Predictors of traumatic intracranial lesions in clinically diagnosed mild head injury patients at Mulago Hospital
MetadataShow full item record
Background: Head injury is one of the most common causes of admission at Mulago hospital with the bigger proportion of it being categorised as mild head injury (MHI). Study Aims: The main aims of this study were to determine the prevalence of traumatic intracranial lesions (TICLs) and to identify the most reliable clinical predictors of TICLs in clinically-diagnosed adult MHI patients at Mulago hospital Methods: This was a cross-sectional study of 204 clinically diagnosed adult mild head injury patients at Mulago hospital. We did cross-tabulation to study the association between clinical features and a traumatic intracranial lesion (TICL) through a Chi square (X2) test. To determine the strength and direction of association we estimated the odds ratios using simple logistic regression. We finally did a correlation analysis of all clinical characteristics to select one of each correlated clinical characteristics which we included into a multivariate logistic regression and then ran multiple analyses on all variables having a P-Value less than or equal to 0.1(10%). We then did model performance comparison using Akaike Information Criteria (AIC) and the adjusted Macfadden R2 values to identify the best performing predictor model. Results: The prevalence of TICLs identifiable on head CT scan was estimated at 66 % [95% CI (59-72%)]. The most frequent TICL identified was a cerebral contusion with a prevalence of 33.8% [95%CI (25.99%-39.24%)] and the least frequent TICLs were sub-arachnoid and intra-cerebral haemorrhages each with a prevalence of 5.4% [95%CI (3.08%-10.5%)]. We identified a set of 6 factors that showed the highest reliability in identifying patients with TICLs (sensitivity 89.63%, specificity of 43.48%) Individual predictors within the set included Severe and moderate headache, a referred patient, male gender, GCS=13, and presence of raccoon eyes. The strongest individual predictor of a TICL in a clinically-diagnosed Mild Head injury (MHI) patient was severe headache with an adjusted odds ratio of 5.7. Conclusion: Our study established 6 clinical features that could be used to identify patients with clinically-diagnosed mild head injury who would benefit from a head CT scan. They included Severe and moderate headache, a referred patient, male gender, GCS=13, and presence of raccoon eyes. These factors could be considered positive indicators for the need of a head CT scan for clinically diagnosed MHI patients and allow clinicians to be more selective in use of the CT scan without compromising patient care.