The UK society score10
| Preoperative factor | Death (%) | Likelihood ratio | Weight of evidence | |
|---|---|---|---|---|
| Overall | 5.2 | 0.05 | −29 | |
| Age (years) | 20 to 39 | 1.5 | 0.4 | −9 |
| 40 to 49 | 2.3 | 0.4 | −8 | |
| 50 to 59 | 3.5 | 0.7 | −4 | |
| 60 to 69 | 5 | 1 | 0 | |
| 70 to 79 | 9.2 | 1.9 | 6 | |
| 80 to 88 | 22.2 | 5.4 | 17 | |
| Sex | Male | 5.1 | 1 | 0 |
| Female | 8.4 | 1.7 | 5 | |
| Angina | None to class IV | 2.7 to 7.7 | 0.5 to 1.5 | −7 to 4 |
| Dyspnoea | None to class IV | 3.8 to 14.8 | 0.7 to 3.2 | −1 to 12 |
| COPD | No | 5.7 | 1.1 | 1 |
| Yes | 9.4 | 2 | 7 | |
| Emphysema | No | 5.8 | 1.1 | 1 |
| Yes | 7.7 | 1.6 | 5 | |
| Asthma | No | 5.8 | 1.1 | 1 |
| Yes | 7.1 | 1.4 | 4 | |
| Pre-op support | None | 5.2 | 1 | 0 |
| IABP | 28.6 | 8.3 | 21 | |
| Ventilated | 25 | 7.8 | 21 | |
| Inotropes | 17.6 | 4.4 | 15 | |
| > 1 of the above | 58.3 | 24.8 | 32 | |
| Renal | None | 4.1 | 0.8 | −2 |
| History | 12.3 | 2.6 | 10 | |
| Dialysis | 12.5 | 3 | 11 | |
| Smoking | Never | 5.6 | 5.6 | 1 |
| Stopped < 5 years | 3.6 | 0.7 | −4 | |
| Stopped > 5 years | 5.7 | 1.1 | 1 | |
| Smoker | 6.3 | 1.2 | 2 | |
| Diabetes | None | 5 | 1 | 0 |
| Diet | 7.6 | 1.6 | 4 | |
| Oral medication | 5.7 | 1.1 | 1 | |
| Insulin | 6.6 | 1.4 | −2 | |
| Hypertension | No | 4.3 | 0.8 | −2 |
| Yes | 6.4 | 1.3 | 2 | |
| Vascular disease | No | 4.1 | 0.8 | −2 |
| Yes | 8.4 | 1.7 | 5 | |
| Cerebral | No | 5 | 1 | 0 |
| TIA | 12.4 | 2.7 | 10 | |
| CVA | 10.5 | 2.2 | 8 | |
| LMS | No | 5 | 1 | 0 |
| Yes | 6.4 | 1.3 | 2 | |
| Last MI | None | 4.7 | 0.9 | −1 |
| < 24 hours | 22.7 | 5.5 | 17 | |
| 1 to 30 days | 9.1 | 1.9 | 6 | |
| > 30 days | 5.2 | 1 | 0 | |
| Last PTCA | None | 5 | 1 | 0 |
| < 24 hours | 10.3 | 2.2 | 8 | |
| 1 to 30 days | 5.6 | 1.3 | 3 | |
| > 30 days | 4.6 | 0.9 | −1 | |
| Ejection fraction | 0 to 29% | 7.1 | 1.4 | 4 |
| 30 to 50% | 4.6 | 0.9 | −1 | |
| > 50% | 4.9 | 0.9 | −1 | |
| Op sequence | First | 4.7 | 0.9 | −1 |
| Re-operation | 13.3 | 2.8 | 10 | |
| Op timing | Elective | 3.5 | 0.7 | −4 |
| Urgent | 6.4 | 1.2 | 2 | |
| Emergency | 14.1 | 3 | 11 | |
| Salvage | 40 | 12.4 | 25 | |
LMS, left main stem stenosis > 50%; COPD, chronic obstructive pulmonary disease.
Bayes theorem is the formula that provides the relation between the prior odds and posterior odds: posterior odds = likelihood ratio × prior odds; where the likelihood ratio expresses how much more likely it is that a patient with a given condition should fall among those who die rather than survive. Because this is repeated for each risk factor assuming they are contributing separately, it usually leads to an overprediction of mortality and the contribution of each factor should be damped down accordingly.
For this study we have used: posterior odds = likelihood ratio0.8 × prior odds.
To turn the risk factor table into a risk prediction a weight of evidence is calculated that expresses how much an observed feature adds to or subtracts from the evidence of mortality where weight of evidence = 10 log (likelihood ratio).
The scoring system therefore consists of adding up the weights of evidence for the observed features and adding the results to the staring score of −29, which is the 10 log (number of deaths/number of survivors). This gives a total score “S” that may be transformed to a predicted percentage mortality by the equation: p = 100/[1 + exp (−S/10)].