INS NYC 2024 Program

Poster	Poster Session 04 Program Schedule 02/15/2024 12:00 pm - 01:15 pm Room: Majestic Complex (Posters 61-120) Poster Session 04: Neuroimaging \| Neurostimulation/Neuromodulation \| Teleneuropsychology/Technology Final Abstract #96 Building a Surgical Database from Decades of Medical Records for Prospective Study of Perioperative Neurocognitive Disorders: Framingham Heart Study Eilis Reardon, Boston University School of Medicine, Boston, United States Ala Nozari, Boston University School of Medicine, Boston, United States Ting Fang Ang, Boston University School of Medicine, Boston, United States Chenglin Lyu, Boston University School of Medicine, Boston, United States Maxwell Baker, Boston University School of Medicine, Boston, United States Rhoda Au, Boston University School of Medicine, Boston, United States Catherine Price, University of Florida School of Medicine, Gainesville, United States Ashita Gurnani, Boston University School of Medicine, Boston, United States Category: Assessment/Psychometrics/Methods (Adult) Keyword 1: dementia - Alzheimer's disease Keyword 2: delirium Keyword 3: medical disorders/illness Objective: Perioperative Neurocognitive Disorders (PND) is among the most concerning surgical complications, particularly in older patients with pre-existing cognitive decline. Surgeries are sources of inflammatory exposure which is a risk factor for Alzheimer’s disease and other related dementias (ADRD). The understanding of PND’s relationship to ADRD risk has been hampered by contradictory findings and methodological limitations including small sample sizes, use of insensitive cognitive tests, types of surgeries, lack of detailed preoperative clinical history, and short follow-up for incident cognitive impact. The Framingham Heart Study (FHS), initiated by the National Heart, Lung, Blood Institute (NHLBI) to identify determinants of cardiovascular disease (CVD) and stroke, has only documented history of cardiac or vascular surgeries. There is a need for a comprehensive, longitudinal surgical database that includes the type and timing across different surgical procedures and perioperative clinical parameters. The current presentation provides an overview of the data curation protocols and quality control metrics used to harmonize and curate data extracted from decades of medical records across the FHS cohorts. Participants and Methods: FHS participants have been followed for incident dementia since 1976 and administered a comprehensive neuropsychological battery every 4-6 years since 1999. PND experts designed a standardized data collection REDCap form to extract data from medical records dating as far back as 1927, including surgery type, perioperative measurements/events and procedure-related symptoms/medications. Reviewers were trained using a structured training protocol and assessed using five standardized medical charts, prepared by PND experts. 5% of the completed charts were randomized monthly for re-reviews as post-training quality assurance, to monitor inter- and intra-reviewer variability. Findings of deviations from training protocol were discussed during team meetings and retraining of reviewers was conducted to ensure data consistency. To test the validity of the data collection and training protocols, a pilot study was conducted on a subset of older participants with known history of coronary artery bypass graft (CABG) and/or percutaneous transluminal coronary angioplasty (PTCA), based on verified cardiovascular endpoint reviews. Results: The REDCap form collects information on 23 preoperative, 22 intraoperative, 33 postoperative and 10 medication metrics for each procedure. Average training time for each reviewer was 30 hours including assessment using training charts (four hours each). 103 charts were reviewed for the pilot study, with surgical records spanning from 1930 to 2022. Participants’ first and last procedures occurred on average at age 47 (± 20 years) and at age 75 (± 9 years), respectively.. Reviewers extracted information from cardiovascular procedures with 100% accuracy and no major deviations were detected during quality assurance. Colonoscopy was the most common of 853 recorded procedures recorded, and occurred 56 times across 39 participants. Conclusions: Retrospective chart review is a powerful tool for collecting longitudinal data from medical records to establish a prospective data resource to study the relationship between PND and risk for incident cognitive decline.This data extraction infrastructure is enriching the robust characterization of FHS participants by adding consistently coded measures of surgical procedures, mitigating the necessity for prolonged periods of prospective data collection.

Poster

02/15/2024
12:00 pm - 01:15 pm
Room: Majestic Complex (Posters 61-120)

Poster Session 04: Neuroimaging | Neurostimulation/Neuromodulation | Teleneuropsychology/Technology

Final Abstract #96

Building a Surgical Database from Decades of Medical Records for Prospective Study of Perioperative Neurocognitive Disorders: Framingham Heart Study

Eilis Reardon, Boston University School of Medicine, Boston, United States
Ala Nozari, Boston University School of Medicine, Boston, United States
Ting Fang Ang, Boston University School of Medicine, Boston, United States
Chenglin Lyu, Boston University School of Medicine, Boston, United States
Maxwell Baker, Boston University School of Medicine, Boston, United States
Rhoda Au, Boston University School of Medicine, Boston, United States
Catherine Price, University of Florida School of Medicine, Gainesville, United States
Ashita Gurnani, Boston University School of Medicine, Boston, United States

Category: Assessment/Psychometrics/Methods (Adult)

Keyword 1: dementia - Alzheimer's disease
Keyword 2: delirium
Keyword 3: medical disorders/illness

Objective:

Perioperative Neurocognitive Disorders (PND) is among the most concerning surgical complications, particularly in older patients with pre-existing cognitive decline. Surgeries are sources of inflammatory exposure which is a risk factor for Alzheimer’s disease and other related dementias (ADRD). The understanding of PND’s relationship to ADRD risk has been hampered by contradictory findings and methodological limitations including small sample sizes, use of insensitive cognitive tests, types of surgeries, lack of detailed preoperative clinical history, and short follow-up for incident cognitive impact. The Framingham Heart Study (FHS), initiated by the National Heart, Lung, Blood Institute (NHLBI) to identify determinants of cardiovascular disease (CVD) and stroke, has only documented history of cardiac or vascular surgeries. There is a need for a comprehensive, longitudinal surgical database that includes the type and timing across different surgical procedures and perioperative clinical parameters. The current presentation provides an overview of the data curation protocols and quality control metrics used to harmonize and curate data extracted from decades of medical records across the FHS cohorts.

Participants and Methods:

FHS participants have been followed for incident dementia since 1976 and administered a comprehensive neuropsychological battery every 4-6 years since 1999. PND experts designed a standardized data collection REDCap form to extract data from medical records dating as far back as 1927, including surgery type, perioperative measurements/events and procedure-related symptoms/medications. Reviewers were trained using a structured training protocol and assessed using five standardized medical charts, prepared by PND experts. 5% of the completed charts were randomized monthly for re-reviews as post-training quality assurance, to monitor inter- and intra-reviewer variability. Findings of deviations from training protocol were discussed during team meetings and retraining of reviewers was conducted to ensure data consistency. To test the validity of the data collection and training protocols, a pilot study was conducted on a subset of older participants with known history of coronary artery bypass graft (CABG) and/or percutaneous transluminal coronary angioplasty (PTCA), based on verified cardiovascular endpoint reviews.

Results:

The REDCap form collects information on 23 preoperative, 22 intraoperative, 33 postoperative and 10 medication metrics for each procedure. Average training time for each reviewer was 30 hours including assessment using training charts (four hours each). 103 charts were reviewed for the pilot study, with surgical records spanning from 1930 to 2022. Participants’ first and last procedures occurred on average at age 47 (± 20 years) and at age 75 (± 9 years), respectively.. Reviewers extracted information from cardiovascular procedures with 100% accuracy and no major deviations were detected during quality assurance. Colonoscopy was the most common of 853 recorded procedures recorded, and occurred 56 times across 39 participants.

Conclusions:

Retrospective chart review is a powerful tool for collecting longitudinal data from medical records to establish a prospective data resource to study the relationship between PND and risk for incident cognitive decline.This data extraction infrastructure is enriching the robust characterization of FHS participants by adding consistently coded measures of surgical procedures, mitigating the necessity for prolonged periods of prospective data collection.