Language-concordant automated telephone queries to assess medication adherence in a diverse population: a cross-sectional analysis of convergent validity with pharmacy claims

BMC Health Services Research

Margaret A. Handley¹  Judy Quan¹  Neda Ratanawongsa¹  Urmimala Sarkar¹  Dean Schillinger¹ 

[1] General Internal Medicine and UCSF Center for Vulnerable Populations at San Francisco General Hospital and Trauma Center, University of California;

关键词: Diabetes;    Health information technology;    Medication adherence;    Safety net clinics;    Medicaid;    Managed care;   

DOI  :  10.1186/s12913-018-3071-4

来源: DOAJ

【 摘 要 】

Abstract Background Clinicians have difficulty accurately assessing medication non-adherence within chronic disease care settings. Health information technology (HIT) could offer novel tools to assess medication adherence in diverse populations outside of usual health care settings. In a multilingual urban safety net population, we examined the validity of assessing adherence using automated telephone self-management (ATSM) queries, when compared with non-adherence using continuous medication gap (CMG) on pharmacy claims. We hypothesized that patients reporting greater days of missed pills to ATSM queries would have higher rates of non-adherence as measured by CMG, and that ATSM adherence assessments would perform as well as structured interview assessments. Methods As part of an ATSM-facilitated diabetes self-management program, low-income health plan members typed numeric responses to rotating weekly ATSM queries: “In the last 7 days, how many days did you MISS taking your …” diabetes, blood pressure, or cholesterol pill. Research assistants asked similar questions in computer-assisted structured telephone interviews. We measured continuous medication gap (CMG) by claims over 12 preceding months. To evaluate convergent validity, we compared rates of optimal adherence (CMG ≤ 20%) across respondents reporting 0, 1, and ≥ 2 missed pill days on ATSM and on structured interview. Results Among 210 participants, 46% had limited health literacy, 57% spoke Cantonese, and 19% Spanish. ATSM respondents reported ≥1 missed day for diabetes (33%), blood pressure (19%), and cholesterol (36%) pills. Interview respondents reported ≥1 missed day for diabetes (28%), blood pressure (21%), and cholesterol (26%) pills. Optimal adherence rates by CMG were lower among ATSM respondents reporting more missed days for blood pressure (p = 0.02) and cholesterol (p < 0.01); by interview, differences were significant for cholesterol (p = 0.01). Conclusions Language-concordant ATSM demonstrated modest potential for assessing adherence. Studies should evaluate HIT assessments of medication beliefs and concerns in diverse populations. Trial registration NCT00683020, registered May 21, 2008.

【 授权许可】

Unknown