Skip to main content

Table 4 Adjusted hazard and rate ratios for new Z-drug prescription and new prescriptions, GP consultations, and hospital admissions for people with dementia

From: Adverse effects of Z-drugs for sleep disturbance in people living with dementia: a population-based cohort study

Outcome   Comparator
Z-drug (n = 3532) Sleep disturbance, no sedative-hypnotic (n = 1833) No Z-drug, proximal GP consultation (n = 10,214) Benzodiazepine (n = 5172)
Number of hospital admissions
 Rate per 100PY (events) 126.7 (1944) 93.0 (1671) 107.7 (3563) 129.4 (2441)
 Age, sex-adjusted RR (95% CI) NA 1.34 (1.20–1.49) 1.26 (1.15–1.39) 0.98 (0.89–1.09)
 Fully adjusted RR (95% CI)a NA 1.26 (1.13–1.40)b 1.17 (1.07–1.27)b 0.92 (0.84–1.01)
Number of GP consultations
 Rate per 100PY (events) 1387.8 (21292) 1124.6 (20209) 1511.7 (50021) 1502.6 (28335)
 Age, sex-adjusted RR (95% CI) NA 1.29 (1.22–1.37) 0.97 (0.93–1.01) 0.94 (0.90–0.98)
 Fully adjusted RR (95% CI)a NA 1.17 (1.12–1.23)b 1.07 (1.04–1.11)b 0.96 (0.93–1.00)
New antipsychotic prescription
 Incidence rate per 100PY (events) 38.7 (331) 10.3 (130) 15.2 (227) 75.4 (532)
 Age, sex-adjusted HR (95% CI) NA 2.53 (2.03–3.14) 3.68 (3.10–4.37) 0.79 (0.68–0.90)
 Fully adjusted HR (95% CI)a NA 2.37 (1.84–3.04)b 3.85 (3.18–4.65)b 0.86 (0.74–0.99)
New antidepressant prescription
 Incidence rate per 100PY (events) 23.8 (172) 8.2 (82) 10.3 (199) 30.3 (242)
 Age, sex-adjusted HR (95% CI) NA 2.24 (1.70–2.94) 2.52 (2.05–3.10) 0.86 (0.71–1.05)
 Fully adjusted HR (95% CI)a NA 2.32 (1.65–3.25)b 2.65 (2.09–3.37)b 0.88 (0.71–1.10)
New antibiotic prescription
 Incidence rate per 100PY (events) 109.1 (791) 61.2 (517) 89.5 (1738) 111.0 (1096)
 Age, sex-adjusted HR (95% CI) NA 1.46 (1.30–1.65) 1.27 (1.17–1.38) 1.01 (0.92–1.10)
 Fully adjusted HR (95% CI)a NA 1.34 (1.17–1.52)b 1.19 (1.08–1.30)b 1.05 (0.95–1.15)
  1. Abbreviations: HR hazard ratio, CI confidence interval, PY person-years, RR rate ratio
  2. aAdjusted for all covariates listed in Table 1
  3. bFully adjusted HR remaining statistically significant after controlling the false discovery rate to < 5% (based on 13 outcomes)