How We Got Here: The Failure Of Patient Compliance
The currently used Patient Compliance paradigm, as I’ve subtly suggested in previous posts,1 has not yielded a positive return on the time, money, and effort spent in its pursuit.
The consensus of experience and research in and experience with these efforts to improve adherence to treatment can be succinctly and accurately characterized by these three points:
- Many of the currently used individual compliance-enhancement interventions benefit some of the patients under some conditions some of the time.
- None of the currently used individual compliance-enhancement interventions benefit most patients in most conditions.
- Reliably predicting if a specific compliance-enhancement intervention will benefit a specific patient in a specific situation is an unachieved goal.
From these principles, one can infer that offering several different interventions at once might well be beneficial. And, in fact, there is evidence that a package of several, multi-layered compliance-enhancement interventions can be effective across the patient population. As one would suspect, however, such shotgun approaches are neither efficient or economic. This extract from a Cochrane Review of medication compliance enhancements elegantly summarizes the current state of affairs:
Several complex strategies, including combinations of more thorough patient instructions and counseling, reminders, close follow-up, supervised self-monitoring, and rewards for success can improve adherence and treatment outcomes. However, these complex strategies for improving adherence with long-term medication prescriptions are not very effective despite the amount of effort and resources they consume.2
My conviction is that the failure to develop pragmatic solutions for noncompliance is due to the concept of medical noncompliance itself, manifested in a widespread misunderstanding of that concept, a denial of its complexity, and an obsessive drive to find a panacea for a problem that is more ideological than pathological.
Medical Noncompliance has ultimately become no more than a label we’ve assigned to an extraordinarily wide and varied category of non-optimal behaviors patients exhibit in response to their clinicians’ treatment recommendations. Instances of noncompliance, as already noted, can appear in an infinite variety of forms and be caused by an infinite number of causes. And to flip the late Justice Potter Stewart’s famous statement about pornography,
We can define medical noncompliance
but we often don’t know it when we see it
Why Now: Money – And Saving A Life Or Two
Remarkably, there has been so much emphasis placed on the potential cost-saving to be wrung from increased treatment adherence rates that it seems incumbent to note that the primary goal in improving the effectiveness of medical treatments is to maintain and improve the health of individuals, thus reducing suffering and the loss of life.
That improving the effectiveness of medical treatment by increasing treatment adherence rates will decrease primary healthcare expenses and secondary costs due (e.g., losses due to absenteeism from work) is a pleasant bonus (OK, make that an incredibly wonderful bonus).
That point duly documented, the economic cataclysm created by spiraling healthcare costs during a period of general economic chaos is the tipping point precipitating long overdue changes in how healthcare is managed, politically, professionally, and administratively. Healthcare expenses in some industries are the most expensive budget category other than salaries.3 A 2004 McKinsey Quarterly presciently opened an article on controlling healthcare benefits with the observation that “In a few years, the average Fortune 500 company may be spending as much on health benefits as it earns in profits.”4 Further, responsibility for healthcare costs have arguably become the primary conflict between management and labor. In these circumstances, the inability to efficiently improve medical compliance rates, the most significant cause of unnecessasry healthcare costs has escalated from painful and frustrating to catastrophic.
The healthcare community has a responsibility to find a solution. We should start by looking for that solution in the most likely places.
A passerby walking home late at night sees a drunk on his knees under a streetlight, searching for something. The passerby asks the drunk, “What are you doing?” “Looking for my apartment key,” says the drunk. The passerby, trying to be helpful, walks to the area near the drunk and begins looking for the key. After 15 minutes, it becomes clear that the key is not in the vicinity. The newcomer asks, “Are you sure you lost your key here?” “No, actually I first noticed it was missing when I was walking over there,” the drunk says, pointing to an area a half block away. The passerby, a bit perplexed, asks, “Then, why are you looking here if you lost your key over there?” Responds the drunk, “Because the light is better here, under the streetlight.”
I’ve used that joke before,5 but it fits too well here to pass it up. The point, of course, is that we can’t afford to keep looking for solutions in areas that have repeatedly been shown to be bereft of answers even if those areas are more familiar.
What Has To Change: New Solutions To Noncompliance
One might hypothesize6 that the current system has been perpetuated simply because no alternatives have been found.
In this instance, one would, however, be wrong.
Heck, not only are there promising alternatives but those alternatives are well known and in place – in other fields
Compliance and noncompliance are issues in many areas of human endeavor other than healthcare. The legal system is, in fact, built around the notion of citizens complying with laws. Politicians work to bring us into compliance with their ideas. Adherence is an essential element in educational institutions, military units, and professional communities. And, of course, marketing is all about getting potential buyers to comply with recommendations to buy a particular product or set of goods.
Here’s the key:
Noncompliance is not exclusively a medical issue.
It’s a personality trait that manifests in many areas of life.
Consequently, a reasonable source of potential tactics for managing noncompliance is that set of strategies that have been successful in these other fields.
For today, I’ll ask the reader to be content with one example with the promise (i.e., threat) of several other possibilities to follow in the future.
What Next: Population Segmentation – One Promising Option To Improve Treatment Effectiveness
In keeping with the principles of full disclosure, I want to alert readers to the fact that I am involved with EnrichMap, which offers a system of interventions based on population segmentation (also called patient profiling).
Marketing research uses population segmentation to identify the likely users of a product or service and the specific appeals to which those users are most likely to respond. Within that definition, however, are several varieties of segmentation. Large groups may be broken down into subgroups based on age, ethnic background, economic status, cultural mores, psychological traits, … .
Ironically, Big Pharma is one of many industry sectors that use population segmentation extensively. Most commonly, however, pharmaceutical manufacturers use patient segmentation only for determining the characteristics of the patient group whose needs best match a given product.
Today, sophisticated statistical techniques are used to divide the pharmaceutical marketplace into distinct segments of patients, based on their clinical profiles and the type of medical care they receive. When done well, patient segmentation reveals the size of the market, which segments hold the most value for a brand, and which product attributes will appeal to each patient sector.7
And, population segmentation is used extensively in health care planning, i.e., anticipating future needs, allotting resources, and constructing delivery systems.
To create a valid, reliable, and useful segmentation is not a trivial task; the specific methodology required is, thankfully, beyond the scope of this post. On the other hand, the potential utility of such groupings is intuitively apparent.
If, for example, market segmentation shows that 85-90% of your product, a digital caliper with built-in GPS, is purchased by 45-55 year old Asian women with postgraduate engineering degrees, you can probably turn down that commercial time on SpongeBob SquarePants the ad company is pushing you to buy, regardless of how low the ad rates or how huge the show’s audience (unless there is convincing evidence that the program’s viewership skews to 45-55 year old Asian women with postgraduate engineering degrees).
Similarly, if population segmentation shows that 94% of a working mothers of school-age children strongly believe keeping their kids’ immunizations up to date is essential but find it difficult to keep track of current recommendations, document which immunizations their children have already had, and arrange the logistics of getting their offspring immunized, public health officials might consider (1) decreasing the monies spent on educating mothers about the importance of immunizations (2) developing an automated means of alerting parents to changes in immunization recommendations, and (3) constructing a system, perhaps in cooperation with the schools or other community institutions, to conveniently provide and document the immunizations.
And, on a doctor-patient level, if a physician knows that a patient is part of a population segment that typically becomes ill at ease, uncertain, and confused when dealing with healthcare decisions or treatment instructions, the doctor might decide to institute special measures for that patient that might not be used for others with the same diagnosis. After the explanation of a treatment plan, for example, the patient might be asked to describe the recommendations in his own words or to complete a simple test about the treatment to check his understanding of the plan. The clinician might also provide that patient with auxiliary calendars indicating dosing times and dates, arrange for the patient to enroll in a system of automated reminders or, if the risk from the disorder is high, use an automated medication dispenser that alerts care-givers and the clinical staff if the medication is not dispensed, schedule more frequent follow-up appointments, and request that staff call the patient the next day to assure that the treatment is being correctly implemented.
How Segmentation Differentiates Compliance Interventions
Patient Segmentation matches the specific patient with the interventions most likely to succeed in improving treatment implementation rather than searching for one panacea for all noncompliance. While one might argue that all the measures listed in the case described in the preceding paragraph could be beneficial to every patient, patient segmentation shows that a significant percentage of patients will implement treatment correctly without these aids, another significant percentage of patients will be unlikely to correctly follow any treatment whether or not they are provided these aids, and still another significant percentage of patients will demonstrate a meager improvement with the described assistance but will show a more robust response to other interventions.
Further, segmentation techniques can identify patients that are comparatively more responsive to incentives, personal interventions by the physician, comprehensive explanations, peer support groups, and a myriad of other interventions now applied to a population on an all or nothing basis.
Patient Segmentation focuses on the realistic goal of managing noncompliance rather than the incredibly difficult if possible at all goal of curing noncompliance.
And, given that (1) improving communications between physician and patient can improve patient behaviors (including implementation of prescribed treatment) and (2) communications with patients can be managed more efficaciously than attempting to directly manage behaviors of patients, reaching the ultimate goal – improved treatment effectiveness – is more likely to be accomplished by shifting from the current efforts to coerce higher compliance rates to using Patient Segmentation to enhance the doctor-patient communication systems.
What’s The Catch?
Patient Segmentation sounds pretty good, doesn’t it?
So, why isn’t this kind of tool in use everywhere now?8
The reflexive response is that there is no definitive proof that interventions based on segmentation effectively improves the successful implementation of treatment – which only transforms the question to why a patient segmentation approach to treatment adherence isn’t being vigorously explored.
Why are clinical organizations, academics, the government, pharmaceutical companies, and other healthcare stakeholders far more willing to fund studies of traditional compliance tactics which have been repeatedly shown to be unsuccessful than segmentation strategies, which have been demonstrated effective in other fields?
It’s a puzzlement.
Future posts will address other alternatives (i.e., in addition to population segmentation) to the conventional patient compliance model.__________
- See, for example, The Tragedy Of Patient Compliance, Patient Compliance – So Wrong For So Long, Why Today’s Treatment Adherence Paradigm Must Be Destroyed – Part 1, Why Today’s Treatment Adherence Paradigm Must Be Destroyed – Part 2, Ending The Damage Caused By The Treatment Adherence Paradigm.↩
- Haynes RB, McDonald H, Garg AX, Montague P. Interventions for helping patients to follow prescriptions for medications. (Reprint of a Cochrane review 2003, Issue 4) Update Software Ltd, Summertown Pavilion, Middle Way, Oxford OX2 7LG, UK. 9↩
- Sick and Getting Sicker, By Simona Covel. WSJ. July 13, 2009.↩
- How To Control Health Benefit Costs, Lynn Dorsey Bleil, James Kalamas, and Rayman K. Mathoda. McKinsey Quarterly. February 2004.↩
- See Why Today’s Treatment Adherence Paradigm Must Be Destroyed – Part 2↩
- I’m assuming this is the same “one” who, in the first section of this post, correctly inferred that offering several different compliance-enhancement interventions at once might well be beneficial and now feeling on a roll, has thus been rendered a tad overconfident and perhaps even foolhardy when it comes to reaching suppositions.↩
- A New Measure for Segmenting Patients by Their Total Health Care Experience By John Iacoviello, PhD, and Jim Carroll. Product Management Today. Vol 18, No 3. March 2007↩
- To be fair, segmentation is used here and there.↩