Clark Fisher, MD, PhD
T32 Research Fellow, Anesthesiology
Yale School of Medicine
Diurnal Variation
in Anesthetic Care
Adapted from an Image Generated by DALL·E
DOI: 10.1136/qshc.2005.017566
Anesthetic Adverse Events
Duke 2000-2004
• Perioperative outcomes vary across the day
• Many potential causes
• One might be anesthetic care
Does anesthetic care vary
across the OR Day?
ASPIRE Measures
Temperature
Pulmonary
Postoperative
Nausea and
Vomiting
Neuromuscular
Monitoring
Blood Pressure Pain Management
Design
Inclusion Criteria
• Adult patients (≥18 years-old) in 2022 Standardized Data File
• January 1, 2016 – December 31, 2021
Exclusion Criteria
• Cardiac, obstetric and non-operative procedures
• ASA 5, 6, or E classes
• BMI < 15 | BMI > 60
• Weight < 30 kg | Weight >200 kg
• Fails inclusion/exclusion criteria >= 1 considered ASPIRE measure
• Inclusions include GETA, NMB, predicted surgical duration > 1 hr, predicted
ventilation > 45 minutes, extubation in OR, patients not taken to ICU
Exposure: Anesthesia Start Time
Early Late
Early Late
Exposure: Anesthesia Start Time
Outcome: Measure Adherence Rates
BP-02 Avoided monitoring gaps
BP-03 Avoided MAP < 65 mmHg
NMB-02 Reversed neuromuscular blockade
PAIN-02 Used non-opioid analgesics
PONV-05 Used graded PONV prophylaxis
PUL-03 Used appropriate PEEP
TEMP-02 Monitored core temperature
Results
Quality Measure Pass Rate Across the Day
Quality Measure Pass Rate Across the Day
Patients and Cases Vary Over the Day
Standardized Mean Difference (Late - Early)
Propensity Score Matching Achieves Balance
Standardized Mean Difference (Late - Early)
Effect of Late Start on Quality Measures
Effect of Late Start on Quality Measures
Effect of Late Start on Quality Measures
Interim Conclusions
•Shared trend of case start times
•Biggest case difference across day: case complexity and duration
•Many quality metrics remain constant across the day, however:
• Marginal decrease in use of non-opioid analgesics and PONV
prophylaxis as day progresses
•Improvement in hypotension avoidance as day progresses (NNT = 15)
Reflections on the MPOG
Standardized Data File
≈ MPOG Flat File
Case ID Date Institution …
e671f0fb-60 1/1/2016 1 …
f847f04x-g5 1/1/2016 4 …
f893750w-25 1/1/2016 19 …
… … … …
>200 Features
6 Years of Cases
• Uses robust phenotypes
• Less waiting for data
• No minute-by-minute
physiology or medication
data
https://mpog.org/sdf/
Do you need >12,000,000 cases?
Consider the time involved and account for the opportunity cost.
Do you need to check your data?
YES.
Thank You
Yale Anesthesiology
Robert B. Schonberger
Hung-Mo Lin
Yale Center for Analytical Science
Julian Zhao
Yanhong Deng
Multicenter Perioperative
Outcomes Group
Michael Mathis
MPOG Research Team
Support
NIGMS 5T32GM086287-14
MPOG Outcomes Research Fellowship
