Engaging Providers in Medication Adherence: A Health Plan Case Study

Drugs don’t work in patients who don’t take them.”¹—C. Everett Koop, MD, ScD, former US Surgeon General

Prescription drug therapy is the mainstay of treatment for many chronic conditions, such as hypertension, hyperlipidemia, diabetes, asthma, and attention-deficit/hyperactivity disorder (ADHD). It is well known that adherence to a prescribed drug regimen is crucial for positive health outcomes. Medication adherence—the extent to which patients take medication as prescribed by their healthcare provider—requires a shared responsibility between patient and physician.¹ Nonadherence is sometimes referred to as America’s “other drug problem.”² This widespread problem is often overlooked; nonadherence is common, costly, and complex. One of every 4 medical visits results in patients not following the advice they were given.³ Each year, nonadherence contributes between 33% and 69% of medication related hospital admissions.1 Studies show that within 1 year of being prescribed a medication, patients stop taking 50% of their long-term medications, and up to 1 of 5 new prescriptions is never filled.^4,5 In addition, nonadherence is estimated to contribute to 125,000 deaths annually in the United States, and nonadherence costs an estimated $100 billion annually in direct and indirect healthcare costs.^4,5

Based on the available literature, adherence depends to a large extent on the patient’s perceived health benefit; however, the healthcare provider must work with the patient and/or caregiver to educate, cooperate with, and help the patient to focus on such benefits.

Barriers to Medication Adherence
Many patients engage (intentionally or not) in negative medication-taking behaviors—resulting from forgetfulness, cost, inconvenience, and attitudes—that may result in negative health outcomes. These behaviors include pill splitting, drug holidays, taking medications every other day, or never filling a prescription, all of which adversely affect health outcomes.⁶

An important barrier to adherence is low health literacy. Health literacy can be defined as the ability to read, understand, and make appropriate health decisions.⁷ Poor or low literacy level results in a patient’s inability to follow treatment recommendations and a reduced ability to navigate through the prescribed treatment plan (in this case, following prescribed medication regimens).⁷

Contributing to nonadherence is the high cost of prescription medications and out-of-pocket (OOP) expenses. It is known that 1 of 5 patients chooses not to fill a prescription because of cost.⁸ Financial barriers cannot be overlooked; therefore, influencing patient behavior with monetary incentives or other rewards for appropriate adherence to medication regimens may be useful.⁹ However, essential to improving patient compliance and adherence—to ultimately improve health outcome—is good communication between the patient and the entire healthcare team.¹

Previous studies have shown that by engaging the patient and forming collaborative relationships, healthcare professionals can significantly influence adherence.^2,10,11

In their recent commentary, Cutler and Everett suggest that with an investment of time and resources, healthcare providers can optimize and reconcile patient medication regimens in an attempt to manage nonadherence directly.⁹ Cutler and Everett further note that using proved adherence assessment tools at the time of the visit can predict risk for nonadherence. Having this information can enable providers to encourage adherence at the point of prescribing and during any followup contact with patients.⁹

Participating in any negative prescription-taking behavior can and does affect all patient types, yet noncompliance remains largely unrecognized in clinical practice.⁸

Highmark’s Project Design
To address this problem, in the fall of 2008, Highmark, Inc, a large regional managed care health plan located in Pittsburgh, PA, with more than 4.5 million health insurance members and approximately 2.5 million prescription drug benefit members, embarked on a project to increase the awareness of its providers to nonadherence. The authors are Highmark employees who concentrate on building relationships with network providers to improve the quality of care for the health plan members.

A total of 5 distinct network primary care practices agreed to take part in this study. The goal was to illustrate to the physicians, by the use of prescription drug claims, the extent to which many of their patients were not following their medication regimens as prescribed. Through conversation and discussion with these network providers we found that physicians in general believe that nonadherence is a global issue but does not directly affect their own patient population. This begged the question as to who is accountable for the resolution of the problem. Is it the clinician, the health plan, or the patient? As previously suggested, nonadherence is a joint responsibility that needs to be addressed by all parties.¹

This current case study was conceived as a way to engage and challenge these Highmark network primary care physicians (PCPs) and staff to develop steps to improve medication adherence rates of plan members and ultimately of all their patients. Because each group differed in population and targeted disease state, the project developed according to the dynamics of the individual practice. To minimize the time a practice would need to commit to this project, patient samples were kept to a minimum. However, once the prescription drug data were shared with each group, their interest peaked.

This study was conducted from the fall of 2008 through the spring of 2009; it began with a medication gap analysis using prescription drug claims data from the previous 18 months. Once practice-specific barriers to adherence were identified, solutions were developed for the varied populations. Collaboration between the providers and the health plan consultants was key in these processes.

These claims data helped to identify potential nonadherent patients from the 5 primary care practices— 1 internal medicine, 1 family practice, and 3 pediatric groups. Each group was provided with a small sample of patients

Table 1

These groups were selected based on their positive relationship with the Highmark team and on their willingness to join in this process improvement initiative that could be submitted to QualityBLUE, the health plans’ physician pay-for-performance (P4P) incentive program.

Once potential nonadherent patients with hypertension, hyperlipidemia, or ADHD were identified, the following questions were considered:

1. Are these patients truly nonadherent as illustrated by the claims data?
2. What are their reasons for nonadherence (barriers to adherence)?
3. What can the practice do to resolve these?
4. Is quality of care being affected by nonadherence?

The providers had to research patient charts to verify the prescription drug information and answer these questions.

To determine the specific barriers to adherence, the clinicians were asked to reach out to their patients. Some clinicians chose to contact each patient by telephone to discuss the patient’s medication regimen; others opted to use a survey tool provided by the health plan for a more structured approach. Each practice was encouraged to create a plan of action to deal with the identified barriers.

Methodology
The first step involved data collection for each of the 5 group practices; prescription drug claims were pulled for all of 2007 and for the first 6 months of 2008 (ie, 18 months). Data were extracted using membership of each practice, followed by therapeutic drug categories for hypertension, hyperlipidemia, and ADHD. This ensured that prescriptions ordered by a specialist other than the PCP would be included. Next, the project managers analyzed the information, identifying each member, and then looked for gaps or lack of claims that seemed excessive. We designated <80% adherence as the determining factor.

Sample spreadsheets were developed as a visual tool for each practice to illustrate the gaps in therapy (Figure). Each practice had to determine whether the gaps in claims corresponded to true gaps in therapy or were a result of another reason that could not be determined by claims analysis, such as physician discontinuation of medication, a patient leaving the practice, or a change in insurance coverage.

Figure

After sharing the findings with the plan consultants, a chart review was prepared for each group, and the next steps and interventions suggested. This is where each group differed.

Results
Individual Practice Group
Group A: Family Practice. This group was a large, urban, multiprovider practice with 3 locations and fully automated electronic medical records (EMRs). The practice administrator reviewed each of the 51 electronic charts of patients identified as nonadherent and was skeptical of these results. The administrator personally called every patient to discuss the importance of taking prescribed medications or contacted the pharmacy when the patient was unavailable. She found that 71% (n = 36) of these members were in fact getting their prescriptions filled. However, their claims were being paid for in cash and were not being submitted by the retail pharmacies, because they were part of a promotional generic drug program. This alerted the authors to a claims submission problem.

Indirectly, the lack of these claims also signified that the high cost of prescription drug therapy could be a barrier to adherence. However, these patients were satisfied with getting a generic product filled that reduced their OOP spending. In addition, of patients identified as nonadherent, 12% (n = 6) were getting sample medication from the PCP or the specialist; samples do not generate a claim and, therefore, appear as a gap in therapy. Patients could fall into more than 1 of these categories, because they were taking multiple medications. Other reasons for apparent gaps are listed in Table 2.

Table 2

As a result, group A was confident that a high percentage of the targeted patients were adherent to their medication regimens. The authors found that flaws, errors, and discrepancies exist in the claims data, but the drug data presented a starting point for identifying nonadherence.

Group B: Internal Medicine. This rural group consisted of 4 physicians, with a cardiology subspecialty, and had a partially automated EMR. The practice manager examined each of the 51 electronic charts identified. In some cases the physician discontinued the medication in question, and in others, a lack of claim was the use of promotional generic drug programs in the marketplace (in which case claims are not submitted under a prescription benefit). Other reasons included a patient transfer to a skilled nursing facility and medication samples being provided to defray high OOP costs (which might have related to a tiered drug benefit, a nonformulary agent, or a coinsurance design).

This information again alerted the Highmark team that errors in drug data could be misleading. Another important finding was that 33% (n = 17) of the planidentified patients stopped the medication on their own, because of high cost. In each case, the clinician suggested an alternative therapy or recommended a patientassistance program. Without these data, this practice would not have known that those patients had stopped taking their medications.

Overall, the clinicians’ awareness was heightened, and they realized that chart review alone was an insufficient source of information for a subset, approximately 30% (n = 14), of the 51 patients who had been initially identified as nonadherent. A more formal approach to identifying specific barriers to adherence was needed.

The Highmark team recommended the ASK-20SM Survey as a tool to identify and address barriers to adherence. This validated survey, originally developed by GlaxoSmithKline (and used with permission for this study), examines categories related to medication-taking behaviors. The categories used in this study and the results are listed in

Table 3 

Again, the number of patients in this group was small, but the findings were astounding. The lead physician realized that changes were needed to reduce nonadherence among their patient population. The next step for group B was to create a gap-free system through their e-prescribing and EMR.

This group is now developing an electronic alert that is built into the EMR system, which would be activated when the patient does not get a medication filled or refilled in a timely manner. This would enable the office staff to reach out to the patient and determine the reason for nonadherence.

Groups C, D, and E: Pediatrics. Three pediatric groups participated in the adherence project, using ADHD as the chronic condition studied. Pediatric group C was a large, rural, multiprovider site with multiple locations and a partially automated EMR system, which was closely aligned with a hospital system. Pediatric group D was a large, independent, suburban, multiprovider group with multiple locations yet without an EMR system. Pediatric group E was a large, independent, urban, multiprovider practice with multiple locations. In group E, a behavioral health physician specialist is employed in the practice, and it has a fully automated EMR system.

The findings for pediatric groups C and D were similar. After careful chart review identified nonadherent patients (Table 1), a few common barriers were identified

Table 4

Some of the plan-identified gaps were determined to be appropriate, such as a change of medication by a physician or discontinuation of therapy. The providers rationalized that because of the high sensitivity of this diagnosis, there was no easy method of intervention to improve adherence.

Pediatric group E was the most engaged participant. The behavioral health specialist was intrigued by the data, took ownership of the issue, and ultimately created an updated, modified management approach to ADHD for the group. Using claims data as a starting point, she reviewed each of the 32 electronic patient charts originally identified by the plan and found similar barriers to adherence compared with the other pediatric practices. However, 11 (33%) of the 32 patients identified were found to be of college age, which encompassed a unique set of barriers to adherence.

Table 5

Discussion
The main goal of this study was to increase the physicians’ awareness of medication nonadherence. Although it is well-documented in the medical and pharmacy literature that nonadherence is a significant problem, these physicians did not realize that this was a problem for some of their own patients. Overall, the physicians were surprised to learn that nonadherence existed among their patient population. Despite the small number of patients involved, the value of this study was the providers’ heightened awareness of nonadherence.

Each practice learned something about nonadherence, from simply gaining awareness of the problem to finding specific barriers to medication adherence. Some practices developed methods to improve adherence and worked toward implementing change. Practices with EMRs realized the advantage of technology in defining a problem as well as in the development of an improvement process. They learned that a gap analysis can be a valuable tool.

The providers realized that improved communication with their patients is necessary to determine adherence barriers and how to alleviate them; good communication with the patient and the entire healthcare team is essential for improving patient adherence.¹⁵ Literacy level, English-language proficiency, understanding of basic medical instruction, and socioeconomic status are factors that affect communication and adherence.⁷ Each factor needs to be taken into consideration when informing and educating patients on the importance of medication regimens.

Adherence issues cross over all populations of patients, regardless of age and/or condition,^1,2 and interventions will vary based on that population. It is crucial that programs targeted to improve adherence are customized in a manner meaningful to the specific population being addressed.

The information depicted in the Figure may be useful when talking to physicians about adherence among their own patient populations. Health plans may be able to create a similar visual tool from their claims data that could be shared with providers to engage them in medication adherence. An unexpected value of this sample tool (Figure) was that other Highmark employees may replicate it and share it with their physician groups.

It has been shown that the quality of the doctor– patient relationship is one of the most important factors affecting adherence.¹⁶ A supportive relationship that features encouragement, reinforcement, and motivation from the provider will improve adherence; conversely, poor provider communication can contribute to medication nonadherence.¹⁶

If left unattended, nonadherent patients might have experienced negative health outcomes, and consequently medical costs for the plan could have escalated. ⁵ In the scope of this case study, cross-referencing with medical claims was not completed. However, other health plans interested in a more quantitative approach may take the next step of cross-referencing medical and pharmacy claims to define the costs associated with nonadherence.

The clinicians did recognize the value of collaboration with the health plan consultants who could provide data as identifiers of a problem. Groups C and D were interested in reviewing similar claims data for other chronic conditions related to pediatrics, namely, asthma, but no action was taken by either group.

Unique Features of Group E
The common denominator among group E’s small subset population was that they were all college-aged students. This alerted the behavioral health specialist that college-aged patients have additional barriers to adherence that she and her colleagues were unaware of, and so this subset of patients was falling through the cracks. For each of these patients, the physician created a detailed care plan.

Just as important, the variation among the 12 pediatricians in group E—all of whom treated and managed patients with ADHD—was evident. The behavioral health physician realized yet another significant problem: ADHD was not being managed consistently among the providers in the group. Using this information, she created a quality improvement guide for ADHD management, which was to be used by all physicians in the practice, at all locations, when diagnosing patients with ADHD and interacting with their families.

This guide addressed all facets of ADHD management, including errors of monitoring and addressing compliance. The steps involved in addressing adherence in the practice include:

1. At every visit inquire about barriers to medication therapy
2. Emphasize benefits of daily medication adherence
3. At time of medication check, encourage scheduling of next appointment
4. Offer reminder options, such as an e-mail or a postcard
5. To address appointment no-show, patient or parent will be contacted by telephone to reschedule
   • If unable to reach after 2 attempts, a certified letter will be sent, requesting patient/parent call to schedule an appointment
6. The practice will develop a policy for college-aged patients taking ADHD medication
   • To monitor efficacy and side effects
   • To provide for timely refills to reduce nonadherence.

To improve adherence and the understanding of the value of daily medication regimens, future enhancements to the practice’s website are planned. Links to quality, reliable, evidence-based sites will be accessible; downloadable medication questionnaires will be made available for patients to fill out and bring to each appointment; and medication question-and-answer sessions, similar to a chat room, will be scheduled.

As a result of this project, group E’s physician awareness of nonadherence was increased. This pediatric practice transformed their approach to ADHD management. They implemented new practice guidelines, disseminated the information to the clinical staff, and used their EMRs for theirs and their patients’ advantage. With these improvements, communication between patient and provider is expected to be enhanced, reduce nonadherence, and improve patient outcomes.

Participating in this case study also enabled group E to document improvements regarding medication adherence to the health plan, which fulfilled a P4P requirement.

Collaboration and Care Coordination
We realized the value of physician buy-in and collaboration as critical elements to the success of this project. It became evident that the practice with a physician champion—one who was very engaged and passionate— made the greatest effort to develop process improvements. In addition, we learned that it is crucial to identify the responsible parties, develop a work plan with targeted dates, and use telephone and e-mail reminders as a way to guide and direct the participants, while being careful to appreciate time constraints common to busy primary care practices.

With current shifts in healthcare and payment methodology, adherence is being affected by a lack of coordination of care. This includes provider and patient communication, increased use of technology (including data-sharing among providers), and expanded P4P programs, which promote patient-centered medical homes and care transitions. Collectively, advancements in these areas will improve coordination of care and healthcare delivery, particularly in primary care, where the crisis of nonadherence is most evident.⁹

Limitations
Various limitations to this case study exist. We recognize that the limitations of prescription drug claim data (eg, nonsubmission of claims, errors of membership or eligibility, and lack of integration of medical and pharmacy claims) may have affected the results. However, the claims data were useful as a starting point.

The small sample size was another limitation of the project, but this was necessary to keep the workload manageable for the staff.

In addition, medical claims and pharmacy claims were both available but were not cross-referenced because of the limited scope of this study. This proved to be a limitation in that escalation of care costs was not truly determined in this case study; the care costs were simply assumed. Also, it was not validated that quality of care was being affected by the nonadherence.

Finally, a major limitation of this case study was the realization that this problem is widespread and there is no simple solution. Recognizing this, some physician groups chose to avoid the challenge; they were not pressured to assume the responsibility of improving nonadherence. This reinforces the question of who is accountable for the resolution of the problem—is it the clinician, the health plan, or the patient?

Conclusions
Ultimately, all involved parties learned that nonadherence is a reality and affects the patients they treat. The authors learned that using prescription drug data can provide valuable information, including the need to tailor any intervention to the individual patient’s characteristics (ie, age, disease state, socioeconomic status).

Three key potential benefits of improved adherence are a decrease in total healthcare costs, a decline in hospitalizations, and improved health outcomes. However, for these benefits to be realized, providers must become a part of the solution. By identifying barriers to adherence, steps to improve adherence rates can be put into place. These include collaboration, open communication between healthcare professionals and patients regarding medication regimens, as well as education provided to patients about the benefits of medication adherence.^12-14

Changing patient behavior is difficult. It involves more than simply telling patients what to do or prescribing a medication. Healthcare professionals need to identify barriers to adherence and then help motivate patients to take control of their disease. Much of this revolves around medication-taking behavior and education on the disease state. Increasing primary care physicians’ and staff awareness of medication nonadherence and working toward improving medication adherence rates may lead to a decrease in total healthcare costs, a decline in hospitalizations, and improved patients’ health outcomes.

Disclosure Statement
Ms Scott and Ms McClure have nothing to disclose.

References

1. Osterberg L, Blaschke T. Adherence to medication. N Engl J Med. 2005;353:487-497.
2. National Council on Patient Information and Education. Enhancing Prescription Medicine Adherence: a National Action Plan. August 2007. www.talkaboutrx.org/ documents/enhancing_prescription_medicine_adherence.pdf. Accessed July 14, 2009.
3. DiMatteo MR. Variations in patients’ adherence to medical recommendations: a quantitative review of 50 years of research. Med Care. 2004;42:200-209.
4. Peterson AM, Takiya L, Finley R. Meta-analysis of trials of interventions to improve medication adherence. Am J Health Syst Pharm. 2003;60:657-665.
5. The Task Force for Compliance. Noncompliance with medications: an economic tragedy with important implications for health care reform. Baltimore, MD; 1993.
6. Zwillich T. Patients, Doctors Don’t Discuss Rx Prices. WebMD Health News. March 17, 2009. www.webmd.com/news/20090317/patients-doctors-dont-discuss-rxprices. Accessed July 14, 2009.
7. American Society on Aging and American Society of Consultant Pharmacists Foundation. Adult Meducation. Overview of medication adherence. www.adult meducation.com/OverviewofMedicationAdherence.html. Accessed July 14, 2009.
8. Taylor H, Leitman R, eds. Out-of-pocket costs are a substantial barrier to prescription drug compliance. Health Care News. 2001;1:1-2.
9. Cutler DM, Everett W. Thinking outside the pillbox—medication adherence as a priority for health care reform. N Engl J Med. 2010;362:1553-1555. Epub 2010 Apr 7.
10. Lowes R. Patient-centered care for better patient adherence. Fam Pract Manag. 1998;5:46-47,51-54,57. www.aafp.org/fpm/980300fm/patient.html. Accessed November 10, 2010.
11. Haynes RB, Yao X, Degani A, et al. Interventions to enhance medication adherence. Cochrane Database Syst Rev. 2005;(4):CD000011.
12. Hahn SR, Park J, Skinner EP, et al. Development of the ASK-20 adherence barrier survey. Curr Med Res Opin. 2008;24:2127-2138. Epub 2008 Jun 12.
13. GlaxoSmithKline. ASK-20 Taking Medicine: What Gets in the Way [Survey]. August 2008. www.takingmeds.com/pdf/hcp/HMI719R0_ASK20SurveyEngl.pdf. Accessed August 25, 2010.
14. Matza LS, Yu-Isenberg KS, Coyne KS, et al. Further testing of the reliability and validity of the ASK-20 adherence barrier questionnaire in a medical center outpatient population. Curr Med Res Opin. 2008;24:3197-3206.
15. Burnier M. Long-term compliance with antihypertensive therapy: another facet of chronotherapeutics in hypertension. Blood Press Monit. 2000;5(suppl 1):S31-S34.
16. Krueger KP, Berger BA, Felkey B. Appendix D. Medication adherence and persistence. In: Wu HW, Nishimi RY, Kizer KW, eds. Improving use of prescription medications: a national action plan. Washington, DC: National Quality Forum; 2005. www.qualityforum.org/Publications/2005/10/Improving_Use_of_ Prescription_Medications__A_National_Action_Plan.aspx. Accessed Novem ber 16, 2010.