i
Eye injury and demographic parameters associated with poor visual outcome Lésion oculaire et paramètres démographiques associés à un faible résultat visuel
-
May 20 2019
-
-
Source: J Fr Ophtalmol. 42(8):864-873
[PDF-366.37 KB]
Details:
-
Alternative Title:J Fr Ophtalmol
-
Personal Author:
-
Description:Background :
Eye injuries can result in long-term disability, and healthcare providers need better tools to predict outcomes. Few prognostic models for poor visual acuity have been examined using variables usually present in very severe injuries, which creates a gap in prognosis. Therefore, a model associated with severe and less severe injuries is examined.
Methods :
Eye injuries hospitalized in Bosnia and Herzegovina from 2006 through 2014 were included. A total of 298 eye injuries met the inclusion criteria of being an acute mechanical, chemical or physical eye injury. Prognostic variables were grouped by patient characteristics, eye injury characteristics and eye injury diagnosis. Poor final visual acuity was the main outcome measure (vision less than 20/200). Multivariate regression analysis used stepwise selection to identify the strongest set of predictive variables.
Results :
Lens subluxation (95 % CI : 2.09–14.83), vitreous prolapse (95 % CI : 2.76–26.87), vitreous hemorrhage (95 % CI : 1.71–10.03), posterior segment intraocular foreign body (95 % CI : 1.19–39.09), and vitritis (95 % CI : 0.97–11.12) were significantly associated with poor final visual acuity. The predictive model identified the combination of age over 36, lens subluxation, vitreous prolapse and hemorrhage, vitritis, and macular hemorrhage as the combination most likely to have poor visual acuity. The final model resulted in a strong fit as measured by AIC, log likelihood, goodness-of-fit, and the c-statistic.
Conclusions :
This model can be used in clinical practice to evaluate severity and predict final visual acuity in both severe and less severe eye injuries. The model accounts for characteristics of the injury as well as the patient. Additional studies with larger samples could further verify this model.
-
Subjects:
-
Source:
-
Pubmed ID:31122763
-
Pubmed Central ID:PMC6778009
-
Document Type:
-
Funding:
-
Place as Subject:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type:
