One of the most interesting aspects of working as a community-based pharmacist is the insight you gain into the actual effectiveness of the different health interventions. You can see the most elaborate medication regimens developed, and then see what happens when the “rubber really hits the road”: when patients are expected to manage their own treatment plan. Not only do we get feedback from patients, there’s a semi-objective measure we can use — the prescription refill history.
The clinical trial, from where we derive much of our evidence on treatments, is very much an idealized environment. The relationship to the “real world” may be tenuous. Patients in trials are usually highly selected, typically those that are able to comply with the intervention planned. They may need to be free of any other diseases which could complicate evaluation. Patients that qualify for enrollment enter an environment where active monitoring is the norm, and may be far more intense than normal clinical practice. All of these factors mean that trial results may be meaningful, but not completely generalizable to the patient that may eventually be given the intervention. It’s for this reason we use the term “efficacy” to describe clinical trial results, while “effectiveness” is what we’re more interested in: those real-word effects that are far more relevant, yet more elusive to our decision-making. Efficacy measures a drug’s effect on an endpoint, and estimates risk and benefit in a particular setting. Effectiveness adds in real-world tolerance, the ability to tolerate the regimen, and all the other factors that are present when real patients take a drugs under less-than-ideal conditions. Consequently, effectiveness is a much more useful predictor of outcome than efficacy. Unfortunately, measurements of real-world effectiveness, possibly as a “phase 4″ or real-world trial, are rarely conducted.
But the gap between clinical trials and the real world is obvious: medications just aren’t taken regularly. The extent of non-compliance with prescribed therapy varies between reviews, but an estimate of about 25% is probably fair. Patient with HIV, arthritis, gastrointestinal disorders and cancer tend to be the most adherent to therapy — not surprising, perhaps because of the seriousness of the condition, or perhaps the responsiveness of symptoms to treatment. Least compliant are patients for treatments of lung disease, diabetes, and sleep disorders. In general, it’s safe to assume that if symptoms are mild or non-existent (e.g., high blood pressure or high cholesterol), compliance will be poor, compared to more symptomatic diseases.
Failing to adhere to a prescribed medication course is a significant challenge — after all, you won’t get the outcome you want if the medication isn’t taken. Or will you? The relationship between the two that has always been considered to be causal, but there is some evidence to suggest that adherence to a medication schedule may be a proxy for healthy lifestyles and generally healthy behaviours. I wanted to look closer into adherence after reading a paper highlighted by Mark Crislip a few weeks ago: Adherence to placebo and mortality in the Beta Blocker Evaluation of Survival Trial (BEST). This post-hoc analysis studied the placebo group of a double-blind, placebo-control trial that evaluated the drug bucindolol versus placebo. Those deemed “adherent” to their placebo (taking >75% of doses) were compared to the non-compliant (who took <75% of doses). The authors adjusted for all known modifiable, non-modifiable and psychosocial variables. The surprising result?
Adherent participants had a significantly lower total mortality compared to less-adherent participants (HR = 0.61, 95% Confidence Interval: 0.46–0.82). Adjusting for available confounders did not change the magnitude or significance of the estimates. When considering cause-specific mortality, CVD and pump failure showed similar associations.
Those are impressive gains, apparently from a placebo. So what’s the cause? It’s not any confounding factor already explored by the authors. And we know the placebo has no meaningful physiologic effects. So by increasing adherence to prescribed medications, are we getting therapeutic effects from the intervention, or simply by modifying some other factor that affects outcomes? Or are there other factors at play? In the absence of a prospective randomized trial, we can’t say with any certainty.
While the correlation may not be understood yet, it seems reasonable to continue to promote adherence, given the relationship with improved outcomes. Surprisingly, despite the widespread nature of the problem, the literature base on interventions to improve adherence don’t paint a clear picture on how to change the situation. Well designed trials that evaluate interventions are rare. Probably the most comprehensive summary of the literature is from the Cochrane review, which concludes that even more effective interventions don’t lead to significant improvements in clinical outcomes. Before we get into what works, we should be clear on the terminology [PDF], as it can be source of confusion:
Adherence or Compliance refers to the act of following recommendations made by health professionals, with respect to drug, dose, and medication schedule. The ideal means of measuring adherence is with electronic devices that measure daily dosing. As these are usually not available, the prescription refill frequency is a useful, if imperfect, proxy.
Concordance is a related term used to describe a shared agreement between a health professional and a patient about therapeutic goals. It’s less a measure, and more a philosophical approach to implementing treatment plans.
Persistence refers to the duration of conformance to a particular treatment plan, and is usually defined by the interval between when therapy is started, and when it is discontinued. Adherence to the exact dosing schedule is not necessary in a measure of adherence.
Most research on adherence evaluates the effect of a specific intervention on some measure of medication-taking. In a few studies there are also evaluations of clinical outcomes that are a consequence of (non)adherence — which is far more relevant, but unfortunately, less frequently evaluated:
Technical interventions simplify dosing by modifying how medications are actually taken. You could reduce the number of daily doses (e.g., the extended-release tablet) or give a tablet or dosage form which contains two or more different drugs (e.g., many HIV medications). Technical interventions also include pharmacy-driven services like bubble-packing tablets, or by filling specialized pill “medication reminder” devices.
Technical interventions have been evaluated in multiple reviews. Obviously, blinding is a problem in these circumstances. In general, simplifying medication regimens helps with adherence: Taking one pill per day is easier and more acceptable than one pill every six hours.
Behavioural interventions prompt medication use through the use of memory aids, reminders (manual or electronic), monitoring, feedback and rewards. These types of interventions seem to have beneficial effects, an observation that’s been made in several different studies. There’s a lot of potential to use mobile technology to support medication use, now that smart phones are becoming ubiquitous.
Education interventions are the pharmacist’s bread and butter: patient teaching and knowledge. In general, there’s a positive correlation between patient knowledge, medication adherence, and outcomes, although the effects don’t seem as impressive as technical and behavioural interventions.
Social support interventions offer practical or emotional support to patients. While it sounds attractive, there’s less data available to describe which interventions are the most effective.
Structural interventions are more extensive tactics usually used to manage chronic disease. Workplace specialty care programs (e.g., hypertension) or disease management programs (e.g., diabetes) are examples of this intervention.
There are several caveats to keep in mind. Studies that have prospectively looked at adherence typically enroll patients that again, may not be fully representative of the general medication-taking population. In addition, studies that have looked at adherence suggest that many interventions that work can be time consuming and expensive. Finally, the relevance to long-term medication use is unclear in some cases.
Adherence to therapy is a medical challenge, where the root causes are not always well understood. While the idea of a healthy patient effect seems real enough, the factors that lead to better outcomes in those that adhere to therapy isn’t clear. Given the importance of adherence to the treatment of most chronic diseases, it’s a unfortunate that strategies to improve adherence have not been more thoroughly evaluated. In the absence of good data, interventions to simplify medication schedules, provide education on the importance of adherence, and provide reminders on dosing make sense. When the “rubber hits the road” when planning treatments, it’s important to recognize that adherence may be one of the biggest influences on health outcomes.