Skip to main content

Table 3 Survival functions assigned to specified health states: sources, limitations and strengths

From: Why choice of metric matters in public health analyses: a case study of the attribution of credit for the decline in coronary heart disease mortality in the US and other populations

  Data used Data source Limitations of assumption Strengths of assumption
Treatments CHD patients. Heart failure (HF) in the community is assumed to result in one third the fatality as HF in hospital and hypertension life expectancy is 80% that of US general population MediCare [4] May overestimate the life years lost due to deaths caused by HF in the community and hypertension. May overestimate the impact of treatments preventing and postponing deaths due to HF in the community and hypertension therefore making the estimated relative contribution of risk factors conservative.
Risk Factors CHD patients. Median survival was assumed to be that of the post-AMI patients MediCare [4] May underestimate the benefit of reductions in risk factor prevalence Provides a conservative estimate
  Healthy population. survival was assumed to be that of general US population US Bureau of the Census [17] May both underestimate and overestimate the benefit or harms in terms of survival for each of the changes in risk factor prevalence Avoids methodological issues of non-additivity and double-counting
  Asymptomatic CHD. Survival was estimated as half-way between that of post-AMI patient and General US life expectancy MediCare and US Bureau of the Census [4, 17] Arbitrary and may underestimate the benefit of reductions in risk factor prevalence Provides a conservative estimate