Drugs don’t work in patients who don’t take them

Adherence to the prescribed treatment and the retention of participants for follow‐up evaluation (which obviously impacts adherence) can and will likely affect the validity and statistical power of a clinical trial.1 Medication adherence is defined as the extent to which a patient takes prescribed medications according to the dosage and frequency recommended.2 In a clinical trial, it is the degree to which a study participant follows the treatment protocol. Maximizing adherence in a trial involves a multi‐faceted strategy. Attention to adherence and retention ideally begins even before a participant is enrolled in the trial, that is, during the screening process and continues throughout the trial. Potential participants who cannot attend or complete study visits and/or take study medication as indicated should not be enrolled. Even in a well‐designed trial, nonadherence to protocol‐specified intervention/treatment can lead to underestimation of the possible therapeutic and toxic effects. Retention refers to continued involvement of study participants through the intervention and clinic follow‐up phases of the protocol, including attendance at clinic visits, participation in telephone contacts, and allowing the completion of medical assessments and procedures. Loss to follow‐up in randomized trials can compromise study findings by reducing the power of a study to detect a true difference between the intervention and control groups. Differential loss to follow‐up may lead to bias through exaggerated effects in favor of one of the trial groups. Like poor adherence and retention, loss to follow‐up can also affect the studies validity, and affect the generalizability of the trial results.3 In a recent study of Medicare beneficiaries, adherence to statin therapy among Medicare those ≥66 years of age with a MI hospitalization in 2007‐2010 who filled a high‐intensity statin (atorvastatin 40‐80 mg, rosuvastatin 20‐40 mg) within 30 days following discharge were analyzed. At 182 and 730 days after discharge, 54.8% and 36.7% of beneficiaries, respectively, remained on high‐intensity statins with high adherence. Whites, beneficiaries with fewer comorbidities, low‐income subsidy, and a high‐intensity statin fill prior to their MI and those who had cardiac rehabilitation and cardiologist visits post‐MI were more likely to remain on high‐intensity statins with high adherence.4

Studies have shown that medication nonadherence remains highly prevalent and is associated with adverse outcomes and higher health care costs.5 For example, one study demonstrated that a year after sustaining an acute myocardial infarction, less than 50% of patients are filling prescriptions for statin, beta‐blocker, and angiotensin‐converting‐enzyme inhibitor medications, even when these medications are offered with no copay.6 In 2003, a World Health Organization statement noted that improved medication adherence “may have a far greater impact on health … than any improvement in specific medical treatments.”7 In one study, it was reported that medication adherence interventions among patients with heart failure were found to significantly reduce mortality (relative risk 0.89, 95% CI: 0.81‐0.99) and to decrease odds for hospital readmission (odds ratio 0.79, 95% CI: 0.71‐0.89); however, other studies were more heterogeneous, and not as conclusive.8 A meta‐analysis regarding adherence to drugs prescribed for the primary or secondary prevention of coronary artery disease showed that approximately one‐third of patients with myocardial infarction and approximately half of patients in primary prevention did not adhere to their medication.9 This problem is also common in the treatment of other cardiovascular diseases such as heart failure, arrhythmia, and hypertension.10 Data from the Reduction of Atherothrombosis for Continued Health (REACH) Registry examined regional differences in cardiovascular medication use and found geographic differences.11 Moreover, compared with the report of Chronic Heart Failure Analysis and Registry in the Tohoku District‐2 (CHART‐2) for the Northeast Region of Japan, their patients had fewer ischemic heart diseases and more non‐ischemic cardiomyopathy. Furthermore, their study was characterized by fewer complications of hypertension and dyslipidemia.13 In many studies (but not all), medication adherence has been noted to decline, as the number of daily doses increases. A meta‐analysis on the effect of the dosing frequency of chronic cardiovascular drugs on adherence also showed that patients were more adherent with once‐daily dosing compared with more frequent daily dosing regimens (two to four times a day).14 Other common reasons for nonadherence include forgetfulness, a personal decision to omit doses, a misunderstanding about medication dosing and emotional factors such as depression, that is, also thought to be a known risk factor for nonadherence to medical treatment. However, Berntson et al reported that depression, as defined by the PHQ‐9, was not associated with self‐reported poorer adherence to antihypertensive drugs or cholesterol‐lowering drugs in a large sample representative of the US population (the 2005‐2010 National Health and Nutrition Examination Survey).15 In the study reported in this Journal, a study of Japanese patients with CVD, more than 2× daily dosing frequency, age <65 years, and active employment were significantly associated with nonadherence (OR 4.42 CI 3.02‐6.48; 1.7 CI 1.23‐2.35, and 1.43 CI1.03‐1.99), respectively; but, depression was not a significant factor in nonadherence.16

Adherence has frequently been defined using a cut‐off point of ≥80%. Self‐reported nonadherence rates (commonly determined using the Morisky scale) are usually lower than the rates obtained with other methods, such as electronic monitoring, because self‐reports are known to be limited by a subject's reluctance to disclose nonadherence.12 Self‐report adherence scales are considered relatively simple and economically feasible to use, and can have the added advantage of soliciting information regarding situational factors which may act as barriers to medication adherence (eg, forgetfulness). For several of these self‐report tools, high reliability and validity have been reported. One of the most commonly used measures of antihypertensive medication adherence is the 8‐item Morisky Medication Adherence Scale (MMAS‐8). This scale has been used as a self‐administered questionnaire as well as through telephone administration. Items on this scale reflect reasons for nonadherent behavior (eg, forgetfulness, health beliefs, and side‐effects) and thus are useful in enabling study personnel to tailor their strategy for improving adherence to the underlying reason for nonadherence.

Despite the prior mentioned clinical definition of medication adherence (>80% compliance), there remains no standard operational definition for medication adherence in health research, especially when using pharmacy claims data. In the literature, there are a wide number of terms and operational definitions used to assess medication adherence. Even when studies use the same terms, the operationalized or practical definition presented, the time frame considered, and the method of calculating adherence often differ. While the level of optimal adherence may differ for different clinical conditions, as prior mentioned, a threshold of 0.80 is conventionally used. The methods of managing multiple concurrent medications (“polytherapy”) to obtain a single measure of adherence for each patient is also problematic, as there are many different operational definitions and methods used to handle polytherapy, and this may result in very different adherence estimates and highly variable conclusions, highlighting the need to formulate an agreeable operational definition of adherence. Also, it is generally agreed that adherence is greater in closely monitored clinical trials than in clinical practice.

In the current issue of JCC, Lefort et al report on gender differences in adherence to antihypertensive treatment in a study of French men and women >55 years of age.16 (gender differences in medication adherence have not been well studied). Their study is a cross‐sectional survey based upon 20 000 households recruited by quota sampling “in order to ensure representativeness of the French population.” Although the Morisky Scale was not used, the Girerd compliance test (a validated Scale used in many European studies) was used.17 The cross‐sectional design and the fact that it is a survey limit the usefulness of the findings. Additionally, generalization beyond French citizens over age 55 is obvious. Finally, as the author's state, the results of their study should be interpreted with caution due to the cross‐sectional sectional design and self‐declared adherence. An original finding as stated by the authors is the gender difference in the threshold of the number of anti‐HTN tablets when adherence drops. Specifically, adherence decreased gradually for each additional tablet in men whereas in women, adherence only dropped when taking three or more tablets. More studies of adherence are needed and the need to learn how to improve adherence even more so. As to the present large study, this information should be added to the accumulating knowledge about medication adherence.

The inclusion of diverse populations in clinical trials is central to generating generalizable findings. Although not included in the NIH definition of “minorities,” it has been recognized that women and the elderly have also not been adequately represented in past clinical trials. For instance, Cherubuni et al investigated the extent of exclusion of older individuals in clinical trials of heart failure, a condition in which the elderly is inordinately represented.18 They found that among 251 trials, 64 (25.5%) excluded patients by an arbitrary upper age limit, and that such exclusions were more common among European trials than those conducted in the United States. Overall, they concluded that “109 trials (43.4%) had 1 or more unjustified exclusion criteria that could limit the inclusion of older individuals.”

In conclusion, close examination of predictors of adherence and retention can serve future investigations by aiding in the development of new strategies for adherence and retention by identifying risk factors for nonadherence and study discontinuation, and by helping to better understand biases that result from low adherence and retention in these risk factor groups.

CONFLICT OF INTEREST

None.