February 24, 2004.

Walsh’s Pharmacy • Ireland • August 2004 • Andrew Frasco, PharmD

40

Journal of the American Pharmacists Association

www.japha.org

January/February 2005

Vol. 45, No. 1