Participant attrition in longitudinal studies of cochlear implant outcomes in children

Speech and language scores obtained from 183 prelingually deafened children followed longitudinally at Indiana University School of Medicine were retrospectively analyzed. All participants were assessed preimplant and at regular postimplant intervals. Fifty-two children withdrew from participation before completing 3 years of cochlear implant (CI) use (mean CI use at withdrawal was approximately 2 years). The remaining 131 children participated through 3 or more years of device use (mean≈5.5 years). Data obtained from the 2-year postimplant interval were examined qualitatively and quantitatively to identify differences in the distribution of scores and overall performance between groups. Vocabulary and open-set speech recognition did not differ significantly between the two groups. However, qualitative differences were noted. Vocabulary scores were bimodally distributed for children who withdrew from the study but normally distributed for long-term participants. Similar trends were noted for open-set word recognition; the 75th percentile was 10% for children who withdrew and 30% for children who participated more than 3 years. Our findings suggest that reports of longitudinal speech and language outcomes in children with CIs represent data from children at all levels of performance, and do not only represent the children who are performing well with their cochlear implant.