Summary Measures of Population Health
ration and a disability weight. The burden from short duration conditions for which disability weights were derived from annual profiles was cal- culated by multiplying the number of incident cases by a one-year dura- tion and the disability weight. In the example of the one week with gonnorrhoea in an annual profile of a year this means that for each case the duration of disability was taken as 52 weeks and coupled with the dis- ability weight for the annual profile. In personal communication, two of the authors of the Dutch study have stated that it is not valid to “back- calculate” the values of the constituting elements of a profile from the profile value. However, from the way these weights have been applied in the Dutch burden of disease study, I can only conclude that an annual profile weight for one week with gonnorrhoea must be the equivalent of 1/52 of the weight implied for the one week of disability. The only rea- son not to do so would be to argue that the 51 weeks following an epi- sode of one week of gonnorrhoea are different from 51 weeks of healthy life. With some creativity one could argue that a bout of gonnorrhoea may cause anxiety about marital strife or the risk of having contracted other sexually transmitted diseases leading to loss of health-related quality of life beyond the one week of symptoms. If this is considered an important “health state”, I would argue that it ought to be valued separately to be matched with epidemiological data on how frequently it occurs. It is a lot harder to imagine that a bout of nasopharyngitis would cause disability beyond the symptomatic period. Yet, each incident case of common cold in a year was multiplied with the annual profile’s weight. If the Dutch annualized weights for common cold, sinusitis and tonsillitis had been used in the Victorian Burden of Disease Study the contribution of upper respi- ratory tract infections to Years Lived with Disability (YLD) would have increased six-fold and raised the ranking from 54 to 12. Similarly, lower respiratory tract infections (pneumonia and influenza) would increase eight-fold and change in YLD ranking from position 45 to 9. This indi- cates that the adoption of annualized weights can have more than trivial consequences on burden of disease results.
Let us examine the Dutch weights for annual profiles after “back-cal- culation” to a value for the implied severity for the duration of the short illness. This involves a simple transformation of the weights to the com- mon GBD notation where 0 indicates full health and 1 complete loss of health and then multiplication by 52 divided by the duration in weeks. Figure 1 shows the implied weights for the short duration of these condi- tion and illustrates how the annual profiles have failed to produce plau- sible weights. First, the weights for pneumonia and cystitis are greater than one and thus imply that these conditions are worse than being dead. Sec- ond, it is hard to imagine that the members of the valuation panels would have valued, for example, the loss of health due to urethritis as worse than that of paraplegia if they had realized this is the consequence of the value of their annual profile. In general, each of the annual profile disability