Concordance Between Bedside and Electronic Detection of Intracranial Pressure Crises: Insights From the Brain Tissue Oxygen Monitoring and Management in Severe Traumatic Brain Injury II Trial

OBJECTIVE: Traumatic brain injury (TBI) research and intracranial pressure (ICP) management depends on bedside ICP (B-ICP) crisis identification. We analyzed background-collected electronic ICP (E-ICP) data to study the concordance of this identification. DESIGN: Post hoc comparison of background-collected continuous E-ICP data to routine B-ICP information during severe TBI management in the Brain Tissue Oxygen Monitoring and Management in Severe TBI (BOOST II) randomized trial. SETTING: Ten U.S. ICUs. PATIENTS: Seventy of 110 randomized BOOST II severe TBI patients with complete datasets for this study. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: We studied "minimalist" B-ICP episodes lasting less than or equal to 60 minutes and requiring only tier 1 treatments (83% of total BOOST II I isolated ICP episodes). Bedside clinicians (BCs) identified and treated 509 minimalist B-ICP episodes, defined by 5 minutes of B-ICP greater than or equal to 20 mm Hg (B-ICP 5 ). Corresponding E-ICP during this defining period (E-ICP 5 ) confirmed only 47% of these. The 241 "concordant" B-ICP episodes (both B-ICP 5 and E-ICP 5 ≥ 20 mm Hg) had average E-ICP values for the entire B-ICP episodes (E-ICP AVG ) less than 20 mm Hg in 38%. The 286 "discordant" B-ICP episodes (B-ICP 5 ≥ 20 mm Hg but E-ICP 5 < 20 mm Hg) had E-ICP AVG values less than 20 mm Hg in 76% and both E-ICP AVG and maximal E-ICP's (E-ICP MAX ) values less than 20 mm Hg in 31%. Testing for confounding (e.g., brief, first, or easily controlled B-ICP episodes) did not provide explanations. Insufficient data were available to evaluate temporal asynchrony between BC and electronic datapoints. Study limitations were lacking a rigid end-of-B-ICP episode definition and inability to fully control for temporal synchrony confounding. CONCLUSIONS: Our findings suggest that current methods of clinical identification of B-ICP episodes may not reliably distinguish true episodes of sustained ICP for the most common B-ICP episode type. We suggest developing open-source, real-time, temporally synchronized electronic B-ICP episode definition methods to direct future treatment and research.