Overuse of computed tomography for mild head injury: A systematic review and meta-analysis

Abstract

Background: Computed tomography (CT) scan is a common imaging technique used to evaluate the severity of a head injury. The overuse of diagnostic interventions in the health system is a growing concern worldwide. Objectives: The aim of this systematic review is to investigate the rate of CT scan overuse in cases of mild head injury.

Methods: Eligibility criteria: We encompassed observational studies-either designed as cohort, case-control, or cross-sectional investigations-that reported on CT scan overuse rates for mild head injuries. Studies had to be published in peer-reviewed, English-language sources and provide full content access Information sources: Web of Sciences, Scopus, Medline via PubMed, the Cochrane Library and Embase were searched from inception until April 1, 2023. Studies were included if reporting the overuse of CT scans for mild head injuries using validated criteria. Risk of bias: We used the Risk Of Bias In Non-randomised Studies - of Interventions (ROBINS-I) tool to evaluate the risk bias assessment of included studies. Two independent reviewers evaluated the eligibility of studies, extracted data, and assessed study quality by using the Newcastle-Ottawa Scale. Synthesis of results: Overuse estimates were calculated using a random-effects model. Subgroup analyses were performed to investigate any sources of heterogeneity. Point rate of overuse of CT scans for mild head injuries was the main outcome measured as percentage point estimates with corresponding 95% CIs.

Results: Included studies: Of the 913 potentially relevant studies identified, eight studies were selected for the final analysis. Synthesis of results: The pooled rate of CT scan overuse in patients with mild head injury was found to be 27% [95% CI: 16-43; I2 = 99%]. The rate of CT scan overuse in mild head injury cases varied depending on the criteria used. The rate of CT scan overuse was 37% [95% CI: 32-42; I2 = 0%] with the Glasgow Coma Scale (GCS), 30% [95% CI: 16-49; I2 = 99%] with the Canadian computed tomography head rule, and 10% [95% CI: 8-14; I2 = 0%] with the Pediatric Emergency Care Applied Research Network criterion (PERCAN). Based on subgroup analyses, the rate of CT scan overuse in mild head injury cases was observed to be 30% with the Canadian computed tomography head rule criterion, 43% with the National Institute for Health and Clinical Excellence criterion, and 18% with the New Orleans criterion.

Conclusion: Limitations of evidence: The restricted number of included studies may impact generalizability. High heterogeneity was observed, leading to subgroup analyses based on age, assessment criteria, and study region. Absent data on overuse causes hinders drawing conclusions on contributing factors. Furthermore, this study solely addressed overuse rates, not associated harm or benefits. Interpretation: The overuse of CT scans in mild head injury patients is concerning, as it can result in unnecessary radiation exposure and higher healthcare costs. Clinicians and policymakers should prioritize the implementation of guidelines to reduce unnecessary radiation exposure, healthcare costs, and potential harm to patients.

Trial registration: The study protocol of this review was registered in PROSPERO under the identification code CRD42023416080. https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42023416080.

Fig 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram of the search, screen and selection of studies focusing on the overuse of computed tomography in mild head injury.

Of the 913 potentially relevant studies identified, eight studies were selected for the final analysis.

Fig 2. Summary of risk of bias.

We assessed risk of bias in included studies utilizing the “Risk Of Bias In Non-randomised Studies ‐ of Interventions” (ROBINS-I) tool. In studies, low risk of bias was observed in confounding and participant selection. One study had low bias in intervention classification, while others had uncertain or high risk. Bias in deviations from intended interventions varied, and most studies had low risk. Bias related to missing data and outcome measurement was generally low or unclear. Selection bias in reported results was low in most studies, with one unclear study.

Fig 3. Forest plot showing the pooled rate of overuse of computed tomography in mild head injury based on decision criteria.

Random effects model is used for analysis (95% confidence interval). The overall rate of CT overuse in mild head injury was estimated to be 27% [95% CI: 16–43; I² = 99%].

Fig 4. Assessment of heterogeneity of included studies for overuse of computed tomography in mild head injury using.

Baujat plot. Most studies showed similar overuse rates, but one study (Cellina) had a significantly higher rate. Excluding it, the rate was 22% (95% CI: 15–32; I2 = 98%).

Fig 5. Forest plot of assessment the effect of excluding or including retained studies using sensitivity analysis.

In this analysis, each study is excluded and in its absence, the impact on the rate amount is evaluated. This analysis showed that the results did not change and the of CT overuse in mild head injury was estimated to be 27% [95% CI: 16–43].

Fig 6. Forest plot to assess trends in rate of overuse of computed tomography in mild head injury changes over time using cumulative analysis.

The result of this analysis showed that the rate of CT overuse in mild head injury was 27% [95% CI: 16–43].

Fig 7. Funnel plot for the assessment of publication bias among different studies.

The funnel plot exhibits asymmetry, and the presence of publication bias was not statistically significant: using Egger’s test for small-study effects did not reach significance, as the bias coefficient was 16.76 [95% CI: [-7.09: 40.62] and the P-value 0.136.