Table 2

Sequence symmetry results of the 1000 simulations for different scenarios of medicine utilisation trends and different effect sizes, for a population size of 1 million and 5% incidence of use of each drug in the population
Crude sequence ratio Adjusted sequence ratio
Medicine utilisation scenario Pairs (N) Null SR Crude SR (95% CI) Cover-age (%) Power* (%) Relative bias (%) Adjusted SR (95% CI) Cover-age (%) Power* (%) Relative bias (%)
Expected ASR = 1.0
No trend 225 1.00 1.01 (0.75-1.30) 94.3 1.8 2.42 1.01 (0.75-1.30) 94.5 1.6 2.42
Gradual 41 0.85 0.90 (0.31-1.60) 86.8 0.0 −4.33 1.06 (0.37-1.88) 94.5 0.1 12.18
Constant 100 0.84 0.85 (0.52-1.23) 81.1 0.1 −12.58 1.01 (0.61-1.46) 94.7 0.2 3.73
Rapid 113 0.88 0.90 (0.57-1.27) 87.3 0.0 −7.75 1.03 (0.65-1.45) 95.2 1.3 4.97
Expected ASR = 1.2
No trend 247 1.00 1.21 (0.91-1.55) 95.9 20.7 2.84 1.21 (0.91-1.55) 95.7 20.8 2.84
Gradual 44 0.85 1.08 (0.40-1.91) 90.6 0.0 −3.73 1.27 (0.47-2.24) 96.9 0.5 12.90
Constant 110 0.84 1.04 (0.65-1.47) 81.5 1.3 −11.45 1.23 (0.77-1.75) 94.9 7.9 5.05
Rapid 124 0.88 1.07 (0.70-1.49) 87.4 1.4 −8.70 1.22 (0.79-1.70) 95.8 7.8 3.90
Expected ASR = 1.5
No trend 280 1.00 1.51 (1.15-1.92) 95.6 85.5 2.54 1.51 (1.15-1.92) 95.4 85.8 2.52
Gradual 50 0.85 1.36 (0.56-2.33) 86.7 0.6 −3.58 1.59 (0.66-2.74) 95.6 5.6 13.10
Constant 124 0.84 1.29 (0.84-1.82) 81.1 13.3 −11.39 1.54 (1.00-2.16) 96.2 45.4 5.12
Rapid 139 0.88 1.35 (0.90-1.87) 86.9 24.6 −7.53 1.54 (1.03-2.13) 95.9 52.3 5.20
Expected ASR = 2.0
No trend 337 1.00 2.03 (1.57-2.55) 96.1 100.0 2.91 2.03 (1.57-2.55) 96.2 100.0 2.89
Gradual 60 0.85 1.79 (0.81-3.01) 87.7 17.7 −4.59 2.10 (0.95-3.53) 95.6 40.6 11.87
Constant 146 0.84 1.72 (1.15-2.39) 80.6 76.1 −11.44 2.04 (1.37-2.84) 96.1 97.0 5.08
Rapid 166 0.88 1.79 (1.23-2.45) 87.5 90.7 −7.96 2.04 (1.40-2.79) 97.0 98.7 4.71
Expected ASR = 3.0
No trend 450 1.00 3.03 (2.40-3.77) 96.4 100.0 2.84 3.03 (2.40-3.77) 96.2 100.0 2.85
Gradual 78 0.85 2.67 (1.30-4.47) 87.8 90.8 −3.94 3.14 (1.52-5.24) 96.3 97.9 12.66
Constant 192 0.84 2.59 (1.79-3.55) 80.5 100.0 −11.16 3.07 (2.12-4.21) 97.5 100.0 5.39
Rapid 219 0.88 2.67 (1.90-3.59) 83.4 100.0 −8.50 3.03 (2.16-4.09) 96.0 100.0 4.12
Expected ASR = 0.6
No trend 300 1.00 0.61 (0.46-0.76) 95.6 99.0 1.66 0.61 (0.46-0.76) 95.6 99.0 1.68
Gradual 56 0.85 0.53 (0.23-0.85) 86.8 77.9 −9.71 0.62 (0.27-1.00) 95.2 57.0 5.92
Constant 136 0.84 0.51 (0.33-0.70) 79.2 98.2 −14.48 0.61 (0.39-0.83) 94.9 90.0 1.46
Rapid 154 0.88 0.53 (0.35-0.71) 83.7 98.6 −11.09 0.60 (0.40-0.81) 94.5 91.1 1.14
Expected ASR = 0.8
No trend 254 1.00 0.81 (0.61-1.03) 95.4 44.5 2.09 0.81 (0.61-1.03) 95.7 44.5 2.10
Gradual 46 0.85 0.72 (0.28-1.21) 89.1 32.7 −6.68 0.84 (0.33-1.42) 95.9 16.6 9.43
Constant 114 0.84 0.69 (0.43-0.97) 79.5 60.4 −13.09 0.82 (0.50-1.15) 95.0 29.1 3.14
Rapid 129 0.88 0.71 (0.46-0.98) 83.2 55.8 −9.95 0.81 (0.52-1.12) 94.1 30.9 2.47

*Note: When the expected Sequence ratio =1, Power is the type I error.

Pratt et al.

Pratt et al. BMC Medical Research Methodology 2014 14:66   doi:10.1186/1471-2288-14-66

Open Data