Communicating risks at the population level: application of population impact numbers
BMJ 2003; 327 doi: https://doi.org/10.1136/bmj.327.7424.1162 (Published 13 November 2003) Cite this as: BMJ 2003;327:1162
Data supplement
Corrected calculation of population attributable risk (PAR) in table 2
The formula used for PAR needs to be adjusted as follows if the component relative risks (RR) relate levels of an exposure to a subset of the population rather than to all other people in the population.1 The relative risks used in the table were a polytomous series relating outcomes for people in specific ranges of serum cholesterol concentration to outcomes for people with lower values of serum cholesterol,2 rather than dichotomous relative risks relating people at an exposure level to all other people in the population.
For a dichotomous relative risk:
For a polytomous series of relative risks:
PAR = population attributable risk (Levin definition3)
RR = relative risk
Pe = proportion of population exposed to the risk factor (level)
- Hanley J. A heuristic approach to the formulas for population attributable fraction. J Epidemiol Community Health 2001;55:508-14.
- McPherson K, Britton A, Causer L. Coronary heart disease. Estimating the impact of changes in risk factors. London: Stationery Office, National Heart Forum, 2002.
- Levin K. The occurrence of lung cancer in man. Acta Unio Inter Contra Cancrum 1953;19:531.
Table 2 The impact of blood cholesterol concentration on premature death from coronary heart disease among people aged <75 years over three years in a typical UK general practice population, split by sex
Blood cholesterol concentration
Relevant population size out of 10 000 (n)
Predicted three year incidence of outcome locally (Ip3)
Estimated local prevalence of risk factor (Pe)
Relative risk from best evidence (RR)
Population proportion of outcome attributable to risk factor (PAR)
Estimated total number affected locally in the next three years (nd)
Population impact number by eliminating risk factor (PIN-ER-3)
5.2-6.5 mmol/l:
Men
4664
0.003624
0.41
1.75
0.17
16.90
2.87
Women
4591
0.001464
0.39
1.75
0.16
6.72
1.04
6.5-7.8 mmol/l:
Men
4664
0.003624
0.21
2.57
0.18
16.90
3.08
Women
4591
0.001464
0.22
2.57
0.18
6.72
1.23
>7.8 mmol/l:
Men
4664
0.003624
0.07
3.46
0.10
16.90
1.61
Women
4591
0.001464
0.1
3.46
0.13
6.72
0.88
PIN-ER-t = n × Ip ×PAR
PARi = {Pei ×(RRi-1)}/{1 + ∑i=1 to k[Pei × (RRi-1)]}
Risk factor = high blood cholesterol
Outcome = death from coronary heart disease
Relative risks are relative to values lower than the levels shown
Cited by...
- Model-based economic evaluation of ice cleat distribution programmes for the prevention of outdoor falls among adults from a Swedish societal perspective
- Predictive risk algorithms in a population setting: an overview
- Evidence-based commissioning: using population impact measures to help primary care trusts estimate the benefit of interventions in diabetes and heart failure
- Communicating the benefits of breast feeding
- The population effect of crime and neighbourhood on physical activity: an analysis of 15 461 adults
- Potential population impact of the UK government strategy for reducing the burden of coronary heart disease in England: comparing primary and secondary prevention strategies.
- Alcohol-Related Injuries: Evidence for the Prevention Paradox