Cost-effectiveness of tuberculosis infection screening at first reception into English prisons: a model-based analysis

Abstract

Background: The World Health Organization recommends systematic screening for tuberculosis in incarcerated populations, which are consistently at high risk of tuberculosis relative to the general population. In England, new receptions into prisons do not receive screening for tuberculosis infection, and evidence from economic evaluations is lacking.

Methods: We performed a cost-effectiveness analysis of introducing systematic screening for tuberculosis infection at first reception into English prisons from a health systems perspective. We used a tuberculosis transmission model calibrated to public data on prison populations and flows. We developed decision tree models of prison-specific tuberculosis care pathways and their costs, informed by stakeholders and pilot studies. Sensitivity analyses included eliminating loss to follow-up (LTFU) in care cascades, zeroing extramural escort costs, and targeting screening to those born in countries with higher tuberculosis incidence (over 40 per 100,000 per year).

Findings: In our base case analysis, the intervention had an incremental cost-effectiveness ratio (ICER) of £78,000 per quality-adjusted life-year (QALY) gained. Reducing LTFU and avoiding prison escort costs would substantially improve cost-effectiveness, to ICERs of £70,000 and £54,000 per QALY gained, respectively. Targeting those born in higher incidence countries was predicted to be cost-saving.

Interpretation: Universal tuberculosis screening and preventive treatment for new receptions into English prisons is not cost-effective by the usual threshold of £30,000. However, targeting high-risk groups could be cost-saving. Tuberculosis interventions should explore ways to reduce LTFU and extramural healthcare in order to meet the needs of those incarcerated while minimizing costs.

Funding: UKHSA from 9/2023 to 12/2024.

Keywords: Cost-utility analysis; Economic evaluation; Inclusion health; Mathematical model; Persons deprived of liberty; TB; TPT; Tuberculosis preventive therapy.

Fig. 1

Transmission model structure. A: Detention states and flows between them. The intervention is situated at the flow into remand (blue). B: The tuberculosis (TB) states and transitions, with infectious states in red. States under the blue arrow experience worse health-related quality of life (HRQoL) and increased risk of tuberculosis or death. C: States representing tuberculosis preventive therapy (TPT). The full compartmental model is the product of these three domains, comprising 5 × 8 × 3 = 120 ordinary differential equations.

Fig. 2

Intervention cascades. ATT, anti-tuberculosis treatment (for disease); TPT, tuberculosis preventive treatment; TB, tuberculosis; GP, general practitioner; NHS, National Health Service; IGRA, interferon gamma release assay (for tuberculosis infection).

Fig. 3

Monetary net benefit for targeted screening approaches. For each pre-screen sensitivity and specificity midpoint, yellow text shows the corresponding tuberculosis infection (TBI) prevalence and proportional size of the group screened-in for further assessment. Fill colour shows the mean expected net monetary benefit for each tile, with red colours corresponding to positive net benefit (i.e. cost-effective at a threshold of £30,000 per quality-adjusted life-year gained); blue colours correspond to negative net monetary benefit. Net monetary benefit does not indicate whether an intervention is cost-saving. Magenta numbers ①, ② locate the likely location of exemplar target groups defined in the text.