Abstract
Objectives: To assess determinants of loss to follow-up (FU) at 2 time points of an inception traumatic brain injury (TBI) cohort.
Design and Participants: The PariS-TBI study consecutively included 504 adults with severe TBI on the accident scene (76% male, mean age 42 years, mean Glasgow Coma Scale 5). No exclusion criteria were used.
Main Measure: Loss to FU at 1 and 4 years was defined among survivors as having no outcome data other than survival status.
Results: Among 257 1-year survivors, 118 (47%) were lost to FU at 1 year and 98 (40%) at 4 years. Main reasons for loss to FU were impossibility to achieve contact (109 at 1 year, 52 at 4 years) and refusal to participate (respectively 5 and 24). At 1 year, individuals not working preinjury or with nonaccidental traumas were more often lost to FU in univariate and multivariable analyses. At 4 years, loss to FU was significantly associated with preinjury alcohol abuse and unemployment. Relationship with injury severity was not significant.
Conclusions: Socially disadvantaged persons are underrepresented in TBI outcome research. It could result in overestimation of outcome and biased estimates of sociodemographic characteristics' effects. These persons, particularly unemployed individuals, require special attention in clinical practice.