Global Tuberculosis Control 2010 - Florida Department of Health

the estimated number of incident cases that occurred in
the same year.
Estimates of the incidence of sputum smear-positive
pulmonary TB and the related CDR for these cases are
not published in this report. There are several reasons
for this, which are summarized in the main part of the
report ( ). Among the reasons noted in this box
are the findings and associated recommendations from
the 18-month expert review of methods used to estimate
disease burden described in above, and
associated updates to methods used to estimate disease
burden that have been applied in a series of regional
workshops and country missions (). A fuller
explanation is provided here.
In regional workshops and country missions, the
starting point for estimates of the incidence of TB was
an estimate of the CDR for all forms of TB. This was
because, for many countries, estimates of the incidence
of all forms of TB are used to produce indirect estimates
of TB mortality and TB prevalence (these do not require
estimates of the CDR for smear-positive TB).
The incidence of sputum smear-positive pulmonary
TB can theoretically be estimated by multiplying the
incidence of all forms of TB by the estimated proportion
of all incident cases of TB that have sputum smear-positive
pulmonary TB. The CDR for sputum smear-positive
pulmonary TB would then be estimated by dividing
reported notifications of sputum smear-positive cases
of pulmonary TB by the estimated incidence of sputum
smear-positive pulmonary TB. This approach was used
in global reports published up to December 2009. Subsequently,
the findings of one of the systematic reviews
conducted by the expert group have cast doubts on this
approach. A systematic review of the evidence about the
proportion of TB cases that have sputum smear-positive
pulmonary TB found considerable uncertainty around
the best estimate of this proportion. 1
It may appear that an alternative method is to assume
that, for any given country, the proportion of all incident
cases with sputum smear-positive pulmonary TB
is equal to the proportion of all notified cases that had
smear-positive pulmonary TB in the region of reference
(this could be an epidemiologically-defined region or a
WHO region, for example). However, use of this method
will mean that, for many countries, changes in the estimated
CDR for sputum smear-positive pulmonary TB
will occur from one year to the next simply because of
a change in the share of regional notifications that are
smear-positive, without any real change in the level of
case detection in the country itself (since a change in the
regional proportion will cause the estimated incidence
of smear-positive TB in a given country, the denominator
of the CDR, to go up and down even when the country’s
notifications are stable). This is wrong and likely to
cause confusion.
Another alternative is to assume that, in any given
country, the proportion of sputum smear-positive cases
of pulmonary TB among all notified cases reflects the
true proportion of sputum smear-positive TB among
incident cases of TB in the country. In this case, the estimated
CDR for all forms of TB and the estimated CDR
for sputum smear-positive cases are the same.
If needed for reporting purposes, it is suggested that
the CDR for smear-positive TB is assumed to be similar
to the CDR for all forms of TB.
The prevalence of HIV among incident cases of TB was
directly estimated from country-specific and empirical
data wherever possible. For the estimates published in
this report, suitable data (as defined in ) were
available for a total of 440 country-year data points, up
from 288 country-year data points in the previous year
(for details of the source of data used for each country,
please see the ).
For the 3785 country-year data points for which surveillance
data were either not available or for which the percentage
of TB patients tested for HIV was below 50%,
the prevalence of HIV was estimated indirectly according
to the following equation:
t =
hρ
l + h(ρ – l)
In this equation, t is HIV prevalence among incident TB
cases, h is HIV prevalence among the general population
(from the latest time-series provided by UNAIDS) and ρ
is the incidence rate ratio (IRR) (defined as the incidence
rate of TB in HIV-positive people divided by the incidence
rate of TB in HIV-negative people). 2 To estimate
ρ from empirical data (that is, from countries that have
an empirical estimate of both t and h), the equation was
rearranged as follows:
ρ =
t(l – h)
h(l – t)
We then let logit(t) be log(t/(1-t)) and logit(h) be log(h/
(1-h)). Using data from countries where HIV prevalence
1
Research Institute of Tuberculosis, Tokyo 2010 (data not shown).
2
Data on HIV prevalence in the general population are unpublished
data provided to WHO by UNAIDS.