Table 3

Interrupted time series analysis modelling adjusted relative risk and crude monthly mortality for all alerts (diagnoses and procedures), AMI alerts and sepsis alerts

All alerts
Risk ratio95% CIChange (%)Risk ratio95% CIChange (%)Risk ratio95% CIChange (%)
Modelling adjusted risk
 Prealert trend1.05(1.04 to 1.05)5***1.05(1.02 to 1.08)5***1.04(1.02 to 1.06)4***
 Level change (after lag)0.39(0.35 to 0.44)−61***0.43(0.28 to 0.67)−57***0.41(0.29 to 0.59)−59***
 Postlag trend0.99(0.98 to 1.00)−10.99(0.94 to 1.04)−11.00(0.96 to 1.05)0
Modelling crude risk
 Prealert trend1.07(1.06 to 1.07)7***1.08(1.05 to 1.11)8***1.04(1.03 to 1.06)4***
 Level change (after lag)0.25(0.23 to 0.28)−75***0.27(0.18 to 0.41)−73***0.31(0.24 to 0.39)−69***
 Postalert trend0.99(0.98 to 1.00)−10.98(0.95 to 1.01)−20.99(0.97 to 1.01)−1
  • Risk ratios are the model estimated ratios of relative risk of death (observed number/expected number) in the adjusted model and the rate ratios of death rate (observed number/admission number) in the crude model. Trend risk ratios are monthly increases/decreases. Our model measures the trend prior to an alert, the level change during varying lag periods and postlag trend in relative risk of death. Models are adjusted for autocorrelation. The model uses generalised estimating equations and the Wald test statistical significance was  ***p<0.001. The 172 alerts were generated between January 2011 and November 2013 and sent to 93 acute National Health Service trusts in England.

  • AMI, acute myocardial infarction.