Abstract
In November 2006, a panel of experts participated in a teleconference to discuss the issue of adherence to antihypertensive therapy. The panel was chaired by JoAnne M. Foody, MD, of the Yale University School of Medicine, Section of Cardiovascular Medicine, New Haven, CT. Also participating were Joshua S. Benner, PharmD, ScD, ValueMedics Research, Falls Church, VA, and William Frishman, MD, New York Medical College, Valhalla, NY.
DR FOODY: Unfortunately, a large proportion of patients for whom we prescribe cardiovascular medications discontinue their therapies or take the drugs inappropriately. And we know from large cohort studies in a variety of patient populations that there are very high rates of discontinuation of these therapies and ultimately poor adherence to drug regimens. Not surprisingly, given this, we fail to reach the kinds of targets we are asked to achieve through national guidelines, whether it be with respect to lipid lowering or blood pressure, in our current day‐to‐day practices.
Importantly, this nonadherence in our patient populations has significant impact on our patients' outcomes. We know that patients who don't take their medications as prescribed are much more likely to be hospitalized than patients who do follow their prescriptions. We also know that patients with low adherence rates have greater health care costs and health care burdens, even taking into account the savings in drug costs, than patients who have greater adherence.
So given the significant burden of cardiovascular disease and the known benefits of the therapies that we prescribe but the failings of our current system to promote adherence in our patients, we need to improve adherence in our patients, which is likely to have a major impact on outcomes.
Dr Benner, the terms compliance, adherence, and persistence are confusing terms to many clinicians. Could you briefly define these terms?
DR BENNER: The term patient compliance was the standard term to describe patients' medication use for a number of years and continues to be used in many circles because the term is the Medical Subject Heading (MESH) that's used by the National Library of Medicine to classify studies in MedLine on this topic. But the word compliance has been supplanted in recent years by adherence and persistence. One reason for this trend was that behavioral psychologists rightly pointed out that the decision to continue medications is often made by patients with influence from not only their doctor, but also their pharmacist, nurse, family, and caregivers. And since an important limitation of the most commonly used research methods is that we often can't ascribe the gaps in therapy or discontinuation to a specific reason, adherence and persistence have come into favor among many researchers. In our research, we use adherence when talking about the intensity of use of the medication (usually on a percentage scale over a defined period of time). We use persistence to describe duration of therapy, measured in days or months.
DR FOODY: Dr Benner, given your experience with adherence, what are some of the issues that have come to bear with regard to those patients or characteristics that impact adherence or nonadherence?
DR BENNER: Much of the research in this area has been devoted to answering 2 driving questions: How big is the problem? and Who is at greatest risk of becoming nonadherent? With respect to the first question, we've found that across populations, in studies using large databases of patients, about 1 in 3 patients is still highly adherent to lipid‐lowering or antihypertensive medication 12 months into therapy. After that 12‐month period, patients who are still adherent are actually much less likely to discontinue.
On the second question, 5 types of characteristics can help predict a given patient's adherence.
First, previously nonadherent patients (to any medication) are likely to be nonadherent in the future. Second, there is an observable pattern of nonadherence in patients who are at lower absolute risk of cardiovascular events. This would include patients who are younger than 55 years, people with no prior history of angina, coronary heart disease (CHD), or stroke, as well as patients who have fewer comorbid risk factors such as hypertension, dyslipidemia, diabetes, or congestive heart failure.
Third, we find adherence to be low in patients with evidence of competing health problems that are of immediate concern. These might include mental health conditions such as depression, anxiety, and dementia, or the presence of multiple comorbid conditions and medications. In some cases, these might be patients who are facing end‐of‐life issues or acute medical conditions that take their attention away from preventive medications such as antihypertensives and lipid‐lowering therapy.
Demographic characteristics can also predict nonadherence. For example, patients who are of nonwhite race tend to be about 60% more likely to be nonadherent than those who are white. There is a nonlinear association with age; that is, patients younger than 55 and those older than 75 are more likely to be nonadherent than those patients between these age cutoffs. And finally, although the evidence is mixed, I believe there is sufficient evidence to conclude that lower socioeconomic status is associated with lower adherence.
Finally, there is evidence that characteristics of the medication regimen itself can predict nonadherence. Adherence is lower with medications that must be taken multiple times per day, with medications that cause burdensome side effects, and with regimens that exacerbate “pill burden.” This is a term that applies to the total number of medications (and thus the total monthly expense) a patient faces. For example, in a recent study, we found that patients taking no other medications when they started anti‐hypertensive and lipid‐lowering medications were about twice as likely to be adherent compared with patients taking 6 or more other medications.
In that same study, we also found that patients who start antihypertensive and lipid‐lowering treatments at the same time are about 34% more likely to be adherent compared with those who start them between 60 and 90 days apart. So there seems to be a benefit to synchronizing the initiation of therapy in patients who need both of those risk factors modified.
DR FOODY: Dr Benner, thank you. Just in follow up, you've talked a lot about those things that cause nonadherence. Could you emphasize those characteristics or predictors of improved adherence or persistence in patients?
DR BENNER: The literature contains a number of positive examples of adherence‐enhancing interventions, and some of these can be easily implemented by a physician at the point of care. Patient education programs and close follow‐up from a pharmacist, nurse, physician, or dietitian have all been shown to improve adherence. It really doesn't seem to matter which provider delivers the message, as long as it's a trusted and influential health professional. Frequency and intensity are the keys here, so the more time spent in person with those providers, the better. Mailed refill reminders have been shown to have modest impact. Regimen selection has been shown to be important, so reducing daily dosing frequency, picking medications with fewer adverse events, lessening the out‐of‐pocket cost, using unit‐dose packaging for ease of use and adherence, and even using fixed‐dose combinations have all been associated with better adherence.
Our observational research suggests that one thing every physician can do is to follow patients more closely with office visits and laboratory tests. We've found that patients who receive follow‐up visits and laboratory tests within 3 months after starting lipid‐lowering therapy have considerably higher likelihood of being adherent over the following 3 years, even after controlling for their adherence in the first 3 months. In patients who were re‐tested, those who saw the most significant reductions in low‐density lipoprotein (LDL) cholesterol had the highest probability of being adherent over the following 3 years.
DR FOODY: Thank you, Dr Benner. Dr Frishman, based on your experience, how do these research findings jibe with what you've seen in your practice or research?
DR FRISHMAN: This is a big issue in clinical practice, because often you are dealing with patients with both high blood pressure and hyperlipidemia. You are dealing with patients with essentially an asymptomatic problem that is potentially life‐threatening, especially if the conditions occur together. The issue is, how does one get a patient who feels relatively well to comply with treatment? Patients often look well and don't show signs, at least externally, of active disease, so getting an individual to adhere to a regimen for years when he or she feels good is quite a challenge. I think that some of the issues Dr Benner highlighted—the strategies, the relationship you have with the patient, how much time you spend with them, the positive reinforcements that come from calling a patient on the telephone and telling him or her that blood lipids are normal or certainly at a more favorable level, or blood pressure is down—all provide positive feedback to keep people adherent. But this is a very big challenge in practice.
One of the more interesting observations Dr Benner mentioned is that when patients have both hyperlipidemia and hypertension, which is quite common in middle‐aged and certainly older Americans, starting an individual on 1 drug and then the other will often lead to poor adherence with at least 1 of the medications. One strategy that has been developed to try to deal with this issue of taking multiple medications is the concept of the polypill, which is essentially 1 pill that may treat multiple conditions. This is especially important in cardiovascular care, in which patients are often on multiple drugs, especially if they have coronary disease. Multiple drug regimens in cardiac patients include aspirin, angiotensin‐converting enzyme (ACE) inhibitors, additional treatments for high blood pressure, β‐blockers, and lipid‐lowering therapies.
So the polypill has been developed. We have 1 available here in the United States, based on the principle of treating 2 conditions simultaneously with 1 pill. Two well‐known agents are in this pill: amlodipine, which has been used to treat hypertension for years and is a fairly well tolerated antihypertensive medication, and the well‐known statin drug atorvastatin.
There is an argument for this approach beyond the issue of adherence. In the Anglo‐Scandinavian Cardiac Outcomes Trial (ASCOT), which looked at the utilization of antihypertensive therapy as well as lipid‐lowering therapy, independent of the effects on blood pressure, in a hypertensive population aggressive lipid lowering makes a difference to the point that we probably should consider the hypertensive population the same way we have come to consider groups with diabetes, coronary disease, or renal disease, in whom we recommend aggressive lipid lowering. In fact, ASCOT researchers found that getting the LDL below 100 was associated with appreciable benefits on subsequent cardiovascular risk.
So, just identifying hypertensive patients, most of whom will have LDL cholesterol levels above 100 mg/dL, would provide an ideal opportunity to use an agent to treat not only the high blood pressure but also hyperlipidemia at the same time. This would serve to improve adherence for both conditions by using a once‐a‐day pill to treat 2 commonly associated conditions seen in both cardiovascular and general internal medicine practices.
DR FOODY: Dr Frishman, thank you. While it seems very intuitive to use these combination agents—whether it's combining an antihypertensive with a lipid‐lowering agent, multiple antihypertensive therapies, or multiple diabetic therapies—there are significant resistance and clinical inertia on the part of the medical community to use these agents. What advice can you give physicians about concerns regarding combination pills?
DR FRISHMAN: Most clinicians have been taught to avoid combination medications. We have been taught to use individual drugs and up‐titrate them and only add a second drug if one does not reach the achievable goal. The problem is that our patients are now on multiple drug regimens, often with multiple co‐pays, and these patients are looking for a less complex approach. As mentioned, the more complicated the drug regimen, the less likely patients will adhere to the entire drug regimen. We must educate clinicians about the advantages of combination treatment for high blood pressure, utilizing multiple drugs working by different mechanisms to reduce blood pressure as well as 1 formulation to treat multiple conditions. The amlodipine/atorvastatin combination pill is the prototype of the latter.
In order to improve adherence, we must provide education and highlight findings such as those provided by Dr Benner, emphasizing if you prescribe treatment for hypertension and then for hyperlipidemia afterwards, you are going to have poor adherence to at least 1 part of that regimen. Combination therapy is essential, especially now with data showing that patients with high blood pressure should also be targeted with aggressive lipid lowering. It is not part of the formal guidelines yet, but I think it will be in due time.
DR FOODY: Excellent. Dr Frishman, you brought up the point about co‐pays and, Dr Benner, I know you've done a fair amount of research with regard to cost pressures and how they may impact adherence. Could you reiterate some of that for our readers?
DR BENNER: From an economic perspective, there appears to be some elasticity of demand for these cardio‐protective medications, and that elasticity seems to be higher than for medications for our children or a medication for an acute medical condition. The research on statins and antihypertensives suggests that as co‐pay burden goes up, adherence goes down; that's been confirmed with a number of studies over the past several years. However, other studies show that when the medication is completely free, adherence is also lower than if there is some moderate cost associated with it. So I would infer from this research that there's something to be said for asking the patient to buy in, so to speak, at least a moderate amount to optimize adherence with treatment.
DR FOODY: Thank you. Dr Frishman, as we look at these issues around hypertension, have you really moved to a different paradigm looking at global risk now in your hypertensive patients? How do you address typical hypertensive patients?
DR FRISHMAN: Well, clearly, hypertension can't be looked at alone. Very often it's associated with other conditions, ie, obesity, diabetes, and hyperlipidemia. The presence of hypertension actually identifies an individual at risk, but in addition one really has to look at the other associated risk factors to try to maximize benefit. Just treating high blood pressure and letting someone continue to smoke or not emphasizing the importance of smoking cessation is really not accomplishing very much. So we have to look at the whole individual and deal with the multiple problems that a hypertensive patient often manifests.
It's unusual in practice to find a “pure” hypertensive patient who doesn't have associated risk factors, some problem with weight, blood sugar, or lipids. I think if one realizes this quickly and then initiates therapy in a global way rather than just for 1 risk factor at a time, one would also have a greater chance of achieving long‐term adherence in patients, as Dr Benner showed from his research.
DR FOODY: Dr Benner, can you comment on this from a research perspective?
DR BENNER: What Dr Frishman is describing from a clinical perspective is also well supported by the research. In fact, that is why I am such a big fan of the concept of “global cardiovascular risk.” If you go back and ask patients who have discontinued antihypertensive and lipid‐lowering medications why they discontinued, they most often report quitting these medications because they are not convinced that they need the medication.
The problem is that we are asking patients to take a medication that lowers risk of an event that might happen or might not happen at some future point in time. Global risk helps patients understand how much benefit they are going to get from treatment and what might happen if they don't take the treatment. It also illustrates how multiple modifiable risk factors contribute to the overall risk.
I believe that the construct of global cardiovascular risk, if we were to practice it more commonly, would address the greatest barrier to adherence with lipid‐lowering and antihypertensive medications by giving patients an appreciation for the synergy of risk factors and the importance of treating risk globally.
DR FRISHMAN: I concur. If you look at patients who have multiple risk factors and address them all together, you can show that if all of them are treated successfully, you can reduce cardiovascular risk not just 20%, which you might see with just 1 risk factor manipulation, but 40% or 50%, by taking the global risk approach. I think you will get a much greater buy‐in to treatment right from the onset rather than by just addressing each risk factor separately.
DR FOODY: Given these important comments around the synergy of risk and this concept of global risk, I think we often practice medicine in silos such that our hypertensive patients are identified as just that, a hypertensive patient, and we secondarily may or may not address their other risk factors. But as both of you have highlighted, these patients are at tremendously high cardiovascular risk. This concept of global risk and the synergy of multiple risk factors are really essential as we look to optimizing outcomes in these patients. Given the burden of risk in our society, we have a strong case for the polypill concept, and certainly we have the prototype in amlodipine/atorvastatin. Overall, we need to address these patients globally and think beyond single risk factors and single treatment approaches. We have a large body of evidence supporting this approach in clinical practice.
Dr Frishman, any comments in closing?
DR FRISHMAN: Let me highlight that 1 of 4 Americans is hypertensive. We clearly have to pay attention to their associated risk factors. As ASCOT showed, treating hypertensive patients with lipid‐lowering medications even when LDL levels are “normal” improves outcomes. We must view high blood pressure as a target condition for aggressive lipid lowering, as we already do with coronary artery disease, diabetes, and renal disease. In the future, we are also looking for a polypill that might address even more cardiovascular risk factors.
DR FOODY: Dr Frishman, given your forward‐looking stand, what do you foresee with the next generation of guidelines from the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC)?How much of this do you think will be integrated into these newer guidelines for hypertension, lipids, or cardiovascular risk?
DR FRISHMAN: I think what really should be done is just a global risk recommendation rather than looking at risk factors as you mentioned, in silos, in which each condition looks like it's more important than the other conditions. Rather than having the JNC 8, I would like to see a global risk statement that would address all the risk factors for cardiovascular disease in 1 major recommendation. In my opinion, some of the problems have been caused by the fact that we practice in silos rather than looking at the patient overall. So I would welcome a clinical recommendation with a global viewpoint rather than looking at individual risk factors.
DR FOODY: Dr Benner, any other comments before we conclude?
DR BENNER: I concur with Dr Frishman's comment completely. I think the sooner we help patients understand that treatment aims to minimize a quantifiable risk rather than trying to manage 3 or 4 seemingly unrelated numbers, the better adherence will be, and the better off our patients will be.
DR FOODY: In conclusion, we need to think about hypertension acting synergistically with multiple risk factors, whether it be lipid lowering, the obesity epidemic that we are currently dealing with, or the lack of physical activity in our society. We know hypertension clusters with multiple other cardiovascular risk factors; unfortunately, in the current population, these risk factors are highly prevalent.
We also know that by lowering lipid levels and blood pressure, we can dramatically improve outcomes, at least in clinical trials. But the issue has been how to translate clinical trial results into day‐to‐day practice, as all these drugs are underutilized in clinical practice. Even when we prescribe medications appropriately and consistent with evidence based guidelines, patients do not adhere to these therapies. Reported discontinuation rates are quite high—upwards of 60% to 70% in multiple studies of patients on antihypertensives and lipid‐lowering agents.
So clearly there is an urgent need to develop and apply strategies to improve adherence as rates of cardiovascular death and morbidity continue to rise. We have highlighted some important approaches in this discussion. First and foremost, we need to change the paradigm with which we approach our hypertensive patients. We need to think about global risk in these patients and address that consistently. Among the strategies is the polypill, which addresses multiple risk factors in a single pill. Research has shown that the polypill amlodipine/atorvastatin improves adherence and that simplifying drug regimens and reducing pill counts can do that as well. We must address nonadherence regularly in clinical practice and look for ways to improve adherence. Given the burden of nonadherence in our society, we really need to incorporate these strategies into our clinical practice.