Clare Walker

Accurate diagnosis of latent tuberculosis in children, people who are immunocompromised or at risk from immunosuppression and recent arrivals from countries with a high incidence of tuberculosis: systematic review and economic evaluation

BACKGROUND Tuberculosis (TB), caused by Mycobacterium tuberculosis (MTB) [(Zopf 1883) Lehmann and Neumann 1896], is a major cause of morbidity and mortality. Nearly one-third of the worlds population is infected with MTB; TB has an annual incidence of 9 million new cases and each year causes 2 million deaths worldwide. OBJECTIVES To investigate the clinical effectiveness and cost-effectiveness of screening tests [interferon-gamma release assays (IGRAs) and tuberculin skin tests (TSTs)] in latent tuberculosis infection (LTBI) diagnosis to support National Institute for Health and Care Excellence (NICE) guideline development for three population groups: children, immunocompromised people and those who have recently arrived in the UK from high-incidence countries. All of these groups are at higher risk of progression from LTBI to active TB. DATA SOURCES Electronic databases including MEDLINE, EMBASE, The Cochrane Library and Current Controlled Trials were searched from December 2009 up to December 2014. REVIEW METHODS English-language studies evaluating the comparative effectiveness of commercially available tests used for identifying LTBI in children, immunocompromised people and recent arrivals to the UK were eligible. Interventions were IGRAs [QuantiFERON(®)-TB Gold (QFT-G), QuantiFERON(®)-TB Gold-In-Tube (QFT-GIT) (Cellestis/Qiagen, Carnegie, VA, Australia) and T-SPOT.TB (Oxford Immunotec, Abingdon, UK)]. The comparator was TST 5 mm or 10 mm alone or with an IGRA. Two independent reviewers screened all identified records and undertook a quality assessment and data synthesis. A de novo model, structured in two stages, was developed to compare the cost-effectiveness of diagnostic strategies. RESULTS In total, 6687 records were screened, of which 53 unique studies were included (a further 37 studies were identified from a previous NICE guideline). The majority of the included studies compared the strength of association for the QFT-GIT/G IGRA with the TST (5 mm or 10 mm) in relation to the incidence of active TB or previous TB exposure. Ten studies reported evidence on decision-analytic models to determine the cost-effectiveness of IGRAs compared with the TST for LTBI diagnosis. In children, TST (≥ 5 mm) negative followed by QFT-GIT was the most cost-effective strategy, with an incremental cost-effectiveness ratio (ICER) of £18,900 per quality-adjusted life-year (QALY) gained. In immunocompromised people, QFT-GIT negative followed by the TST (≥ 5 mm) was the most cost-effective strategy, with an ICER of approximately £18,700 per QALY gained. In those recently arrived from high TB incidence countries, the TST (≥ 5 mm) alone was less costly and more effective than TST (≥ 5 mm) positive followed by QFT-GIT or T-SPOT.TB or QFT-GIT alone. LIMITATIONS The limitations and scarcity of the evidence, variation in the exposure-based definitions of LTBI and heterogeneity in IGRA performance relative to TST limit the applicability of the review findings. CONCLUSIONS Given the current evidence, TST (≥ 5 mm) negative followed by QFT-GIT for children, QFT-GIT negative followed by TST (≥ 5 mm) for the immunocompromised population and TST (≥ 5 mm) for recent arrivals were the most cost-effective strategies for diagnosing LTBI that progresses to active TB. These results should be interpreted with caution given the limitations identified. The evidence available is limited and more high-quality research in this area is needed including studies on the inconsistent performance of tests in high-compared with low-incidence TB settings; the prospective assessment of progression to active TB for those at high risk; the relative benefits of two-compared with one-step testing with different tests; and improved classification of people at high and low risk for LTBI. STUDY REGISTRATION This study is registered as PROSPERO CRD42014009033. FUNDING The National Institute for Health Research Health Technology Assessment programme.

General practitioners providing non-urgent care in emergency department: a natural experiment

Objective To examine whether care provided by general practitioners (GPs) to non-urgent patients in the emergency department differs significantly from care provided by usual accident and emergency (A&E) staff in terms of process outcomes and A&E clinical quality indicators. Design Propensity score matched cohort study. Setting GPs in A&E colocated within the University Hospitals Coventry and Warwickshire NHS Trust between May 2015 and March 2016. Participants Non-urgent attendances visits to the A&E department. Main outcomes Process outcomes (any investigation, any blood investigation, any radiological investigation, any intervention, admission and referrals) and A&E clinical indicators (spent 4 hours plus, left without being seen and 7-day reattendance). Results A total of 5426 patients seen by GPs in A&E were matched with 10 852 patients seen by emergency physicians (ratio 1:2). Compared with standard care in A&E, GPs in A&E significantly: admitted fewer patients (risk ratio (RR) 0.28, 95% CI 0.25 to 0.31), referred fewer patients to other specialists (RR 0.31, 95% CI 0.24 to 0.40), ordered fewer radiological investigations (RR 0.38, 95% CI 0.34 to 0.42), ordered fewer blood tests (0.57, 95% CI 0.52 to 0.61) and ordered fewer investigations (0.93, 95% CI 0.90 to 0.96). However, they intervened more, offered more primary care follow-up (RR 1.78, 95% CI 1.67 to 1.89) and referred more patients to outpatient and other A&E clinics (RR 2.29, 95% CI 2.10 to 2.49). Patients seen by GPs in A&E were on average less likely to spend 4 hours plus in A&E (RR 0.37, 95% CI 0.30 to 0.45) compared with standard care in A&E. There was no difference in reattendance after 7 days (RR 0.96, 95% CI 0.84 to 1.09). Conclusion GPs in A&E tended to manage self-reporting minor cases with fewer resources than standard care in A&E, without increasing reattendance rates.