Barriers to adherence and hypertension control in a racially diverse representative sample of elderly primary care patients

Pharmacoepidemiol Drug Saf. 2009 Aug;18(8):672-81. doi: 10.1002/pds.1766.


Purpose: To examine the effect of antihypertensive adherence on blood pressure and barriers to adherence in racially diverse elderly patients.

Methods: Telephone survey of a representative sample of 300 of all 3416 hypertensive patients aged >70 from four urban primary care practices. From electronic records, we calculated subjects' annual mean systolic blood pressure. We asked about the last missed antihypertensive dose in six time intervals. Based on association with blood pressure control, non-adherence was defined as missing any dose in the past 3 months. Subjects were also asked about six domains of adherence barriers: health, personal support, drug coverage, medication filling and use, doctor-patient interaction and knowledge. All models adjust for demographics, treatment regimen and sampling weights.

Results: The 202 subjects (67% response rate) were: female (65.9%), black (64.8%), mean age 77.4 years (5.49) and on mean 2.4 (SD 1.3) antihypertensive drugs. Mean annual systolic pressure for non-adherent subjects (22% of the cohort) was higher than adherent subjects (137.7 vs.133.4 mmHg, p = 0.065). After adjustment, the association between adherence and blood pressure was stronger in black than white patients (p = 0.007). In an initial model, being unaware of Medicare Part D had a lower adjusted odds ratio (AOR) of adherence (p < 0.05). In the final model, adherence barriers were: medication filling/use (run out of pills [AOR 0.25, CI 0.09-0.66] and 28% reduction per each of eight barriers); doctor-patient interaction (less important to discuss hypertension [AOR 0.32, CI 0.12-0.84]); and knowledge (38% lower AOR per incorrect answer about diseases unrelated to hypertension).

Conclusion: Self-reported adherence was associated with a higher blood pressure, especially in elderly black patients. To promote adherence, our data suggest targeting: filling prescriptions, prioritizing hypertension care and educating about effects of hypertension.

Publication types

  • Multicenter Study
  • Research Support, Non-U.S. Gov't

MeSH terms

  • African Americans / statistics & numerical data
  • Aged
  • Aged, 80 and over
  • Antihypertensive Agents / therapeutic use*
  • Blood Pressure / drug effects*
  • European Continental Ancestry Group / statistics & numerical data
  • Female
  • Health Care Surveys
  • Health Services for the Aged / statistics & numerical data*
  • Humans
  • Hypertension / drug therapy*
  • Hypertension / ethnology
  • Hypertension / physiopathology
  • Linear Models
  • Logistic Models
  • Male
  • Medication Adherence / statistics & numerical data*
  • New York City / epidemiology
  • Odds Ratio
  • Pennsylvania / epidemiology
  • Primary Health Care / statistics & numerical data*
  • Risk Assessment
  • Risk Factors
  • Surveys and Questionnaires
  • Treatment Outcome


  • Antihypertensive Agents