Table 3

Multivariable GEE logistic regression analysis evaluating the association between measures of drug use and inadequate BP control among primary care patients with unhealthy drug use (n=229 observations)*

Main independent variableUnadjusted modelsAdjusted models†Final adjusted model†,
OR (95% CI)p ValueOR (95% CI)p ValueOR (95% CI)p Value
Number of days of drug use in past 30 days
 Highest tertile (23–30 days)0.82 (0.44 to 1.52)0.520.70 (0.37 to 1.32)0.280.67 (0.31 to 1.46)0.31
 Middle tertile (3–22 days)0.88 (0.47 to 1.63)0.680.82 (0.43 to 1.57)0.550.72 (0.36 to 1.41)0.34
 Lowest tertile (0–2 days)111
Drug type
 Cocaine1.44 (0.75 to 2.78)0.281.44 (0.75 to 2.78)0.281.12 (0.52 to 2.40)0.78
 Opioids1.61 (0.67 to 3.85)0.281.76 (0.71 to 4.36)0.221.77 (0.70 to 4.48)0.23
 Other0.81 (0.08 to 8.25)0.861.00 (0.09 to 10.67)0.9980.84 (0.07 to 9.60)0.89
 Marijuana111
Drug use severity
 Highest1.48 (0.74 to 2.93)0.271.47 (0.69 to 3.12)0.321.50 (0.68 to 3.30)0.31
 Middle0.99 (0.51 to 1.95)0.990.97 (0.49 to 1.93)0.931.14 (0.53 to 2.47)0.74
 Lowest111
  • *Inadequate BP control defined as systolic BP of 140 or greater or diastolic BP of 90 or greater as specified by the National Committee for Quality Assurance (NCQA) and others. Three separate GEE (generalised estimating equations) logistic regression models evaluated the association between each drug use measure (days use of the main drug in past 30 days, drug type, and drug use severity) and inadequate BP control.

  • †Adjusted models included age, sex, race/ethnicity, health insurance, depression, and Charlson score.

  • ‡One multivariable model with same covariates and all three drug use measures. Age was the only statistically significant covariate in the days of drug use model (AOR=0.96; 95% CI 0.93 to 0.997, p=0.034) and in the drug type model (AOR=0.97; 95% CI 0.93 to 0.99, p=0.047). There were no significant covariates in the drug use severity model or in the final adjusted model.

  • AOR, adjusted OR; BG, blood glucose; BP, blood pressure; GEE, generalised estimating equations.