Understanding the impact of disease
control: TB epidemiology and the
GFATM
Questions
Routine
OR
Surveys
Research
Size of
the
problem
Notifications
Cure rate
Deaths
Global Report
Private sect.
Vital Reg.
Prevalence
Incidence
Active CF
S. Africa
Review
S. Korea
Notifications
Tr. outcomes
Morocco
Sent. sites
Prevalence
Incidence
S. Korea
Modelling
HIV; MDR;
Diagnostics
Access
Nairobi
Risk factors
Social issues
PPM
India
HIV
MDR
Duration dis.
S. Africa
S. Africa
Notifications
Cure rate
Deaths
Peru
Diagnosis
Drug supply
Qual. control
Czech.
Prevalence
Incidence
DOTS
IPT
India
Czech.
Direction
of change
Reasons
for change
Impact of
Control
Chennai
World
Size/Routine
Highest TB rates per capita are in Africa
per 100 000 population
< 10
10 to 24
25 to 49
50 to 99
100 to 299
300 or more
No Estimate
The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the World
Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or
boundaries. Dotted lines on maps represent approximate border lines for which there may not yet be full agreement.
© WHO 2002
Size/OR
TB deaths = Incidence  Case fatality rate
Table 1. Estimating case detection rates from vital registration data. Red:
Statistics South Africa. Blue: TB programme. Errors are fractional errors.
TB deaths (k)
All deaths (k)
Unknown causes (k)
Deaths due to HIV
Completeness of reporting (%)
Proportion of HIV deaths due to TB
TB deaths (k)
Proportion of cases HIV positive (%)
CFR negative (%)
CFR positive (%)
CFR (%)
Incidence (k)
Notifications (k)
CDR (%)
2001
Error (%)
50.9
451.9
56.2
9.2
90.0
15.0
66.1
59.0
16.0
35.0
37.0
242.9
192.9
80.0
20
20
5
50
10
50
50
11.0
Ghana
Kenya
Gambia
Liberia
Malawi
Mauritius
Mozambique
Nigeria
Botswana
South Africa
Sierra Leone
Tanzania
Ethiopia
Uganda
Zambia
Brazil
Columbia
Afghanistan
Cyprus
Iraq
Pakistan
Somalia
Tunisia
Syria
Turkey
Netherlands
Bangladesh
India
Indonesia
Malaysia
Myanmar
Nepal
Sri Lanka
Thailand
Brunei
Cambodia
China
Philippines
R. Korea
Japan
Samoa
Viet Nam
Size/Surveys
Prevalence surveys: National blue; sub-national red.
Kolin
Size/Research 1
SS+ TB incidence/100k
300
200
South Korea
ARTI50
100
Civil servants
0
1960
1970
1980
1990
Size/Research 2
Calculating sample sizes
1,000,000
Precision (%)
1
Sample size
100,000
2
4
8
10,000
16
32
1,000
64
100
10
0.01
0.1
1
Percent positive
10
100
Direction/Routine
Morocco:
PTB incidence projected to 2015
70
Incidence rate/100K
60
50
on aging population
40
30
on 1994 age-structure
20
10
0
1980
Decline in TB in Morocco:
trend1990
in age1995
structure
of
1985
2000
2005cases
2010
2015
0.99
%cases >15 yrs
0.98
0.97
0.96
0.95
0.94
1980
1985
1990
1995
2000
2005
Direction/OR
DOTS reduces prevalence of culture+
TB by 37% in less than a decade in China
Prevalence culture+ TB/100,000
250
DOTS
Other
200
150
100
50
0
1990
2000
Incidence or death rate/10K/year
Direction/Surveys
Decline in TB, Alaska 1950-73
6
cases 13%/yr
t1/2 = 5 yr
5
4
3
2
1
0
-1
-2
1950
deaths 30%/yr
t1/2 = 2.3 yr
1955
1960
source: Grzybowski Tubercle 1976
1965
1970
1975
Direction/Research
TB: elimination by 2050?
1600
Incidence/million/yr
1400
1200
1000
800
600
Projected incidence
100x bigger than
elimination
threshold in 2050
GP2: incidence falls
5-6%/yr 2010-2015
400
200
0
1990
2000
2010
2020
2030
2040
2050
Vaccines, drugs and risk factors?
Reasons/Routine
Nairobi
Reasons/OR
Smoking and TB in Chennai
27k deaths and 16k controls, 1994-1997. 2k TB deaths
Smoker
Non-smoker
Odds
TB deaths
1454
386
3.76
Controls
6430
10058
0.64
OR = 5.9 (4.5)
F = 0.79
RR -1
PAF =
× F = 0.66 (0.61)
RR
60% of all TB deaths among men in Chennai are
attributable to smoking
Gajalakshmi, V., Peto, R., et al. Smoking and mortality from tuberculosis and other diseases in India:
retrospective study of 43000 adult male deaths and 35000 controls Lancet (2003) 362 507–515.
Reasons/Surveys
Drug Resistance in Retreatment Patients
(n = 1 508)
SA rates weighted by Province
Mpumalanga
Kwazulu-Natal
Eastern Cape
North West
Limpopo
Gauteng
Western Cape
Free State
South Africa
0
5
10
MDR
15
Any R resistance
20
25
Any H resistance
30
35
Reasons/Research 1
Annual incidence (%) .
10
HIV-
9.4
HIV+
8
5.9
6
4
2
2.2
1.0
1.1
1.1
0
1991-1994
1995-1997
1998-1999
TB incidence among gold miners in SA
DDR
Reasons/Research 2
Smear positive disease in
South African gold miners
Incidence (%/yr)
Prevalence (%)
Dis.Duration (yr)
HIVHIV+
0.48
2.87
0.55
0.44
1.15
0.15
Ratio
6.01
0.80
0.13
Corbett et al. 2003
Impact/Routine
Pulmonary TB cases/100,000
Dynamics of pulmonary TB in
Peru 1980-2000
220
DOTS 1990
200
case finding
180
160
140
120
100
1980
PTB falling at 6%/yr
1985
1990
1995
2000
Impact/OR
"Model DOTS Project" reduces TB prevalence in
south India
Prevalence/100K
2000
source: TRC Chennai
fall ~10%/yr in
MDP
1000
Male C+
Male S+
Female C+
Female S+
0
68- 71- 73- 76- 79- 81- 8470 73 75 78 81 83 86
Year
99- 0101 03
Impact/Surveys
South Korea
900
Prevalence SS+ TB/100k
800
700
600
500
400
300
200
-0.0797t
P = ae
100
0
1960
1965
1970
1975
1980
1985
1990
1995
Impact/Research
Decline in prevalence
Kolin, Czechoslovakia
Men SS+
Women SS+
Men SSWomen SS-
100
Cases
Men: 20%  10%/yr
50
Women: 26%  21%
0
1960
1961
1962
Year
1963
1964
China
The last word…
We must eradicate tuberculosis, and we must
do it now, … All available resources must be
used. Chemotherapy, computers, prophylaxis
and prevention, case finding and kindness can
be blended in a properly constructed
epidemiological model which will tell us exactly
where we are going and how fast.
Davies, J.C.A. The Eradication of Tuberculosis in Rhodesia DPH, London School of
Hygiene and Tropical Medicine, 1966
1.
2.
3.
4.
5.
6.
Estimating CDR
Assume that the 1997 estimate of CDR is
correct so that we know the incidence in 1997.
Assume that the trend in total notifications (all
forms; DOTS non-DOTS) gives the trend in
incidence.
Use this trend to work out the incidence of all
forms in the year 2004.
Assume that the SS+ incidence is 45% of the
all-forms incidence for HIV- people; 35% for
HIV+ people.
Calculate the SS+ incidence
Divide the notification rate in DOTS areas by
the SS+ incidence to get the SS+ DOTS case
detection rate.
The Design Effect
Suppose we have k clusters with m people in each cluster so that the total sample size is n = km.
D = 1 + (m  1)ρ
where ρ, the intra-class (or intra-cluster) correlation coefficient, is the ratio of the between-cluster
variance sb to the total variance so that
sb2
r  2
sb  sw2
where sw is the within-cluster variance.
1. All members of a cluster are identical, sw = 0, r = 1 and D = m.
Effective sample size = n/m = k = the number of clusters.
2. Members of each cluster have no particular similarity sb = 0, r = 0, and D = 1.
Effective sample size = n = number of people.
For example, if each cluster contains
m = 10 people and ρ =0.1 then 1 + (m-1)ρ = 1.9.
m = 20 people and ρ = 0.2 then 1 + (m – 1)ρ = 4.8
Even a small value of the intra-cluster correlation coefficient, multiplied by the size of the cluster,
could lead to a substantial increase in variance and reduction in effective sample size.