Medication Adherence: Toppling Barriers With A Learning Health System
Barriers that hinder proper medication use stretch across the individual, family, healthcare provider and community-level. Learning Health Systems bring all these areas under a single umbrella, and thus are uniquely positioned to coordinate, implement, and assess the effectiveness of various adherence improvement strategies. This article describes the implementation of a barrier reduction strategy using a Learning Health System. Examples are provided using an epilepsy population.
More Than A Prescription
Ideally, there would be a direct line going from writing a prescription to the correct and complete use of the medication. Unfortunately, the line is often broken as poor medication adherence is a consistent and sizeable finding across numerous patient populations. It is common to find 40% or more of patients do not take medications as prescribed. In fact, the World Health Organization suggested it may be equally effective to develop strategies that increase medication adherence as compared to develop new therapies (see here). Full appreciation of the factors related to poor adherence nests the individual within a family, healthcare provider and community (see image below). Although the aims stretch well beyond just medication use, Learning Health Care Systems bring together these contributing factors under a single umbrella, and thereby offer considerable promise for coordinating the multiple strategies needed to improve adherence.
This article describes the implementation of one such strategy using a case-example from collaborative work with the Epilepsy Learning Health System (see here). This is an applied, follow-up example to a proposed theoretical Disease and Well-being model previously described, “Learning Health Systems: Removing the Tinted Glasses.”
Because of the multiple factors involved, no single strategy will eliminate all medication non-compliance. Instead, multiple strategies are needed and used in some targeted combination to address the needs of a specific patient. The starting point for this article is with a well-established and effective toolkit developed by the Epilepsy Learning Healthcare System (learn more here). The “Barriers to Medication Adherence Toolkit” (BMAT) is a self-report questionnaire designed to assess the 18 most common medication adherence barriers. The lead investigator for the BMAT is Dr. Avani Modi (profile and contact information).
All strategies have an implementation plan. Often the initial step involves matching the requirements to an appropriate technology.
The main requirements for the Barriers toolkit were:
A Web-Based Patient Portal
Questionnaire data entry associated with a clinic visit
An integrated, secure messaging system (e.g., patient questions or other provider communications)
Automatic delivery of educational materials based on questionnaire results
Facilitate Staff Follow-up
Monitor and track patients reporting a medication barrier
Schedule and document follow-up (e.g., phone call) with patients to address barrier management
Readily add clinics
Aggregate reporting across clinics
Bringing the BMAT into an electronic delivery system offered an opportunity to enhance the questionnaire’s accessibility. Minimizing the reading-level requirement was the main aim and was addressed in three ways (additional form design techniques, see “Eyes vs. Fingers: Form Design Lessons For Accurate Data Entry”).
1) Visual Cues
Often there are few alternatives to a text-heavy approach to questionnaire design. However, all items on the BMAT were readily paired with a visual icon representing the question concept (see example items below). Moreover, pairing items with icons can also help declutter reports (for examples, see the article “Patient Registries: Clinical Report Design Techniques”).
2) Enabling Audio
A recording of the instructions and each item read aloud was created and made accessible by clicking a speaker icon (see above image). This eliminated the reading requirement entirely.
3) Video Instruction
For more difficult concepts, video instruction is an effective mechanism to increase comprehension. For example, a question about syncope on a sport pre-participation survey (see article here). Videos were created for items thought benefiting from additional explanation and directly integrated into the questionnaire (see below). For instance, a brief (< 20 seconds) video demonstrating several examples for the item “Difficulty getting to the pharmacy to pick up medicine”.
Avoiding The Kitchen Sink
Information overload can be overwhelming, often causing patients to become dis-engaged with intervention efforts (for a discussion of related issues see “Disease Information: Making It Relevant To Patients”). Thus, patients reporting two or more barriers were asked to identify the most important barrier. Then, using the BMAT algorithm, appropriate educational materials were automatically gathered into a report format which was immediately accessible to patients by clicking a button (see image below showing part of a report).
In addition to the information materials, identification of a medication barrier also triggered notification to clinical staff for follow-up. Tools were integrated into the system to assist with the follow-up process.
Tools were designed to help clinical staff manage the list of patients requiring follow-up. For instance, monitoring, tracking and documenting contact attempts, as well as the scheduled follow-up phone call or clinic visit.
For some barriers the algorithm generated materials designed to assist patients with developing a plan to address the barrier. For example, a form to organize the needed components of a plan. These forms were made available electronically in the portal. Both staff and patients had access to the forms to facilitate collaboration in the development and documentation of the plan.
There are multiple additional components to the Epilepsy Learning Health System that will be discussed in subsequent articles. Two things of note. First, the BMAT is just one component to a broad range of patient / staff entered and EMR extracted data that can be used to deliver 1) personally relevant disease information and educational materials to patients, and 2) consolidated information for healthcare staff to facilitate clinical decision making. Second, the Epilepsy Learning Health System is housed within an integrated research application (e.g., clinical trials, observational studies, surveys, registries, etc.) so as to help leverage and coordinate the work of others and eliminate redundant effort. More to come…