Objective of model

The objective of the Economics of TB Screening model is to allow users to simulate costs and yields of two screening algorithms and use this information to select the diagnostic pathway with superior impact/cost trade-off.

Scenarios

In the online model cost and yield calculations can be customised by the user for two algorithms.

Scenario 1 Scenario 2
dCXR/CAD as a rapid triage before Xpert MTB/RIF testing All tested directly on Xpert MTB/RIF

Assumptions and estimates are based on recent publications and market intelligence, others are default values for user customization.
In Scenario 1 dCXR/CAD is used to screen risk group members (symptomatic and asymptomatic) and to detect the presumptive TB cases on the basis of the CAD4TB abnormality score. It is assumed that only individuals having a CAD score above the abnormality threshold (e.g. >60 the optimal sensitivity and specificity point on the ROC curve) will be eligible for Xpert MTB/RIF testing.

  • CXR/CAD presumptive TB rate calculation: TP + FP = prevalence * (sensitivity) + (1-prevalence) * (1-specifity).
  • Total number of TB cases detected calculation: total number of people screened per year * TB prevalence rate * sensitivity CAD * sensitivity Xpert MTB/RIF.

Sensitivity and specificity values of CAD and Xpert MTB/RIF can vary with the HIV prevalence rate (percentage of smear negative PTB) in the risk group, which is automatically reflected in the model along the below estimated values:

Assumptions accuracy for HIV prevalence variations Low 0-4% Medium 4-12% High >12%
Sensitivity / specificity CAD [1] 94% / 84% 90% / 75% 85% / 75%
Sensitivity / specificity Xpert [2] 95% / 99% 90% / 99% 86% / 99%

In Scenario 2 no screening with dCXR/CAD is done so all risk group members are diagnosed using Xpert MTB/RIF.


Input

Users can customize the below default variables before pressing [calculate] to reflect the TB program settings. Some values such as depreciation period, maintenance cost and Xpert cartridge cost are fixed values in the model.

Variables to customize

  • TB prevalence % in risk group
  • Actual % of cases notified in risk group
  • Working days per year
  • HIV Prevalence % in risk group (low, medium or high)
  • Digital X-ray cost (stationary system vs. portable)
  • CAD cost per screen (cost depends on volume of the bundle)
  • Annual Salaries Clinician (X-ray) X-ray / Xpert Technician
  • Throughput: people screened per day

Output

As output the model provides the user with the following information for both algorithms.

Economics

  • Total screening cost (per year)
  • Screening cost (per person screened)
  • Screening cost (per case notified)
  • Cases notified (per € 100,000 budget)

Yield

  • Future % of cases notified
  • Future number of cases notified (per year)
  • Net increase of cases notified
  • Number Needed to Screen