Adherence, control of cardiometabolic factors and therapeutic inertia in patients with type 2 diabetes in the primary care setting

Abstract Introduction Studies on treatment adherence to glucose‐lowering drugs among patients with type 2 diabetes (T2D) including concomitant treatment for other cardiovascular risk factors are scarce. We aimed to estimate the prevalence of good adherence to all medications used to control diabetes, hypertension and dyslipidemia and to analyse cardiometabolic control and its associated factors in T2D patients in the primary care (PC) setting. Methods Observational, retrospective study conducted in adult patients with T2D who were followed in the PC setting in Spain. Patients were classified as adherent in a particular category if the summary of the proportion of days covered (PDC) for a particular medication category was ≥80% and were considered globally adherent if the PDC was ≥80% for each of the 3 medication categories. Results A total of 457 evaluable patients were recruited, among which 321 patients (70.3%, 95% CI 65.8 to 74.4) were adherent to the three drug categories. The proportion of patients controlled for the 3 cardiometabolic risk factors was 31% according to the contemporary clinical practice guideline criteria, 58% according to investigator judgment and 36% when the objective for HbA1c was individualized. In a multivariate analysis, presenting comorbidities was associated with a lower likelihood of showing adequate control of dyslipidemia (odds ratio [OR] 0.25, 95% CI, 0.16–0.40) and the three cardiometabolic factors as a whole (OR 0.43, 95% CI 0.26–0.70). In a post hoc analysis, therapeutic inertia was found to be greater for dyslipidemia and hypertension than for T2D. Conclusions Despite a relatively high adherence to all medications for treating diabetes, hypertension and dyslipidemia in patients with T2D in the PC setting in Spain, the control of cardiometabolic risk factors as a whole is far from optimal. This could be related, at least in part, to the high frequency of comorbidity of these patients.


| INTRODUC TI ON
Adherence to medication is defined as the extent to which patients take medications as prescribed by their health care professionals. 1 According to the World Health Organization, poor adherence in individuals with chronic illnesses that require long-term treatment is a worldwide problem with rates that average 50%. Poor adherence is associated with poor health outcomes and increased health care costs and compromises health system effectiveness overall. 2 Poor adherence is a multifactorial phenomenon involving socioeconomic, health care system, patient-related, disease-related and treatment-related factors. 1,3 Among disease-and treatment-related factors, the presence of comorbid conditions and the complexity of treatment regimens are recognized factors that can negatively affect treatment adherence. 1,3 Diabetes is a chronic disease that requires a stepwise treatment approach, with the treatment becoming more complex as the disease progresses. 4 Thus, among ambulatory physician practices in the USA in 2012, in 58% of the visits for diabetes, patients were receiving two or more antihyperglycemic medications. 5 This situation is further complicated by the fact that, consistent with guideline recommendations, 6 patients with diabetes require treatments for the management of cardiovascular risk factors, including hypertension and dyslipidemia. When accounting for all prescribed medications, in 2000, the annual proportion of primary care visits of patients with diabetes listing at least 5 prescription medicines was 30%, 7 with blood pressure and lipid-lowering drugs as major contributors to this increasing complexity of pharmacologic regimens. 8 This situation may explain in part why treatment adherence in patients with diabetes is poor, being one of the chronic diseases with the lowest rate of adherence. 9,10 Overall, good adherence to diabetes treatment ranges from 31% to 80%, [9][10][11][12][13][14][15][16] depending on the study design, prescribed drug, extent of the follow-up and/or definition of adherence.
Poor adherence to antihyperglycemic treatments has been associated with poorer diabetes control, 12,17 increased risk of all-cause hospitalization 15 and all-cause mortality. 15,18 Despite all the available information on adherence to antihyperglycemic agents, much less is known about adherence to the whole complex regimen that patients with diabetes require for the treatment of their disease and the management of cardiovascular risk factors. Lopez-Simarro et al. 13 reported a rate of nonadherence among 320 patients with diabetes seen in primary care in Spain of 36%, 38% and 32% for medications for diabetes, hypertension and dyslipidemia, respectively. Additionally, in the same study, the proportion of patients with T2D with good control of HbA1c, blood pressure and LDLcholesterol was 62.5%, 40.9% and 35.9%, respectively. 19 Lower rates of nonadherence were reported by Ho et al. 15 in patients with diabetes in a US managed care organization (20%, 19% and 25% for antihyperglycemics, antihypertensives and statins, respectively). Many clinical guidelines and expert committees recommend the individualization of glycaemic targets and treatment decisions in the management of type 2 diabetes (T2D) depending on patient's preferences and characteristics, such as frailty or comorbid conditions, with special interest in the presence of cardiovascular or renal disorders. 20,21 This approach has the goal of reducing complications and maintaining quality of life in the context of comprehensive cardiovascular risk management and patient-centred care. We performed a retrospective study whose primary objective was to estimate the prevalence of good adherence to all medications used to control diabetes, hypertension and dyslipidemia in patients with T2D attending PC centres. Secondary objectives included analysing adherence within each of the therapeutic groups; comparing the levels of HbA1c, blood pressure and LDLc between adherent and nonadherent patients; and comparing the proportion of patients controlling these three risk factors between adherent and nonadherent patients in this population. A post hoc objective was to evaluate therapeutic inertia for the treatment of diabetes, hypertension and dyslipidemia and to analyse associated factors.

| Study design, setting and patients
This was an observational, retrospective study. The study was approved by the Ethics Committee of each participating site. Eighty primary care physicians throughout Spain recruited patients consecutively during a single inclusion visit. The index date was established as the date 365 days before the date of the inclusion visit. To be included in the study, patients had to be 18 years or older; have been diagnosed with T2D; be followed by a primary care physician; be prescribed oral antihyperglycemic drugs (OAHDs), antihypertensive drugs for the treatment of hypertension and lipid-lowering drugs for the treatment of dyslipidemia for at least 12 months; and not have changed their residence in the last 12 months. Patients were excluded if they were unable to provide their written informed consent; were dependent; had been participating in a clinical trial at any time during the 1-year study period; had a psychiatric disorder other than a depressive or anxiety disorder; had a severe or terminal disease; were receiving insulin or glucagon-like peptide 1 receptor agonists; or became pregnant or were diagnosed with ketoacidosis, malnutrition-associated diabetes, drug-induced diabetes or gestational diabetes during the 1-year study period.

| Medications and estimation of treatment adherence
Information on medication received was obtained from the Spanish electronic medical prescription system (eReceta). eReceta was K E Y W O R D S adherence, diabetes, dyslipidemia, therapeutic inertia initiated in 2004 and was implemented across all autonomous regions in Spain; the system relies on the patient's electronic health card and the link of the card to his/her medical records in several databases. Medication adherence was calculated as the proportion of days covered (PDC). The PDC was based on the filled eprescriptions during the 1-year study period for each of the three categories of medications: OAHDs, antihypertensives and lipidlowering drugs. To calculate the PDC, we estimated the total days of supplies from the first refill to the last refill during the 1-year observation period, divided by the total days of the treatment interval; the treatment interval was defined as the time elapsed from the date of the first refilled e-prescription to the end of the observation period regardless of whether the patient was maintained on the first drug prescribed or was switched to another drug or, if the drug was discontinued by the physician, until the date the primary care physician recommended discontinuing the drug. The resulting figure was multiplied by 100 to estimate the percentage of PDC.
The PDC was averaged for all drugs within a category. Patients were categorized as adherent to a particular category if the summary PDC for that category was 80% or greater. Patients were considered globally adherent if the PDC was ≥80% for each of the three medication categories.

| Demographics, clinical assessments and definition of disease control and therapeutic inertia
Demographics and clinical data were obtained from the patient's electronic clinical record or, if not available, at the time of the inclusion visit. Demographics included age, sex, race and household status. Clinical data included data from the physical examination (weight, height and body mass index) and diabetes-related complications (retinopathy, nephropathy, neuropathy and diabetic foot) and comorbidities (coronary artery disease, heart failure, peripheral occlusive arterial disease, cerebrovascular disease, depression, osteoarthritis, chronic obstructive pulmonary disease and others); regarding diabetes-related complications and comorbidities, whether they were present at the time of index date or had occurred during the 1-year observation period was recorded.
Information on the degree of cardiometabolic control (ie HbA1c, blood pressure and LDLc) was obtained from the electronic clinical record if it was recorded within 3 months before the inclusion visit or was measured at the time of the inclusion visit by measuring blood pressure and/or performing blood extraction for laboratory analysis.
Glycaemic control was evaluated according to the clinical practice guidelines and based on individualized criteria. Patients were considered controlled according to the clinical practice guideline criteria if HbA1c was <7.0% 22 and based on individualized criteria: if patients were younger than 75 years old with less than 10 years of diabetes duration and had no diabetes-related comorbidities (ie coronary heart disease, heart failure or occlusive peripheral arterial disease) or complications (ie retinopathy, nephropathy, neuropathy or diabetic foot), they were considered controlled if HbA1c was <6.5%; and if patients met any of the latter criteria, they were considered controlled if HbA1c was <7.5%. The guideline-based criteria for categorizing patients as having adequate disease control in terms of hypertension and dyslipidemia were as follows: systolic blood pressure <140 mm Hg and diastolic blood pressure <90 mm Hg 22 and LDLc <100 mg/dl for primary prevention and <70 mg/dl for secondary prevention (as per contemporary guidelines). 23 In addition, physicians were asked to assess whether, regardless of the actual values of patients' cardiometabolic parameters and based on the clinical characteristics, they considered that the patient was controlled for each of these 3 factors.
We considered that there was therapeutic inertia if a patient was adequately controlled based on the physician's criteria but was not controlled according to the cardiometabolic parameters as previously described. 24

| Statistical analysis
To achieve 5% statistical precision in the estimation of a population proportion with an asymptotic normal 95% confidence interval and assuming a nonadherence proportion of 50%, it would be necessary to include 384 patients in the study; assuming a 10% rate of nonevaluable patients, it would be necessary to recruit 428 subjects.
The characteristics of the recruited population are presented with means and standard deviations for continuous variables and absolute and relative frequencies for qualitative variables. To describe the prevalence, the point estimate and the corresponding 95% confidence interval are provided. In the bivariate analysis, characteristics of the adherent and nonadherent patients were compared using Student's t test or the Mann-Whitney test for continuous variables and the chi-squared test or Fisher's exact test for categorical variables.
To explore the factors associated with global adherence, a multiple logistic regression analysis was performed. The dependent variable was global adherence, and the independent variables were age, BMI, HbA1c, LDLc, blood pressure, number of medicines, To explore factors associated with therapeutic inertia for each of the cardiometabolic risk factors, we performed four multiple logistic regression analyses using the presence of therapeutic inertia for each cardiometabolic factor (in the case of T2D using the clinical practice guideline and individualized criteria for defining disease control) as dependent variables; the independent variables were age; sex; BMI; diabetes duration; HbA1c; LDL; systolic blood pressure; diastolic blood pressure; number of daily pills for treating T2D, hypertension and dyslipidemia; total number of pills; whether the patients were   Table 1. Patients were elderly and almost evenly distributed according to sex. The mean T2D duration was 7.3 years, and the mean BMI was 30.9 kg/m 2 . On average, patients were receiving over 4 oral medications for the treatment of T2D, hypertension and dyslipidemia, and overall, they were receiving almost 10 drugs on average ( Table 1). The most frequent T2D-associated complication was nephropathy (12%), and the most frequent comorbidities were osteoarthritis (44%) and, to a lesser extent, depression (19%) and coronary artery disease (18%).

| Adherence to medications for the management of cardiometabolic risk factors
Overall, 321 patients (70.2%, 95% CI 65.8 to 74.4) were adherent to all three drug categories, namely oral antihyperglycemic agents, antihypertensives and lipid-lowering drugs; of the remaining patients, 86 (18.8%) were adherent to two pharmacologic categories, 33 (7.2%) were adherent to only one pharmacologic category, and 17 (3.7%) were completely nonadherent. Adherence was over 80% for each individual pharmacologic category (Figure 1).

| Control of cardiometabolic risk factors and their association with adherence
The proportions of patients who were considered to have controlled cardiometabolic risk factors according to the clinical practice guideline criteria, investigator judgment or individualized objectives for HbA1c are presented in Figure 2. Regardless of the cardiometabolic risk factor, the proportion of patients with a particular risk factor controlled was higher according to investigator judgment than according to the clinical practice guideline (CPG) criteria. The proportion of patients with all three cardiometabolic risk factors controlled was 31% according to the clinical practice guideline criteria, 58% according to investigator judgment, and 36% when the HbA1c objective was individualized.
In the bivariate analysis (Table 2), there was no difference be-  (Table 2).
In the multivariate analysis (Table 3), global adherence was not as-  (Table 3).

When control was evaluated based on individualized criteria for
HbA1c, among those who had adequate control of the three risk factors, 30.8% were globally nonadherent compared with 69.2% who were globally adherent.

| Therapeutic inertia
Therapeutic inertia was greater for dyslipidemia and hypertension than for T2D (Figure 4). Therapeutic inertia for the treatment of T2D dif- In the multivariate analyses, we did not find any factor associated with therapeutic inertia for the treatment of T2D when individualized criteria for HbA1c were considered. Age was directly associated with a higher likelihood of therapeutic inertia for the treatment of were not necessarily exhibiting hypertension and dyslipidemia, and they could be receiving insulin). 15 Lopez-Simarro et al. 13

in a similar
setting as that in our study, reported rates of nonadherence for the individual components of the treatment of cardiometabolic risk factors that were higher than those found in our study. They found that 36%, 38% and 32% of participants were nonadherent to medications for the treatment of diabetes, hypertension and dyslipidemia, respectively, 13 while our corresponding figures were 17%, 11%, and 15%, respectively. In addition, in the same study, patients with T2D with good control for HbA1c, cLDL and blood pressure were more adherent to the respective drug classes, although there was not a statistically significant relationship between the control of those risk factors and therapeutic adherence. 19 Ho et al. 15  F I G U R E 4 Therapeutic inertia in patients with diabetes for diabetes, hypertension and dyslipidemia. Therapeutic inertia exists if a patient was adequately controlled based on the physician criteria but he/she was not controlled according to the actual cardiometabolic parameters. When interpreting the actual cardiometabolic parameters, patients were considered to have controlled diabetes (a) according to the clinical practice guidelines if HbA1c was <7.0%; (b) based on individualized criteria, if patients were younger than 70 years old with less than 10 years of diabetes duration and had no diabetes-related comorbidities (ie coronary heart disease, heart failure or occlusive peripheral arterial disease) or complications (ie retinopathy, nephropathy, neuropathy or diabetic foot), they were considered controlled if HbA1c was <6.5%; and if patients met any of the latter criteria, they were considered controlled if HbA1c was <7.5. The criteria for categorizing patients as having adequate disease control regarding hypertension and dyslipidemia were as follows: systolic blood pressure <140 mm Hg and diastolic blood pressure <90 mm Hg (19); and an LDLc <100 mg/dl for high-risk patients and <70 mg/dl for very high-risk patients not only impairs quality of life and is associated with greater health care utilization 27,28 but, together with chronic kidney disease, comorbidities are also key contributors to mortality among patients with diabetes. 29 It is also well known that targeting these multiple risk factors in patients with T2D is associated with a reduced risk of cardiovascular and microvascular events. 30 In our multivariate analysis, global adherence was not associated with adequate control of the three cardiometabolic risk factors, and the only variable associated with adequate control was the presence of comorbidities, which was associated with a 57% reduction in the likelihood of those factors being adequately controlled. This finding is consistent with the previous results that found that some comor-

ACK N OWLED G M ENTS
The authors thank to Fernando Rico-Villademoros (COCIENTE S.L., Madrid, Spain), who provided medical writing assistance. This assistance was funded by MSD Spain. The study was funded and sponsored by MSD Spain, a subsidiary of Merck & Co., Inc., Kenilworth, NJ, USA.

CLI N I C A L TR I A L R EG I S TR ATI O N
Not applicable.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.