Table 1

Key parameters and assumptions
Parameter Definition Base value [range] Remarks
TB prevalence among the general population and high-risk groups
Prev0 TB prevalence among the general population (per 100,000) country specific Estimated prevalence of all TB among general population based on the estimates in Global TB Control 2010, WHO.
PRi and Previ for Group i (i=1,2,…) where Previ= PRi× Prev0 Assumed prevalence of pulmonary TB for each of the high-risk groups. Defined either by prevalence ratio to the general population or by direct estimate PR1 = 1.0 Although not strictly accurate, risk-group prevalence parameters were set to allow general inference regarding some high-risk groups. (Group 1: General population; Group 2: Poor, urban dwellers; Group 3: Malnourished; Group 4: Diabetics; Group 5: TB contacts, and Group 6: Prisoners and/or other high-risk groups.) Note that these groups are only indicative. Users of the tool should define the prevalence parameters based on the best estimate for the local context.
PR2 = 1.5
PR3 = 2.0
PR4 = 3.0
Prev5 = 4000/105
Prev6 = 6000/105
Proportion of abnormalities among screened subjects
Pr (symp) = a × Previ+ b (i=1,2,…) Proportion of TB symptomatics as a function of Previ a = 2.9829 Pr (symp) and Pr (X-ray) were expressed as a linear function of TB prevalence in the target population. The method was employed due to the observed linear trends in prevalence survey findings. Simple linear regression models were fit for multiple data points from prevalence surveys and resulted intercept and coefficient were used. * The overlap was assumed to be 20%, based on prevalence survey findings (References 5, 8). This parameter is required to calculate the combined suspects, i.e. {Pr (symp) ∪ Pr (X-ray)}.
b = 0.0355
Pr (X-ray) = a × Previ+ b (i=1,2,…) Proportion of subjects with X-ray abnormality as a function of Previ a = 3.0415
b = 0.0377
Overlap* {Pr (symp) ∩ Pr (X-ray)} divided by {Pr (symp) + Pr (X-ray)} 0.20
Proportional yields (Refer to Figure 2)
Proportions of prevalent TB cases who are Sensitivity analysis was conducted for the ranges of PYsymp [0.3-0.5] and PYsmear [0.5-0.7], both are assumed to follow a uniform distribution.
PYsymp TB symptomatic 0.40 [0.3-0.5]
PYsmear smear-positive 0.60 [0.5-0.7]
PYss TB symptomatic and smear-positive 0.24 Assuming PYsymp and PYsmear are independent, PYss = PYsymp × PYsmear
PYsc TB symptomatic and smear-negative 0.16 PYsc = PYsymp × (1-PYsmear)
PYxs Not TB symptomatic and smear-positive 0.36 PYxs = (1-PYsymp) × PYsmear
PYxc Not TB symptomatic and smear-negative 0.24 PYxc = (1-PYsymp) × (1-PYsmear)
Dropout rate for screening/diagnostic test
Symptom screening 0.00 A proportion of individuals who entered in a screening step but did not complete the step to receiving the result. The model assumed a high return rate for culture due to a long turnaround time for solid culture.
Sputum smear microscopy 0.02
Chest X-ray (screening) 0.02
Chest X-ray (diagnosis) 0.20 Similarly, the rate is high if the chest X-ray is used for diagnosis of smear-negative TB as it often requires some lead time for group diagnosis (e.g. TB diagnostic committees) as per national guidelines.
Sputum culture (solid) 0.15
Xpert MTB/RIF 0.00
Unit cost for screening/diagnostic test
Symptom screening in USD per person 0.02 The unit cost of screening and diagnostic tests in USD. They were meant to be direct unit cost excluding capital, equipment and human resources cost. These costs can be included in the analysis in the web-based tool if they are deduced to the cost per test.
Sputum smear microscopy in USD per slide (three slides per person) 0.70
Chest X-ray in USD per test 3.00 [2.00-6.00]
Sputum culture (solid) in USD per test 5.00
Xpert MTB/RIF in USD per specimen 16.86

Nishikiori and Van Weezenbeek

Nishikiori and Van Weezenbeek BMC Public Health 2013 13:97   doi:10.1186/1471-2458-13-97

Open Data