A Systematic Review of Interventions to Improve Adherence to Diabetes Medications within the Patient–Practitioner Interaction
Antoinette Schoenthaler, EdD, and Yendelela L. Cuffee, PhD, MPH
From the Center for Healthful Behavior Change, Department of Population Health, New York University School of Medicine, New York, NY.
• Objectives: To conduct a systematic review of the literature examining interventions delivered by health care practitioners to improve medication adherence in patients with diabetes mellitus (DM).
• Methods: Databases were searched up to 2 July 2013 to identify eligible studies that included interventions that were conducted in a clinic-based setting and delivered by a health care practitioner (eg, nurse, physician, diabetes educator) to improve adherence to diabetes medications (including oral hypoglycemic agents and insulin). Articles were limited to published clinical trials conducted in adults ≥ 18 years of age and published in English-language journals.
• Results: 18 papers were reviewed: 15 trials targeted patients with DM, 3 targeted health care practitioners. 7 patient-focused and 1 practitioner-focused trial demonstrated a beneficial effect of the intervention compared with a control group. The patient-focused trials were complex interventions involving a combination of adherence-enhancing strategies such as individualized patient counseling, tailored patient education, medication reminders, behavioral feedback and reinforcement, and care management by ancillary staff, nurses, text message or telephone-linked system. The practitioner-focused trial employed an electronic feedback system for individualized care and quality improvement. Limitations included the diversity in the measures employed to assess adherence; differing definitions of adherence; the inclusion of patients regardless of baseline adherence status; and the short duration of the trials.
• Conclusions: Additional research is needed to under-stand the conditions under which interventions targeting the patient-practitioner interaction can be implemented in clinic settings to improve medication adherence in patients with DM.
The prevalence of diabetes mellitus (DM) has grown to epidemic proportions. DM affects nearly 26 million Americans, ranking as the 7th leading cause of death in the United States, and is a major risk factor for stroke, heart disease, kidney disease, blindness and lower-limb amputations . Despite the availability of numerous efficacious treatments (eg, oral hypoglycemic agents and insulin) to manage DM, and their proven efficacy in reducing cardiovascular morbidity and mortality, adherence to treatment remains suboptimal . Recent systematic reviews have documented medication adherence rates for oral hypoglycemic agents (as assessed by self-report and objective measures) ranging from 36% to 93% among patients with DM [3,4]. Rates of insulin adherence range from 62% to 64% depending on the population studied . Poor adherence to DM treatments is a major contributor to inadequate glycemic control and is estimated to incur up to $289 billion annually in direct healthcare costs [5–7]. Given these consequences, patient nonadherence has been described as one of the greatest challenges for successful treatment of chronic diseases such as DM .
The definition of medication adherence has taken numerous forms over time. Once regarded as the patient’s responsibility (eg, compliance), the success of the prescribed medication in reaching the intended therapeutic target rested solely on the patient’s ability to understand and execute the health care practitioner’s directions . More recently, adherence has been recognized as a complex behavior that requires patients and practitioners to form a partnership based on a shared understanding of one another’s perspectives within a supportive healthcare delivery system [8,10]. When patients and practitioners engage in a collaborative relationship that includes shared decision-making [11,12], patient-centeredness [13–15], and information exchange[16,17], patients are more likely to report better adherence to their medications.
While much work has been done to identify patient factors, and to a lesser extent practitioner factors, associated with better medication adherence in patients with DM in observational studies, effective intervention approaches that practitioners can incorporate into clinical practice has not been evaluated. In this systematic review, we appraise the existing literature examining interventions delivered by health care practitioners on medication adherence in patients with DM.
The search strategy were similar to the ones used by Cramer et al  (in their systematic review of adherence to diabetes medications) and McDonald et al  (in their systematic review of interventions to improve adherence to medications). To identify manuscripts that met inclusion criteria for this review, we used standard Cochrane Collaboration systematic review techniques . The PubMed and EMBASE databases were searched from inception to 2 July 2013 for our concepts. In addition, key articles from each search were run in PubMed using the “related articles” feature to identify further articles. Additional strategies included searching the bibliographies of eligible articles and searching other systematic reviews and meta-analyses for relevant articles. The concepts for medication adherence, intervention type, and diabetes were included in the search with several keyword synonyms and the subject headings for practitioner-patient relations. Articles were limited to reports of clinical trials published in English-language peer-reviewed journals and conducted in adults (age ≥ 18 years). The full search strategy in PubMed is available online.
Studies were eligible for the review if they met the following criteria: (1) patients had a diagnosis of diabetes;(2) included an intervention to improve adherence to diabetes medications (including oral hypoglycemic agents and insulin); (3) there was assessment of medication adherence as an outcome either through electronic monitoring devices, self-report, pill counts or pharmacy refills; (4) the intervention took place in a clinic-based setting and delivered by a health care practitioner, defined as either a nurse, physician, diabetes educator, or care manager; and (5) the study was a clinical trial (case-control study or randomized controlled trial [RCT]). We excluded studies that included pharmacist-led interventions given that this topic has been addressed in other recent systematic reviews [20–24]. We also excluded studies that did not report adherence rates for diabetes medications or the methods used to determine medication adherence. The outcome of interest was the between-group difference in medication adherence between intervention and control arms.
All titles and abstracts from the search were reviewed independently by the authors. Citations were categorized as potentially relevant, not clearly relevant, or gave insufficient information to make a judgment. Any disagreements about inclusion in the review were discussed by the authors, with all differences resolved by consensus. Percent agreement between the 2 authors was high (86%) across all reviewed citations. Printed copies of all potentially relevant citations were obtained. The authors independently abstracted all data from the eligible citations. Data were abstracted from tables, figures, and text using a structured data collection form. Data including participant and site characteristics, study methods, outcomes, and risk of bias were collected on the form. In addition, 2 authors of potentially relevant citations were contacted about unpublished data. Risk of bias for each study was assessed using the guidelines outlined in The Cochrane Collaboration Handbook for Systematic Review of Interventions . Percent agreement between the authors was high (96%) across all abstracted data.
A total of 1011 articles were identified, 24 of which were extracted for full review. Six of these studies were excluded for one of the following reasons: incomplete data on adherence outcome [25, 26], the intervention was not clinic-based [27–29], or the adherence assessment was not specific to diabetes medications . Thus, 18 trials were included in this systematic review [31–48]. (See Figure summarizing the literature search results.)
The characteristics of the included trials are shown in Table 1. Seventeen of the included studies were RCTs [31–37, 39–48] and 1 was a single-group pre-post study . Among the RCTs, the comparison group was usual care (UC) in 11 trials [32–34, 36, 39–42, 44–46]. An attention control (ie, sessions that are peripheral to the intervention topics to control for additional time and attention) or a minimal intervention was used in 6 trials [31, 35,37,43,47,48]. Approximately one-third of the trials (n = 7) were conducted in the United States [32,33,37,41,42,44,45]. The number of participating clinics per trial ranged from 1 to 86 (mean, 15). The majority of the trials (n = 11) were conducted in primary care clinics [32–37,41,42,44–46]. Seven trials were conducted in specialty clinics (eg, endocrinology clinic) or tertiary hospital settings [31,38–40,43,47,48].
Fifteen of the trials targeted patients [31–35,37–41,43–45,47,48] and 3 targeted practitioners [36,42,46]. The sample size of the trials ranged from 45 to 2458 patients (median, 197) and 40 to 160 practitioners (median, 140). Fifty-four percent of the patients were male with a mean age of 58.3 (SD, 5.2) years. Nine trials targeted patients with poor glycemic control (HbA1c > 7.0%) at baseline [33,34,37,39–42,47,48]; one trial targeted patients newly diagnosed with DM . Three of the trials targeted high-risk patients with comorbid conditions including hypertension  and depression [32,41]. The practitioner-focused trials targeted primary care physicians, physician assistants, and nurse practitioners [36,42,46]. Only one of the practitioner-focused trials reported demographic characteristics .
Eleven of the trials (60%) were complex interventions that used a combination of strategies to target multiple self-management behaviors (eg, self-monitoring blood glucose [SMBG], foot care) [26,33,37–40,43–45,47,48] Eleven of the trials were delivered by practitioners that were existing members of the clinic staff [31,33,34,36,37,40–43,46,48]; 7 trials employed ancillary staff for the purposes of the study [32,35,38,39,44,45,47]. The median duration of the trials was 6 months (range, 1.5–15 months). Follow-up was adequate in majority of the trials (range, 64–100%). Most studies used a single measure of adherence: electronic monitoring devices (EMD) were used in 2 trials [32,34], pill count was used in 2 trials [31,40], pharmacy refills was used in 4 trials [36,41,42,46], and self-report data was collected in 10 trials [33,35,37–39,43–45,47,48]. The Morisky medication adherence scale was the most commonly used self-report measure. Two studies used a combination of objective (eg, EMD or pharmacy) and self-report measures to assess adherence [34,42]. A majority (n = 17) of the studies also assessed the impact of the intervention on change in glycemic control. Four trials also assessed the impact of the intervention on changes in patient trust in the practitioner and patient involvement in decision making  or perceived satisfaction with the practitioner’s communication skills [34,44,48].
Risk of Bias
Fifteen of the 17 reviewed RCTs specified the randomization procedures or included clear description of the allocation concealment procedures [32–38,40–44,46–48]. Eleven of the RCTs included information on blinding of the data collection procedures and/or whether the interventionist was blinded to the study objectives [32–35,37,41–45,47]. One trial used the same clinic staff member to deliver the intervention and control arms . The majority of studies adequately reported how incomplete outcome data was handled with most (n = 10) using an intent-to-treat analysis [32–35,37,41,42,44–46] Finally, all of the studies adequately reported data for the primary outcome. A table showing the ratings of risk of bias for each article is available online.
Effects of Included Interventions on Medication Adherence
The effects of the interventions are summarized below and in Table 2. A meta-analytic approach could not be used to calculate study effect sizes due to the high heterogeneity of the interventions and measures employed. As such, we only reported whether statistically significant differences were noted in adherence outcomes between the intervention and control groups. Of the 18 included trials, 8 reported statistically significant improvements in medication adherence attributed to the intervention [31,32,34,36,38,43,45,47]. Seven of these trials targeted patients [31,32,34,38,43,45,47] and 1 targeted practitioners . Among the nonsignificant trials, 8 targeted patients [35,37,39–41,44,48] while 2 were practitioner-focused [42,46]. Comparison of participant characteristics across trials that reported such data showed that the positive trials tended to consist of a higher percentage of minority patients (32.7% vs. 17.5%) and patients with a high school education or less (60.2% vs. 43.8%) than patients in nonsignificant trials. Patients also had a longer mean duration of time with diabetes in the positive trials (9.18 vs. 7.87 years); however, they were less like to be prescribed insulin (24.3% vs. 33.0%) as compared with patients in nonsignificant trials. Finally, trials reporting positive findings were more likely to be conducted in primary care clinics [32,34,36,45]. Alternatively, trials reporting negative findings were more likely to be comprised of older male patients (60.8% male, mean 61.4 years of age vs. 46.3% male, mean 55.8 years of age in positive trials). All 3 studies performed in Veterans Affairs Medical Centers also reported negative findings [33,37,44], which could not be explained by differences in baseline HbA1c or participant health status when compared with negative trials in other clinic settings. Below we describe the characteristics of the positive interventions as compared with nonsignificant trials in terms of the intervention content, methods of intervention delivery and dose, assessment of medication adherence, and impact on HbA1c. Descriptions are further categorized by intervention target: patients or practitioners.
Intervention content. In the 7 patient-focused trials reporting positive effects, 3 primarily focused on improving medication adherence [31,32,34]. The interventions were comprised of multiple strategies including medication reminders, care management, individualized counseling to address patient-specific barriers to nonadherence (eg, beliefs about medications, medication cost), patient education, and self-management support [31,32,34]. For example, Bogner et al  randomized 180 patients with comorbid DM and depression to either an integrated care intervention or usual care; adherence was assessed with an EMD for a period of 12 weeks. The integrated care intervention consisted of individualized sessions that addressed patients’ psychosocial and logistical barriers to adherence while also providing patient education about the course and treatment of DM and depression, and assistance with referrals. Care managers also collaborated with patients’ primary care physician to facilitate adherence to guideline-based treatment recommendations. At baseline, the medication adherence rates for oral medications (measured as proportion of patients with ≥ 80% adherence) were the same for both conditions over the past 2 weeks. At the 12-week follow-up, the proportion of patients with adherence rates ≥ 80% was significantly higher for the intervention group than for the control group (65.2% vs. 30.7%, P < 0.001). The intervention group was also significantly more likely to achieve glucose control (HbA1c < 7%) and experience fewer depressive symptoms at 12 weeks .
The remaining positive patient-focused trials targeted several diabetes self-care behaviors in addition to medication adherence and employed intervention strategies that included tailored patient education, development of problem solving skills, action planning, and behavioral feedback and reinforcement delivered either by trained nurses, text message, or a telephone-linked system [38,45,47]. The teach-back method and using pictorial images to explain concepts were also identified as effective strategies for delivering information to patients with low health literacy . Neither of the positive trials that assessed patients’ satisfaction with the practitioners’ communication skills reported significant differences between the intervention and control group at the final follow-up [34,48]. In contrast with the positive interventions, the nonsignificant trials were more likely to use patient education as the primary strategy [33,39,40,48] and target multiple diabetes self-care behaviors (eg, weight management, SMBG) with none focusing solely on medication adherence. However, in 1 trial, patients’ perceived satisfaction with the interpersonal aspects of care was higher among the intervention group than the control group .
Intervention delivery and dose. The method of intervention delivery varied across trials. Research nurses or college-level students delivered the intervention in 4 of the positive trials [32,38,45,47] while existing clinic staff were utilized in 3 trials [31,34,43]. Two trials utilized technology as a means to provide tailored interventions that maximized nurses’ time by focusing on patients with the greatest care needs [38,45]. For example, in the trial by Piette et al , an automated telephone disease management (ATDM) system designed to assess health risks, provide health tips, and allow patients to engage in interactive self-care modules was used to augment nurse case management (NCM). At the 12 month follow-up, medication non-adherence decreased by 21% among patients randomized to the ATDM+NCM group as compared with the UC group .
Interventions delivered by clinic staff tended to include fewer intervention contacts (mean, 2 sessions) for shorter duration (average length of intervention, 45 minutes over 2.5 months). In contrast, interventions delivered by trained research staff ranged from 3 to 12 sessions (mean, 6.67) and were of longer duration (average length of intervention, 111 minutes over the course of 5.25 months). Technology-assisted interventions also had higher intervention contacts with up to 72 contacts in a trial utilizing text messaging . Exposure to the intervention was assessed in 3 of the positive trials [34,43,45]; an average of 90% of patients received the intervention as intended.
Patient-level trials reporting nonsignificant findings were more likely to utilize existing clinic staff (n = 5) [33,37,40,41,48] than research staff (n = 3) [35,39,44]. Two of the trials utilized group-based intervention strategies [33,37], of which 1 used trained peer supporters . One trial used an ATDM system . In contrast to the positive trials, clinic staff–led interventions in the nonsignificant trials had a higher number of intervention contacts (mean,9 sessions) and were of longer duration (mean, 10 months). Interventions led by research staff ranged from 4 to 16 sessions (mean, 10 sessions) for an average of 9 months. Intervention contact time was not consistently reported in these trials. Primary care physicians were involved in 6 of the trials for medication management and to address acute health concerns [33,35,39,41,44,48]. Six trials reported intervention attendance rates with an average of 78% of patients receiving at least some of the intervention as intended [33,35,37,39,41,44].
Assessment of medication adherence. Three of the positive trials (43%) utilized an objective measure [31,32,34] to assess medication adherence while 4 trials utilized validated patient self-report questionnaires [38,43,45,47]. All of the trials assessed patients’ baseline adherence level, of which 3 (43%) included patients with high adherence (≥ 80% adherence at baseline visit) [31,34,47]. In 1 trial adherence was assessed via an EMD and self-report questionnaire . At the final follow-up, a higher proportion of patients reported being adherent to their oral medications than what was “objectively” reported by the EMD. Finally, 2 of the trials assessed medication adherence to oral medications only [32,34], 3 trials assessed both oral medications and insulin usage [38,45,47], and 2 did not specify [31,43]. Two of the trials assessing both types of medicines accounted for occurrences of hypogylcemia in the calculation of medication adherence [45,47]. However, neither study reported differential effects of the intervention on adherence to oral medications as compared to insulin.
In contrast to the positive trials, 75% (n = 6) of the nonsignificant trials utilized self-report measures [33,35,37,39,44,48]. Baseline adherence was assessed in 7 of the trials, of which 4 (57%) included patients with high adherence [28,35,41,48]. Three of the trials assessed medication adherence to oral medications only [35,39,40], 4 trials assessed both oral medication and insulin usage [37,41,44,48], and 1 did not specify . Only the trial by Heisler et al  reported the differential effects of the intervention by medication type. While nonsignificant, patients randomized to the NCM attention control group exhibited greater improvements in adherence to insulin at the 6 month follow-up compared with the reciprocal peer support intervention group (–10% vs. –3% reduction in missed doses over the past week). The remaining 3 trials that collected data on both medication types did not account for insulin usage in the calculation of medication adherence.
Impact on HbA1c. Six of the positive trials examined the impact of the intervention on HbA1c (mean baseline level, 8.47%) [31,32,34,38,45,47]. Four (67%) of the trials demonstrating a positive effect on medication adherence also reported significant improvements in HbA1c (average reduction, –0.88%) [32,38,45,47]. All of the nonsignificant trials also assessed the impact of the intervention on HbA1c (mean baseline level, 8.58%). Half of the trials reported significant between-group differences in HbA1c despite having no effect on medication adherence (average reduction, –0.83%) [37,39,44,48].
Intervention content. Health care practitioners were the primary focus in 3 of the trials [36,42,46]. One of the trials demonstrated a significant benefit from the intervention. In this RCT, Guldberg et al  randomized 86 general practices to receive either a CD ROM–based electronic feedback system or usual care. The electronic feedback system presented practitioners with data on their DM patient population which could be used for individual consultations or quality assurance. Medication adherence was assessed as the initiation and sustainment of treatment as documented in pharmacy records. At the 15-month follow-up, patients followed in intervention practices were significantly more likely to redeem prescriptions for their initial oral medication (mean change, 20.9%; 95% confidence interval [95CI], 7.9–34.8 [P = 0.002]) and insulin treatment (mean change, 21.4%; 95CI, 9.9–32.8 [p < 0.001]) than patients in the control practices . In the 2 nonsignificant trials, practitioners were trained to counsel patients on making shared treatment decisions for their DM by using either motivational interviewing techniques  or a decision aid tool .
Intervention delivery and dose. Participating practitioners delivered the intervention in all the trials, however, training in the intervention content, frequency of contacts, and adherence to the study protocol varied greatly. In the positive trial, the CD ROM–based feedback system was distributed at 3 time points over the course of the 15-month study . Qualitative data from participating practitioners showed that the system was most often used to assess overall quality of their diabetes care rather than for individual consultations . In the motivational interviewing trial, practitioners attended a 1.5-day training session with 2 half-day follow-up sessions over the course of 1 year . More than 90% of participating practitioners attended all of the trainings. This is in contrast to the trial employing a decision aid whereby practitioners received one training session for less than 3 minutes before meeting with the first enrolled patient . In regards to intervention dose, practitioners in the motivational interviewing trial were expected to deliver 3 sessions of 45 minutes each per patient . No data was provided on the percentage of sessions delivered or adherence to the counseling protocol. In the decision aid trial, practitioners were randomized to either use the tool with all of their eligible patients (intervention group) or discuss treatment as usual. Fidelity to the study protocol was assessed by coding videotaped clinical encounters with the OPTION decision-making instrument . While this study did not report a beneficial effect of the decision aid tool on medication adherence, intervention patients were rated as having significantly higher levels of involvement in the clinical encounter than the control group .
Assessment of medication adherence. All 3 trials utilized pharmacy records to assess medication adherence. One trial also used patient self-report to assess adherence to all diabetes medications . Comparisons of the measures showed that patients in both study groups tended to overreport nonadherence by self-report as compared with pharmacy record data (24% and 19% nonadherent by self-report in the intervention and usual care groups vs. 2.5% and 0% by pharmacy record). Baseline adherence levels were only assessed in the motivational interviewing trial; 96% of intervention patients and 100% of control patients were considered adherent by pharmacy records . Two trials collected data on both oral medications and insulin [36,42] and 1 trial assessed adherence solely to oral medications . Only the positive trial assessed the differential impact of the intervention on adherence by medication type . The intervention was equally beneficial for the initiation of oral and insulin prescriptions in intervention practices.
Impact on HbA1c. Most patients had baseline HbA1c levels lower than 8% in the practitioner-focused trials (mean, 7.27%), of which none demonstrated significant between-group differences in HbA1c. A significant within-group reduction in HbA1c was demonstrated in both the intervention and control groups among patients participating in the motivational interviewing trial (mean change, –0.7%, P < 0.01) .
A systematic review of the literature yielded 18 studies that tested the effects of intervention strategies delivered by health care practitioners on adherence to diabetes medications. Patient-focused trials comprised 83% of the reviewed interventions, and 17% were practitioner-focused. Eight of the 18 reviewed trials (44%) reported a statistically significant improvement in medication adherence attributed to the intervention [31,32,34,36,38,43,45,47]. The effective patient-focused interventions (n = 7) were complex, including a combination of patient education, individualized counseling targeting the psychosocial (eg, depression, self-efficacy) and logistical (eg, medication cost, referrals) barriers to adherence, medication reminders, action planning, and reinforcement of positive behavior. The effective practitioner-focused intervention utilized an electronic feedback system to increase the quality of care provided to patients with diabetes .
When comparing the intervention content across trials, we found that effective interventions were more likely to directly target medication nonadherence and be delivered in one-on-one sessions by trained ancillary staff through the use of care managers and/or technology than those reporting nonsignificant findings. Despite improvements in adherence, patients in the control group were equally likely to report high satisfaction with practitioners’ communication skills as the intervention group [34,48]. However, lack of baseline assessment of perceived satisfaction with practitioners’ communication skills make it difficult to determine if nonsignificant findings reflect ceiling effects or other extraneous confounding variables. When comparing the intervention dose, we found that clinic staff in the nonsignificant trials led 4½ times as many intervention sessions (average, 9 sessions vs. 2 sessions) over a longer duration of time (10 months vs. 2.5 months) than in the positive trials. This finding may be explained by the intervention content in the nonsignificant trials, with most primarily focusing on delivering patient education, which is a necessary but insufficient condition for lasting behavior change . For example, previous studies have shown that neither patient’s knowledge of their HbA1c level or high diabetes-related knowledge scores were associated with better medication adherence or reductions in Hb1Ac levels [51,52]. Rather, as demonstrated by the positive trials, patients are more likely to initiate and maintain self-management behaviors when they work collaboratively with health care practitioners to become an active, informed partner in their care . Specifically, patient-centered intervention strategies that provide tailored education and assist patients in developing problem solving and decision making skills through behavioral feedback and reinforcement equip patients with the knowledge, skills, and confidence to manage their health .
Few studies targeted practitioners making it difficult to draw conclusions about effective approaches to improve diabetes medication adherence. Findings from the positive trial suggests that provision of feedback on performance measures related to guideline-concordant care may be more effective (in the short term) for improving medication adherence than intervention approaches that try to modify practitioners’ communication skills. Finally, objective measures of medication adherence were more likely to be utilized in positive trials. However, trials employing objective measures were no more likely to report improvements in HbA1c than those utilizing patient self-report in both the positive and nonsignificant trials.
Four limitations in the methodology of the reviewed trials warrant consideration when examining the effects of interventions on medication adherence. First, the adherence measure varied with only 2 trials [32,34] reporting adherence with the objective electronic assessments. When both an objective and subjective measure of adherence was used, results between the 2 types of measures conflicted (eg, the objective measure rated the patient as less adherent but patients reported perfect adherence). This highlights one of the greatest challenges to the adherence field—no single measure can be considered the “gold standard” for all types of research or clinic settings.
Second, there was great variability in how medication adherence was defined across trials. Definitions of adherence were partly dependent on the measure used and ranged from the percent of prescribed doses taken correctly for objective measures to a summary score derived from self-report measures that use different recall periods. Frequency of the dosing regimen (once-daily vs. twice-daily dosing) and medication type (oral medication vs. injection) also affect the interpretation of one’s findings. For example, Tan et al  showed that failure to account for occurrences of hypoglycemia during the course of the study can underestimate the proportion of adherent patients.
Third, a limitation of the trials is the inclusion of patients irrespective of their baseline adherence status. That is, adherent and nonadherent patients form a single group and as a result are exposed to the same intervention approach. As a result, initially adherent patients maintain their high adherence levels and nonadherent patients do not get the dose or specificity of intervention strategies they may need to improve their behaviors. Trials that included initially adherent patients in this review were also less likely to detect changes in Hba1c, suggesting that other self-care behaviors are important intervention targets for these patients [31,34,35,41,48]. Lack of a baseline assessment of medication adherence in the negative practitioner-focused trials also makes it difficult to distinguish whether nonsignificant findings reflected an inability to alter practitioners’ skills, ceiling effects, or extraneous confounding variables.
Finally, most (75%) of the positive trials were 3 or fewer months in duration as compared with the nonsignificant trials, where 60% of trials lasted at least 12 months. Thus, the question remains whether positive findings were due to a Hawthorne effect in some of the trials. Replication of the findings in trials with longer follow-up periods and more robust assessments of medication adherence are needed to answer this question.
There are limitations of this review. We only included trials with published data in English-language journals and may have missed some trials that met the predefined criteria. We also limited our review to interventions to improve adherence to diabetes medications that are delivered by health care professionals thus, we may have excluded relevant work using different intervention strategies. However, we did include trials reporting nonsignificant findings thereby minimizing the negative publishing bias. Finally, heterogeneity of the interventions and measures used in the studies precluded quantitative meta-analysis of the findings.
Diabetes is a complex chronic condition that requires patients to develop the knowledge, skills, and confidence to execute daily self-management behaviors. A collaborative patient-practitioner interaction is at the forefront of effective patient self-management and can lay the foundation for patient’s acceptance and implementation of a medication regimen . However, increasing demands on practitioners to address patients’ complex medical needs within time-constrained visits greatly diminishes opportunities to engage in discussions about medication adherence. Furthermore, the current health care system is structured to support visit-based, acute care, making it difficult for practitioners to change their communication style or implement multi-component intervention strategies, even when provided with ample training opportunities.
Findings from this review suggest that medication adherence among patients with DM can be improved when practitioners are supported by trained ancillary staff who have the dedicated time to provide individualized patient-centered adherence counseling. Technology also provides an efficient and cost-effective means to enhance medication adherence by offering patients continuous monitoring, feedback, and reinforcement . Practitioners similarly benefit from technology-based systems that utilize continuous quality improvement principles to provide feedback on performance measures .
Application of these findings to real world clinic settings will require the adoption of new models of care such as the patient-centered medical home and panel management that facilitate the provision of high quality, patient-centered care for patients with chronic disease . Core elements of these models include team-based approaches that support physician decision-making by utilizing nonphysician support staff (eg, health coaches), health information technology, and structured work processes to deliver proactive, coordinated care . The Geisinger Health System and Group Health Cooperative serve as examples of innovative delivery systems that are implementing multidimensional adherence strategies that are producing clinical and cost-saving benefits . The lessons learned from these organizations will be invaluable as health care reform moves from law to reality for patients and health care practitioners alike.
Corresponding author: Dr. Antoinette Schoenthaler, Center for Healthful Behavior Change, Dept. of Population Health, NYU School of Medicine, 227 E. 30th St., 634, New York, NY 10016, firstname.lastname@example.org.
Funding/support: This work was supported by K23 HL 098564-01 from the National Heart, Lung, and Blood Institute.
Financial disclosures: None.
1. Center for Disease Control and Prevention. National Diabetes Fact Sheet: National estimates and general information on diabetes and prediabetes in the United States, 2011. Atlanta, GA; 2011.
2. Ho PM, Rumsfeld JS, Masoudi FA, et al. Effect of medication nonadherence on hospitalization and mortality among patients with diabetes mellitus. Arch Intern Med 2006;166:1836–41.
3. Cramer JA. A systematic review of adherence with medications for diabetes. Diabetes Care 2004;27:1218–24.
4. Odegard PS, Capoccia K. Medication taking and diabetes: A systematic review of the literature. Diabetes Educ 2007;33:1014–29.
5. Osterberg L, Blaschke T. Adherence to medication. N Engl J Med 2005;353:487–97.
6. Peterson AM, Takiya L, Finley R. Meta-analysis of trials of interventions to improve medication adherence. Am J Health Syst Pharm 2003;60:657–65.
7. DiMatteo MR, Giordani PJ, Lepper HS, Croghan TW. Patient adherence and medical treatment outcomes: A meta-analysis. Med Care 2002;40:794–811.
8. World Health Organization: Adherence to long-term therapies: evidence for action. Geneva: World Health Organization; 2003.
9. Haynes R, Taylor D, Sackert D, editors. Compliance in health care. Baltimore: Johns Hopkins University Press; 1979.
10. Miller NH, Hill M, Kottke T, Ockene IS. The multilevel compliance challenge: recommendations for a call to action. A statement for healthcare professionals. Circulation 1997;95:1085–90.
11. Heisler M, Bouknight RR, Hayward RA, et al. The relative importance of physician communication, participatory decision making, and patient understanding in diabetes self-management. J Gen Intern Med 2002;17:243–52.
12. Parchman ML, Zeber JE, Palmer RF. Participatory decision making, patient activation, medication adherence, and intermediate clinical outcomes in type 2 diabetes: A STARNet study. Ann Family Med 2010;8:410–7.
13. Fuertes JN, Mislowack A, Bennett J, et al. The physician-patient working alliance. Patient Educ Couns 2007;66:29–36.
14. van Dam HA, van der Horst F, van den Borne B, et al. Provider-patient interaction in diabetes care: Effects on patient self-care and outcomes. A systematic review. Patient Educ Couns 2003;51:17–28.
15. Ratanawongsa N, Karter AJ, Parker MM, et al. Communication and medication refill adherence: The Diabetes Study of Northern California. JAMA 2013;173:210–8.
16. Piette JD, Schillinger D, Potter MB, Heisler M. Dimensions of patient-provider communication and diabetes self-care in an ethnically diverse population. J Gen Intern Med 2003;18:624–33.
17. Schneider J, Kaplan SH, Greenfield S, et al. Better physician-patient relationships are associated with higher reported adherence to antiretroviral therapy in patients with HIV infection. J Gen Intern Med 2004;19:1096–103.
18. McDonald HP, Garg AX, Haynes RB. Interventions to enhance patient adherence to medication prescriptions: Scientific review. JAMA 2002;288:2868–79.
19. Higgins J, Green S, editors. Cochrane Handbook for systematic reviews of interventions. 5.1.0 ed [updated March 2011]. The Cochrane Collaboration; 2011.
20. Chisholm-Burns MA, Graff Zivin JS, Lee JK, et al. Economic effects of pharmacists on health outcomes in the United States: A systematic review. Am J Health Syst Pharm 2010;67:1624–34.
21. Chisholm-Burns MA, Kim Lee J, Spivey CA, et al. US pharmacists’ effect as team members on patient care: systematic review and meta-analyses. Med Care 2010;48:923–33.
22. Hatah E, Braund R, Tordoff J, Duffull SB. A systematic review and meta-analysis of pharmacist-led fee-for-services medication review. Br J Clin Pharmacol;2013.
23. Hatah E, Tordoff J, Duffull SB, Braund R. Pharmacists’ performance of clinical interventions during adherence support medication reviews. Res Soc Admin Pharm 2013.
24. Machado M, Bajcar J, Guzzo GC, Einarson TR. Sensitivity of patient outcomes to pharmacist interventions. Part I: Systematic review and meta-analysis in diabetes management. Ann Pharmacother 2007;41:1569–82.
25. McKee MD, Fletcher J, Sigal I, et al. A collaborative approach to control hypertension in diabetes: Outcomes of a pilot intervention. J Primary Care Comm Health 2011;2:148–52.
26. Pearce KA, Love MM, Shelton BJ, et al. Cardiovascular risk education and social support (CaRESS): report of a randomized controlled trial from the Kentucky Ambulatory Network (KAN). J Am Board Family Med 2008;21:269–81.
27. Zolfaghari M, Mousavifar SA, Pedram S, Haghani H. The impact of nurse short message services and telephone follow-ups on diabetic adherence: which one is more effective? J Clin Nursing 2012;21:1922–31.
28. Hendricks LE, Hendricks RT. The effect of diabetes self-management education with frequent follow-up on the health outcomes of African American men. Diabetes Educ 2000;26:995–1002.
29. Wolever RQ, Dreusicke M, Fikkan J, et al. Integrative health coaching for patients with type 2 diabetes: a randomized clinical trial. Diabetes Educ 2010;36:629–39.
30. Crowley MJ, Powers BJ, Olsen MK et al. The Cholesterol, Hypertension, And Glucose Education (CHANGE) study: Results from a randomized controlled trial in African Americans with diabetes. Am Heart J 2013;166:179–86.
31. Ahmadipour H, Farajzadegan Z, Kachoei A, Pirdehghan A. Secondary prevention by enhancing adherence in diabetic patients. Int J Prevent Med 2010;1:50–5.
32. Bogner HR, Morales KH, de Vries HF, Cappola AR. Integrated management of type 2 diabetes mellitus and depression treatment to improve medication adherence: A randomized controlled trial. Ann Family Med 2012;10:15–22.
33. Edelman D, Fredrickson SK, Melnyk SD, et al Medical clinics versus usual care for patients with both diabetes and hypertension: a randomized trial. Ann Intern Med 2010;152:689–96.
34. Farmer A, Hardeman W, Hughes D, et al. An explanatory randomised controlled trial of a nurse-led, consultation-based intervention to support patients with adherence to taking glucose lowering medication for type 2 diabetes. BMC Fam Pract 2012;13:30.
35. Farmer AJ, Wade AN, French DP, et al. Blood glucose self-monitoring in type 2 diabetes: a randomised controlled trial. Health Technol Assess 2009;13:1–50.
36. Guldberg TL, Vedsted P, Kristensen JK, Lauritzen T. Improved quality of type 2 diabetes care following electronic feedback of treatment status to general practitioners: A cluster randomized controlled trial. Diabetic Med 2011;28:325–32.
37. Heisler M, Vijan S, Makki F, Piette JD. Diabetes control with reciprocal peer support versus nurse care management: A randomized trial. Ann Intern Med 2010;153:507–15.
38. Kim HS, Kim NC, Ahn SH. Impact of a nurse short message service intervention for patients with diabetes. J Nurs Care Qual 2006;21:266–71.
39. Kim HS, Oh JA. Adherence to diabetes control recommendations: impact of nurse telephone calls. J Adv Nurs 2003;44:256–61.
40. Ko GT, Li JK, Kan EC, Lo MK. Effects of a structured health education programme by a diabetic education nurse on cardiovascular risk factors in Chinese Type 2 diabetic patients: A 1-year prospective randomized study. Diabetic Med 2004;21:1274–9.
41. Lin EH, Katon W, Rutter C, et al Effects of enhanced depression treatment on diabetes self-care. Ann Family Med 2006;4:46–53.
42. Mullan RJ, Montori VM, Shah ND, et al. The diabetes mellitus medication choice decision aid: a randomized trial. Arch Intern Med 2009;169:1560–8.
43. Negarandeh R, Mahmoodi H, Noktehdan H, et al. Teach back and pictorial image educational strategies on knowledge about diabetes and medication/dietary adherence among low health literate patients with type 2 diabetes. Prim Care Diabetes 2013;7:111–8.
44. Piette JD, Weinberger M, Kraemer FB, McPhee SJ. Impact of automated calls with nurse follow-up on diabetes treatment outcomes in a Department of Veterans Affairs Health Care System: A randomized controlled trial. Diabetes Care 2001;24:202–8.
45. Piette JD, Weinberger M, McPhee SJ, et al. Do automated calls with nurse follow-up improve self-care and glycemic control among vulnerable patients with diabetes? Am J Med 2000;108:20–7.
46. Rubak S, Sandbaek A, Lauritzen T, et al. Effect of “motivational interviewing” on quality of care measures in screen detected type 2 diabetes patients: A one-year follow-up of an RCT, ADDITION Denmark. Scand J Primary Health Care 2011;29:92–8.
47. Tan MY, Magarey JM, Chee SS, et al. A brief structured education programme enhances self-care practices and improves glycaemic control in Malaysians with poorly controlled diabetes. Health Educ Res 2011;26:896–907.
48. Wong FK, Mok MP, Chan T, Tsang MW. Nurse follow-up of patients with diabetes: Randomized controlled trial. J Adv Nurs 2005;50:391–402.
49. Guldberg TL, Vedsted P, Lauritzen T, Zoffmann V. Suboptimal quality of type 2 diabetes care discovered through electronic feedback led to increased nurse-GP cooperation. A qualitative study. Prim Care Diabetes 2010;4:33–9.
50. Beukelman D. Information alone does not change behavior. In: Michigan Consortium on Assistive Device Delivery Systems: Resource Directory. 2nd ed. 1990:79-82.
51. Heisler M, Piette JD, Spencer M, et al. The relationship between knowledge of recent HbA1c values and diabetes care understanding and self-management. Diabetes Care 2005;28:816–22.
52. Schoenthaler AM, Schwartz BS, Wood C, Stewart WF. Patient and physician factors associated with adherence to diabetes medications. Diabetes Educ 2012;38:397–408.
53. Hibbard JH, Mahoney ER, Stock R, Tusler M. Do increases in patient activation result in improved self-management behaviors? Health Serv Res 2007;42:1443–63.
54. Funnell MM, Anderson RM, Arnold MS, et al. Empowerment: an idea whose time has come in diabetes education. Diabetes Educ 1991;17:37–41.
55. Piette JD. Interactive behavior change technology to support diabetes self-management: Where do we stand? Diabetes Care 2007;30:2425–32.
56. Bu D, Pan E, Walker J, et al. Benefits of information technology-enabled diabetes management. Diabetes Care 2007;30:1137–42.
57. Meyers D, Peikes D, Genevro J, et al. The roles of patient-centered medical homes and accountable care organizations in coordinating patient care. Rockville, MD: Agency for Healthcare Research and Quality; 2010.
58. Margolius D, Bodenheimer T. Transforming primary care: From past practice to the practice of the future. Health Aff 2010;29:779–84.
59. Cutler DM, Everett W. Thinking outside the pillbox–medication adherence as a priority for health care reform. N Engl J Med 2010;362:1553–5.
« Does Bioelectrical Impedance Analysis Provide a Reliable Diagnosis of Secondary Lymphedema in Breast Cancer Patients? Addressing the Sexual Health Concerns of Women with Gynecologic Cancer: Guidance for Primary Care Physicians »