Electronic Data Capture and Coding Tool for Medication Regimen Complexity

Our ongoing research focuses on approaches to increase the efficiency and effectiveness of health care and health care systems. In regard to medications, to obtain effectiveness expected from medications, patients must take them as prescribed while adhering to special instructions. Medication adherence decreases, however, as medication regimen complexity increases (e.g. multiple medications, multiple doses, specific dietary or time requirements). The results can be reduced effectiveness and cost-effectiveness of treatments, poor outcomes for both patients and payers. A common measure of medication regimen complexity (MRC) is a simple count of prescribed medications. However, medication regimen complexity consists of multiple additional characteristics such as the dosage forms, dosing frequencies, and additional directions for use. There is a gap in knowledge about the degree and range of medication regimen complexity and possibly associated risk for patients with chronic diseases.

The objective of the Skaggs Scholars project was to measure medication regimen complexity for defined clinical populations with chronic diseases: hypertension; diabetes; geriatric depression; and HIV. We also tested for differences in medication regimen complexity within and between the clinical populations at two collaborating Skaggs institution study sites at the University of Colorado and the University of California San Diego. Medication regimen complexity measures could be used as a tool for targeting patients who may benefit most from pharmacist interventions.

Libby AM, Fish DN, Hosokawa PW, Linnebur SA, Metz KR, Nair KV, Saseen JJ, Vande Griend JP, Vu SP, Hirsch JD. Patient-Level Medication Regimen Complexity Across Populations With Chronic Disease. Clin Ther. 2013; 35(4): 385-398. doi:pii: S0149-2918(13)00075-1. 10.1016/j.clinthera.2013.02.019. PMID: 23541707

Using the published Medication Regimen Complexity Index (MRCI) tool by George and colleagues—with author’s permission and publisher’s permission from The Annals (George J, Phun, Y-T, Bailey, MJ, Kong, DCM, Stewart, K. Development and validation of the medication regimen complexity index. The Annals of Pharmacotherapy. 2004; 38:1369-1376, DOI 10.1345/aph.1D479)—we developed a Microsoft Access Database for that is a derivative product from the published paper-and-pencil tool. We made minor adaptations to the dosage forms section only to reflect prescribing practices in the U.S., rather than in Australia where the tool was developed. Our electronic data capture and coding tool calculates patient-level total MRCI scores and patient-level sub-scores by drug category (target disease/cohort defining; all other prescription medications not associated with the cohort-defining disease state; over-the-counter medications) that are new contributions to the MRCI. For scholarly purposes, not commercial use, we have made it available for downloading with instructions for how we used it. The links can be found at the bottom of this web page. To enhance comparability across studies if researchers use the tool as it is provided, it should be noted as Version 1.0. The following downloads are available:

1. MRC electronic data capture tool (MS Access)—save it on local computer first and then open it; click the enable button on the top; print the directions; then enter the form;
2. MRC Additional Instructions we used during our study to clarify approaches to coding but were not specified in the original index’s instructions; these supplement the original directions embedded in the MRCI tool and were documented to assure consistent coding between clinician coders; and
3. MRC Training Guide for electronic data capture tool—step-by-step instructions using example cases for coding with this database tool.

Our coding tool also includes an adaptation for a patient-level summary score and subscores. This tool allows uniform forms for three separate subscores of medication regimen complexity: for disease-specific prescription medications (e.g. antiretrovirals for the HIV cohort), non-disease specific prescription medications (e.g. all other prescription medications not antiretrovirals for the HIV cohort), and over-the-counter medications. Programmed reports that applied weights allowed pharmacist coders to conduct quality checks during coding. Patients were given a unique cohort-specific identifier and their medication list was coded by a clinical pharmacist. Random samples of 30 patients within each cohort were coded by a different pharmacist rather than the primary coder to assess inter-rater reliability. Prior to coding, the study team coded case examples together in a group meeting, discussed definitions and rationale, and made decision rules for cases not clearly addressed in the tool instructions, such as when to code a medication as prescription versus over-the-counter in cases where the medication can be obtained from either source. There was minimal and non-significant variability between primary and secondary pharmacist raters. Weighted Kappa statistics between coders one and two for each cohort were: Geriatric: 0.9387; HIV: 0.9307; Diabetes: 0.8457; and Hypertension: 0.9396.