Choosing algorithms for TB screening: a modelling study to compare yield, predictive value and diagnostic burden

Abstract

Background: To inform the choice of an appropriate screening and diagnostic algorithm for tuberculosis (TB) screening initiatives in different epidemiological settings, we compare algorithms composed of currently available methods.

Methods: Of twelve algorithms composed of screening for symptoms (prolonged cough or any TB symptom) and/or chest radiography abnormalities, and either sputum-smear microscopy (SSM) or Xpert MTB/RIF (XP) as confirmatory test we model algorithm outcomes and summarize the yield, number needed to screen (NNS) and positive predictive value (PPV) for different levels of TB prevalence.

Results: Screening for prolonged cough has low yield, 22% if confirmatory testing is by SSM and 32% if XP, and a high NNS, exceeding 1000 if TB prevalence is ≤0.5%. Due to low specificity the PPV of screening for any TB symptom followed by SSM is less than 50%, even if TB prevalence is 2%. CXR screening for TB abnormalities followed by XP has the highest case detection (87%) and lowest NNS, but is resource intensive. CXR as a second screen for symptom screen positives improves efficiency.

Conclusions: The ideal algorithm does not exist. The choice will be setting specific, for which this study provides guidance. Generally an algorithm composed of CXR screening followed by confirmatory testing with XP can achieve the lowest NNS and highest PPV, and is the least amenable to setting-specific variation. However resource requirements for tests and equipment may be prohibitive in some settings and a reason to opt for symptom screening and SSM. To better inform disease control programs we need empirical data to confirm the modeled yield, cost-effectiveness studies, transmission models and a better screening test.

Figure 1

Algorithms composed of one or more screening methods and one or more confirmatory tests. In panel A one screening tool is applied (e.g. symptoms) and screen positives are further evaluated by one confirmatory test with high sensitivity and high specificity (e.g. Xpert MTB/RIF). In panel B one screening tool is applied (e.g. symptoms) and screen positives are further evaluated by a confirmatory test with low sensitivity (e.g. sputum smear microscopy), and persons with a negative test receive a second test or procedure (e.g. clinical diagnosis, or sputum culture). In panel C two screening tools are applied (e.g. symptoms and chest radiography) and screen positives on either one or on both are further evaluated with a confirmatory test. In panel D two screening tools are applied sequentially. Screen positives on the first screen (e.g. symptoms) undergo a second screen (e.g. CXR) and if also positive on the second a confirmatory test is applied. The single confirmatory test in panels C and D could also be replaced by two-steps as in panel B.

Figure 2

Outcomes of an example screening and diagnostic algorithm. Modified after Lonnroth IJTLD 2013. TB = tuberculosis, +ve = positive, -ve = negative.

Figure 3

TB cases detection and requirements for screening chest X-rays and confirmatory tests of each algorithm, assuming 1% TB prevalence among the screened population. CXR = chest X-ray for screening; SSM = sputum smear microscopy; XP = Xpert MTB/RIF; TP = true positive; 1 = first screen; 2 = second screen if the first screen is positive.

Figure 4

Number needed to screen to find one true case of active TB and positive predictive value of each algorithm at different levels of TB prevalence. Panel A: Number needed to screen (NNS) to find one true positive (TP) case; Panel B: Positive predictive value (PPV). CXR = chest X-ray for screening; SSM = sputum smear microscopy; XP = Xpert MTB/RIF; 1 = first screen; 2 = second screen if first is positive.

Figure 5

Effect of uncertainty in the accuracy of screening and diagnostic tests and assumptions about clinical diagnosis on the NNS and PPV, assuming 1% TB prevalence in the screened population. Panel A: Variation in the number needed to screen (NNS); Panel B: Variation in the positive predictive value (PPV). The symbols represent the point estimates and the vertical bars the range due to uncertainty in the model parameter, as specified in Table 2. The specific scenarios are listed in Table 3. CXR = chest X-ray for screening; SSM = sputum smear microscopy; XP = Xpert MTB/RIF; 1 = first screen; 2 = second screen if first is positive. SSA = sub Saharan Africa.