Keith Edwards Scoring System Effectively Diagnoses Childhood Tuberculosis

The Keith Edwards scoring system has high sensitivity and specificity for diagnosing childhood tuberculosis (TB), according to study findings published in Open Forum Infectious Diseases.

Researchers conducted a systematic review and meta-analysis to evaluate the accuracy of childhood TB diagnostic scoring systems and algorithms. They searched the PubMed, CINAHL, Embase, Scopus, and Google Scholar databases for original relevant studies through March 30, 2023.

Eligible studies were published in English and assessed the diagnostic accuracy of childhood TB scoring systems for their sensitivity and specificity compared with a reference standard and had enough information to determine true positive, true negative, false positive, and false negative values. Participants were younger than 15 years of age with presumptive TB and had completed a diagnostic TB scoring test or diagnostic algorithm.

The review included 15 studies with 7327 study participants. A total of 10 diagnostic scoring systems were considered in the subgroup analysis. The Keith Edwards score was used most frequently (5 studies), followed by the Kenneth Jones Criteria and Ministry of Health (MoH)-Brazil score (3 studies each).

Among the 15 studies evaluated in the patient selection domain, 9 (60%) had a low risk of bias, 5 (33.3%) had a high risk, and 1 (0.1%) had an unknown risk. For the index test, the conduct or interpretation was regarded as low risk for 12 (80.0%) studies, with the other 3 studies having an unknown risk of bias.

The scoring systems had a range of sensitivities of 93.35% to 68.18%, and the specificities ranged from 99.40% to 35.71%. The Keith Edwards scoring system had a pooled sensitivity of 81.9% (95% CI, 70.7%-89.5%; I² = 85.73%; P <.001) and a pooled specificity of 81.20% (95% CI, 55.30%-93.80%; I²= 98.92%; P <.001).

Other scores or algorithms with multiple studies in the review showed a pooled sensitivity of 80.1% (95% CI, 63.4%-90.3%; I²= 83.14%; P <.001) and 79.9% (95% CI, 71.70%-86.20%; I²= 0.00%; P <.406) for the Kenneth Jones criteria and MoH-Brazil scoring system, respectively. The pooled specificity was 45.70% (95% CI, 35.60%-56.10%; I² = 0.00; P <.875) for the Kenneth Jones criteria and 73.2% (95% CI, 67.3%-78.5%; I² = 96.42; P <.001) for the MoH-Brazil scoring system.

The researchers noted that most of the scoring systems or algorithms used chest radiography or tuberculin skin test and all included bacteriologic confirmation, which biased their diagnostic accuracy. Also, the analysis could have missed some relevant studies.

“We recommend the Keith Edwards scoring system because it has high sensitivity and specificity and can easily be used at primary health care or lower health facilities,” the study authors concluded.

This article originally appeared on Pulmonology Advisor