Table 2

Risk of events associated with poor adherence

Without time-varying BP*With time-varying BP
HR (95% CI)p ValueHR (95% CI)p Value
Any-cause hospitalisation
 CMA ≥0.50 vs <0.501.081 (1.032 to 1.123)0.00031.076 (1.037 to 1.128)0.0007
 CMA ≥0.60 vs <0.601.062 (1.017 to 1.099)0.00211.058 (1.022 to 1.104)0.0046
 CMA ≥0.70 vs <0.701.060 (1.022 to 1.094)0.00071.057 (1.025 to 1.097)0.0012
 CMA ≥0.80 vs <0.801.037 (1.006 to 1.069)0.02041.034 (1.003 to 1.066)0.0302
 CMA ≥0.90 vs <0.901.022 (0.993 to 1.050)0.11911.021 (0.994 to 1.051)0.1434
 Non-time-varying1.444 (1.356 to 1.537)<0.00011.429 (1.342 to 1.521)<0.0001
 All time-varying§1.049 (1.018 to 1.082)0.00211.047 (1.015 to 1.079)0.0034
 Censored at last fill1.032 (1.001 to 1.065)0.0431.030 (0.999 to 1.063)0.059
Myocardial infarction
 CMA ≥0.50 vs <0.501.616 (1.304 to 2.002)<0.00011.603 (1.293 to 1.987)<0.0001
 CMA ≥0.60 vs <0.601.534 (1.255 to 1.876)<0.00011.525 (1.247 to 1.866)<0.0001
 CMA ≥0.70 vs <0.701.332 (1.101 to 1.612)0.00321.326 (1.095 to 1.605)0.0038
 CMA ≥0.80 vs <0.801.281 (1.079 to 1.522)0.00471.277 (1.075 to 1.516)0.0054
 CMA ≥0.90 vs <0.901.141 (0.973 to 1.339)0.10521.142 (0.974 to 1.340)0.1022
 Non-time-varying1.539 (1.258 to 1.883)<0.00011.532 (1.258 to 1.883)<0.0001
 All time-varying§1.381 (1.161 to 1.642)0.00031.375 (1.156 to 1.635)0.0003
 Censored at last fill1.301 (1.093 to 1.550)0.0031.298 (1.090 to 1.546)0.003
Death
 CMA ≥0.50 vs <0.501.398 (1.160 to 1.685)0.00041.402 (1.163 to 1.689)0.0004
 CMA ≥0.60 vs <0.601.307 (1.095 to 1.561)0.00301.308 (1.095 to 1.562)0.0030
 CMA ≥0.70 vs <0.701.203 (1.016 to 1.424)0.03171.209 (1.020 to 1.433)0.0287
 CMA ≥0.80 vs <0.801.233 (1.061 to 1.433)0.00621.234 (1.06 to 1.436)0.0066
 CMA ≥0.90 vs <0.901.160 (1.008 to 1.334)0.03771.163 (1.010 to 1.340)0.0357
 Non-time-varying1.857 (1.581 to 2.182)<0.00011.857 (1.580 to 2.182)<0.0001
 All time-varying§1.403 (1.211 to 1.624)<0.00011.399 (1.207 to 1.623)<0.0001
 Censored at last fill1.262 (1.076 to 1.480)0.0041.261 (1.073 to 1.481)0.005
  • * Model was adjusted for baseline demographics (age, sex, socieoeconomic index); disease severity (baseline BP categorised as normal, pre-hypertension, hypertension; levels of ischaemic disease categorised as no ischaemic disease and no hyperlipidaemia and no cardiovascular disease risk factors, no ischaemic disease and treated hyperlipidaemia and no cardiovascular disease risk factors, no ischaemic disease and untreated hyperlipidaemia and cardiovascular disease risk factors, ischaemic disease and treated hyperlipidaemia, ischaemic disease and untreated hyperlipidaemia); comorbid conditions (body mass index categorised as underweight, normal weight, overweight, obese; chronic obstructive pulmonary disease; diabetes categorised as no diabetes, controlled diabetes without treatment and no complications, controlled diabetes with oral antidiabetics and complications, uncontrolled diabetes with oral antidiabetics and no complications, uncontrolled diabetes with oral antidiabetics and complications, controlled diabetes with insulin and no complications, controlled diabetes with insulin and complications, uncontrolled diabetes with insulin and no complications, uncontrolled diabetes with insulin and complications, untreated uncontrolled diabetes with no complications, untreated uncontrolled diabetes with complications; cancer diagnosis or treatment; psychiatric disorder categorised as none, treated psychiatric disorders, untreated psychiatric disorders; sleep apnoea; thyroid disorder categorised as none, treated thyroid disorder, untreated thyroid disorder); and number of hospitalisations in the pre-index period.

  • Model was adjusted for all covariates listed for Model * plus time-varying on-treatment quarterly BP level categorised as low (DBP ≤70), normal (SBP <120 and 70<DBP<80), prehypertension (120≤SBP<140 or 80≤DBP<90), stage I hypertension (140≤SBP<160 or 90≤DBP<100) and stage II hypertension (SBP ≥160 or DBP ≥100).

  • Model was adjusted for all baseline covariates listed for Model * plus non-time-varying BP level in the follow-up period; CMA threshold was <0.80 for non-adherent and ≥0.80 for adherent.

  • § Model was adjusted for all baseline covariates listed in Models * and ‡; all covariates were allowed to vary over time, with patients contributing to different levels of each covariate as disease status changed in the follow-up period; CMA threshold was <0.80 for non-adherent and ≥0.80 for adherent.

  • Model was adjusted for all baseline covariates listed in * above; the adherence ratio was calculated as the sum of days on which medication was available divided by the difference in days between the first and last fill, excluding the days supplied with the last fill.

  • CMA, medication adherence ratio (cumulative measure of medication availability); DBP, diastolic blood pressure; BP, blood pressure; HR, hazards ratio; SBP, systolic blood pressure.