Brain Health Screening in Older Surgical Patients: A Multicenter, Retrospective, Observational Analysis and Survey

Guidelines recommend routine screening perioperatively for cognitive impairment, frailty, and delirium for patients at risk. Capturing these 3 geriatric screening variables in multicenter databases would also enable much-needed large-scale pragmatic research. Our primary hypothesis was that the well-curated Multicenter Perioperative Outcomes Group (MPOG) database would have a low rate of retrievable geriatric screening variables. Our secondary hypothesis was that multiple barriers exist that impede clinical implementation of recommended screenings as well as the digital capture of these variables into the MPOG database.

This was a 2-component study. The first component was a retrospective observational analysis using the MPOG database to identify geriatric screening variables in patients over the age of 65 undergoing nonemergent inpatient surgery. The second component was a survey sent to MPOG participant sites (49 institutions) to assess actual screening practices and perspectives.

Of the 908,158 relevant patient records only 8054 (0.89%) were identified as having a preoperative cognitive screen, and 123,114 (13.6%) were identified as having a postoperative delirium screening. No frailty screenings were observed. Forty-3 survey responses (88% response rate) were received. Approximately half of the respondents indicated their institutions perform cognitive screening (n=22; 51.2%), frailty screening (n=17; 44.7%), or delirium screening (n = 16; 45.7%). Only 10 institutions (23.2%) reported performing all 3, and 13 (30.2%) institutions reported performing none. Multiple barriers were identified. The most common significant barrier reported was a lack of available standard screening tools for the electronic health record.

This study identified minimal data collection related to neurocognitive disorders which appears incongruous with clinical practice guidelines. Challenges related to capturing this data locally and in multi-center datasets were identified. Overcoming those barriers may facilitate future pragmatic research studies related to this important public health priority.