Reader Comments

Post a new comment on this article

Survival above age 55?

Posted by ericross11 on 02 Apr 2014 at 21:14 GMT

When producing life expectancy estimates, were mortality rates in the group above age 55 adjusted in any way? Or were the observed mortality rates assumed to apply to everyone with age >55?

As an example, Hogg et al. (Lancet 2008) used risk ratios applied to the general population mortality rates to derive estimates for mortality above age 65, since they had limited data on individuals with HIV at advanced age. So while they observed a mortality rate of 3.3/100PY, they used a mortality rate of 8.9/100PY in calculating life expectancy estimates (see appendix 2).

No competing interests declared.

RE: Survival above age 55?

ericross11 replied to ericross11 on 22 May 2014 at 15:15 GMT

To give a more specific example from this paper:

Figure 1 shows an observed mortality rate of around 27 per 1000 person-years in the oldest age group (55+). If this mortality rate were applied without adjustment, it would suggest a life expectancy from age 55 of 37.0 years (1/0.027).

The CDC estimated life expectancy from age 55 in the general US population in 2008 to be 26.7 years ( So if the observed mortality rates in the oldest age group from this paper were applied without adjustment, then the implication would be that persons living with HIV can expect to live 10 years longer than the general population from age 55 (37.0 years vs. 26.7 years). This seems pretty unlikely! And an excessively long life expectancy from age 55 would contribute to overestimating life expectancy from age 20 in persons with HIV.

So to interpret the results of this paper, it'd be helpful to know if the observed mortality rates in the oldest age group were applied directly or adjusted when making life expectancy calculations. I can't find any details about this in the methods or supplementary materials.

No competing interests declared.