Immigrant screening for latent tuberculosis in Norway: a cost-effectiveness analysis

Abstract

The incidence of tuberculosis (TB) disease has increased in Norway since the mid-1990s. Immigrants are screened, and some are treated, for latent TB infection (LTBI) to prevent TB disease (reactivation). In this study, we estimated the costs of both treating and screening for LTBI and TB disease, which has not been done previously in Norway. We developed a model to indicate the cost-effectiveness of four different screening algorithms for LTBI using avoided TB disease cases as the health outcome. Further, we calculated the expected value of perfect information (EVPI), and indicated areas of LTBI screening that could be changed to improve cost-effectiveness. The costs of treating LTBI and TB disease were estimated to be €1938 and €15,489 per case, respectively. The model evaluates four algorithms, and suggests three cost-effective algorithms depending on the cost-effectiveness threshold. Screening all immigrants with interferon-gamma release assays (IGRA) requires the highest threshold (€28,400), followed by the algorithms "IGRA on immigrants with risk factors" and "no LTBI screening." EVPI is approximately €5 per screened immigrant. The costs for a cohort of 20,000 immigrants followed through 10 years range from €12.2 million for the algorithm "screening and treatment for TB disease but no LTBI screening," to €14 million for "screening all immigrants for both TB disease and LTBI with IGRA." The results suggest that the cost of TB disease screening and treatment is the largest contributor to total costs, while LTBI screening and treatment costs are relatively small. Increasing the proportion of IGRA-positive immigrants who are treated decreases the costs per avoided case substantially.

Assessing the value of screening tools: reviewing the challenges and opportunities of cost-effectiveness analysis

“…Masking 4. Risk 1 + masking No screening Women eligible for a National Breast Screening Program (NBSP) Lifetime NHS 4% 2014 GBP QALYs Yes Risk stratified NBSPs were relatively CE compared to the UK NBSP FP7-HEALTH-2012-INNOVATION-1 Yes Gupta et al [ 28 ] USA Cystic lung disease High-resolution computed tomographic (HRCT) imaging no HRCT screening Patients with Spontaneous Pneumothorax NA Societal 3% 2014 USD QALYs Yes HRCT image screening is CE None Yes Haukaas et al [ 44 ] Norway Tuberculosis (TB) 1. TST + IGRA 2.…”

“…Chowers et al tested different accuracy rates for HIV diagnostic tests with sensitivity analyses [ 21 ]. Other studies assumed specific accuracy estimates (usually 100%) and acknowledged the limitations, such as potentially overestimating cost-effectiveness estimates by excluding pertinent costs associated to misclassified patients [ 22 , 29 , 44 ].…”

“…According to the WHO, screening interventions are expected to provide treatment alternatives for those patients with identified cases of disease [ 1 ]. However, 15 studies (22%) failed to model a treatment pathway [ 22 – 24 , 26 , 29 , 31 , 38 , 44 , 57 , 64 , 70 – 74 ]. The main outcomes captured by these studies were the following: cases detected, missed cases, avoided cases, and identified true positives and true negatives.…”

“…Some studies reported uncertainty around treatment efficacy inputs [ 15 , 44 , 65 , 66 , 75 ]. Sensitivity and scenario analyses were broadly used to account for this uncertainty.…”

“…Not surprisingly, cost-effectiveness estimates were influenced by treatment efficacy of early treatment and uptake [ 65 , 66 , 75 ]. A few studies conducted a value of information analyses to estimate the value of collecting further information to resolve decision uncertainty [ 18 , 44 , 75 ].…”