Purpose: This article presents some issues for consideration before scaling from a pilot study to a larger investigation in longitudinal observational studies of traumatic brain injury (TBI) rehabilitation. Materials and methods: We present a case to discuss protocol improvements in longitudinal TBI-rehabilitation studies. The case was a pilot study conducted at two university hospitals in Denmark investigating 1-year outcomes related to sleep disturbance and agitation during neurointensive care. We included patients with moderate and severe TBI determined by the Glasgow Coma Scale, sleep disturbance was assessed using actigraphy, and agitation was assessed using the Agitated Behavior Scale. Results: Patients (n = 29) were more severely ill and had poorer six-month outcomes in Eastern vs. Western Denmark. Recovery was similar at one-year follow-up. Protocol improvements were needed in relation to inclusion criteria, intervention delivery, patient assessment, and follow-up outcomes. Conclusion: In TBI-rehabilitation studies, we suggest adding the severity of disease score to the initial GCS score and a delirium detection score to the ABS score. Actigraphy should not be used during deep sedation. Established procedures should be in place along all stages of the study protocol, including preparation and periodic assessment of study nurses to optimize data quality.