Table 4

Multivariable GEE logistic regression analysis evaluating the association between measures of drug use and inadequate BG control (n=83 observations)*

Unadjusted modelsAdjusted models†
Main independent variableOR (95% CI)p ValueOR (95% CI)p Value
Number of days of drug use in past 30 days
 Highest tertile (19–30 days)0.43 (0.13 to 1.41)0.160.27 (0.07 to 1.10)0.07
 Middle tertile (5–18 days)0.93 (0.33 to 2.66)0.901.01 (0.38 to 2.69)0.99
 Lowest tertile (0–4 days)11
Drug type
 Cocaine4.44 (1.01 to 19.54)0.0488.82 (1.86 to 41.90)0.006
 Opioids0.56 (0.13 to 2.32)0.420.53 (0.12 to 2.33)0.40
 Other1.67 (0.14 to 20.08)0.694.71 (0.26 to 86.80)0.30
 Marijuana11
Drug use severity‡
 Highest3.05 (1.01 to 9.23)0.0482.57 (0.79 to 8.33)0.12
 Middle2.05 (0.67 to 6.24)0.211.68 (0.48 to 5.85)0.42
 Lowest11
  • *Inadequate BG control defined as glycated haemoglobin of 8% or greater as specified by the National Committee for Quality Assurance (NCQA). Three separate GEE (generalised estimating equations) logistic regression models were used to evaluate the association between each drug use measure (days use of the main drug in the past 30 days, drug type and drug use severity) and inadequate BG control. All available data at baseline and 6-month interview was analysed as repeated measures.

  • †Adjusted models included age, gender, race/ethnicity. Age was the only statistically significant covariate in the days of drug use model (AOR=0.94 95% CI 0.89 to 0.99, p=0.02) and in the drug type model (AOR=0.93 95% CI 0.88 to 0.99, p=0.02). There were no significant covariates in the analysis of drug use severity and inadequate BG control.

  • ‡Alcohol, Smoking, and Substance Involvement Screening Test (ASSIST) global drug score.

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