Table 4.
Simulated population risk estimates using the g-formula. Hypothetical interventions only following high cholesterol, hypertension, diabetes, angina, stroke or CABG
| Intervention | 20-year risk | Population risk ratio | Population risk difference | Number needed to assessa | Average percent intervened uponb | Cumulative percent intervened uponb |
|---|---|---|---|---|---|---|
| (0) No intervention | 3.68 (3.56 to 4.09) | 1 | 0 | – | 0 | 0 |
| (10) Quit smoking | 3.17 (3.04 to 3.56) | 0.86 (0.83 to 0.89) | −0.50 (−0.68 to −0.40) | 199 | 3 | 17 |
| (11) Exercise at least 30 min/day | 3.11 (2.68 to 3.65) | 0.84 (0.71 to 0.94) | −0.57 (−1.13 to −0.24) | 175 | 29 | 72 |
| (12) Keep diet score in the top 2 quintiles | 3.37 (3.22 to 3.73) | 0.92 (0.85 to 0.96) | −0.31 (−0.57 to −0.16) | 323 | 29 | 69 |
| (13) Consume at least 5g alcohol per day | 3.29 (2.98 to 3.81) | 0.89 (0.79 to 0.98) | −0.39 (−0.82 to −0.07) | 259 | 33 | 67 |
| (14) Maintain BMI <25 | 3.72 (3.47 to 4.13) | 1.01 (0.94 to 1.04) | 0.04 (−0.24 to 0.15) | N/A | 32 | 58 |
| (15) ‘Low-risk’ lifestyle (10–12 combined) | 2.55 (2.19 to 2.96) | 0.69 (0.58 to 0.77) | −1.12 (−1.65 to −0.88) | 89 | 40 | 76 |
| (16) ‘Low-risk’ lifestyle (10–12, 14 combined) | 2.47 (2.15 to 2.97) | 0.67 (0.56 to 0.77) | −1.20 (−1.66 to −0.87) | 83 | 48 | 77 |
| (17) ‘Low-risk’ lifestyle (10–13 combined) | 2.25 (1.88 to 2.69) | 0.61 (0.50 to 0.70) | −1.43 (−1.90 to −1.11) | 70 | 47 | 77 |
| (18) ‘Low-risk’ lifestyle (10–14 combined) | 2.25 (1.88 to 2.70) | 0.61 (0.50 to 0.71) | −1.43 (−1.91 to −1.12) | 70 | 51 | 78 |
aOnly those meeting the criteria of having a history of high cholesterol, hypertension, diabetes, angina, stroke or CABG, are intervened upon. We are, however, looking at the 20-year risk in the full population. Thus, this version of the ‘number needed to treat’ represents the number that need to be assessed, and intervened upon if the criteria are met, in order to prevent one event. Because the estimated risk for the BMI intervention was greater than the estimated risk under no intervention, the concept of number needed to assess is not applicable.
bThe average percent intervened on is the average, across all 2-year time periods, of the percent of the population required to change behaviour in a given 2-year time period. The cumulative percent intervened on is the percent of the population required to change behaviour in at least one 2-year time period.