Recovery after injury: an individual patient data meta-analysis of general health status using the EQ-5D

Abstract

Background: General information of health-related quality of life pathways to recovery after injury are largely absent from the literature. This article describes a study which: (1) collated and synthesized individual patient data of injured persons from an earlier systematic review and (2) produced general predictions of health-related quality of life for different injury groups for up to 1 year postinjury.

Methods: A systematic search of literature from January 1990 to December 2008 was completed. Researchers were approached to share their anonymous individual level data. Injuries were grouped into 39 categories based on the Eurocost injury classifications. Multilevel mixed effects models were used to produce predictions across both the five dimensions and the visual analog scale of the EQ-5D measure at 3 days, 30 days, 120 days, and 360 days postinjury.

Results: Individual patient data from 10,496 injured persons (76% of known data worldwide) was retrieved. Predictions were fitted to 27 of the 39 injury categories covering a wide spectrum of injury types. Across most injuries, pain, or discomfort, usual activities and mobility were the most commonly impaired dimensions. Recovery for pain or discomfort was generally more gradual than other health dimensions. For many injury categories, a considerable proportion of people reported residual impairment at 360 days. Regardless of the anatomic location of injury, similar patterns of recovery or persistent impairment were seen for fractures and strains/sprains. Recovery patterns differed and took much longer than estimated in the Global Burden of Disease Study.

Conclusions: This study has produced recovery patterns for 27 injury groups using most of the worldwide individual-level data. For many injury categories, recovery is incomplete and takes much longer than estimated. This study infers that the burden of injury is likely being underestimated.