Michael L. Burns, MD PhD
Assistant Professor, Department of Anesthesiology
Assistant Director of Informatics and Data Analytics
University of Michigan, Ann Arbor, Michigan , USA
MPOG Retreat
Friday, October 21st 2022
I am a co-investigator on project funding from the following
sources to my institution:
Blue Cross Blue Shield of Michigan (BCBSM)
Patient-Centered Outcomes Research Institute (PCORI)
I am a co-inventor on patent No. 11,288,445 B2 entitled
“Automated System and Method for Assigning Billing Codes to
Medical Procedures,” related to the use of machine learning
techniques for medical procedural billing.
Disclosures
Hypothesis
Anesthesiologists primarily operate in care team models
in the U.S., but the association of overlapping
anesthesiologist responsibilities with patient outcomes
remains unexplored.
We set out to study the potential association of this
overlapping care with surgical patient morbidity and
mortality.
Exposure
Staffing Ratio (SR)
Ratio of an anesthesiologist to the overlapping number of
rooms they cover
Each operation was classified into a single "staffing ratio"
group by calculating the time-weighted average of the ratio of
anesthesiologist to overlapping operations.
Exposure
Staffing Ratio (SR)
Ratio of an anesthesiologist to the overlapping number of rooms they cover
120 minutes
Exposure
Staffing Ratio (SR)
Ratio of an anesthesiologist to the overlapping number of rooms they cover
120 minutes
Exposure
Staffing Ratio (SR)
Ratio of an anesthesiologist to the overlapping number of rooms they cover
1:2 1:3 1:4
12 60 48
Exposure
12 60 48
Staffing Ratio (SR)
Ratio of an anesthesiologist to the overlapping number of rooms they cover
[(2 × 12) +(3 × 60) + (4 × 48)]/120 = 3.3
Primary Outcome
The primary outcome was a composite of mortality
and
6 perioperative complication categories, defined by ICD diagnoses, derived
from the Agency for Healthcare Research and Quality's (AHRQ) definitions:
● Cardiac
● Respiratory
● Gastrointestinal
● Urinary
● Bleeding
● Infection
Data
Electronic healthcare data from the Multicenter Perioperative Outcomes Group
Provider sign-in/sign-out data accurately captured due to billing and compliance reqs
Inclusion criteria:
● Elective procedures
● Adult patients (≥ 18 years of age)
● Jan 1, 2010 - Oct 31, 2017
● Surgical types: General,
Gynecologic, Neurological, ENT,
Orthopedic, Urology, Vascular
Exclusion criteria:
● Procedures with a “fixed staffing
ratio”: cardiac, liver transplants,
cataract removal, and obstetrics
● >25% resident involvement
● Missing anesthesia CPT
● Overnight, weekend, holiday cases
Methods
Propensity score matching methods were applied to create four balanced sample
groups
with respect to patient, procedure, and hospital level factors:
- Single case (SR = 1)
- 1-2 overlapping cases (1< SR ≤2)
- 2-3 overlapping cases (2< SR ≤3)
- 3-4 overlapping cases (3< SR ≤4)
Variables included in the propensity score derivation model and used to calculate the
likelihood of being in a particular staffing ratio group included age, sex, type of
operation, surgical service, anesthesia duration, and institution.
Three models were sequentially fit to obtain propensity scores corresponding to each
paired group, using 1< SR ≤2 as the reference.
Anesthesia Duration, 1 vs 1-2 = 0.34
At Teaching Institution, 3-4 vs 1-2 =
0.26
Operative Year, 3-4 vs 1-2 = 0.25
Results
Results
(a) P = .01 for group 1 vs group 1-2.
(b) P = .02 for group 2-3 vs group 1-2.
(c) P < .001 for group 3-4 vs group 1-2.
Limitations
1. 23 U.S. institutions
2. Relatively limited operative set
3. Challenging to address unmeasured confounders
4. Staffing ratios were limited to between 1:1 and 1:4
5. Limited to physician-led anesthesiologist care teams
6. Limited resident involvement to less than 25% in each operation
7. Unable to classify outcome severity
Manuscript Interpretations
● This was a care-team workload study, a topic of importance in every clinical field.
● Given anesthesiologist and CRNA shortages, understanding any potential impact of
workload on patient outcomes is essential to deliver quality care to our patients but
should be weighed against access benefits.
● We are not comparing individual providers, for example CRNAs vs anesthesiologists.
● These national data affirm our current approach to team-based care.
Conclusions
Compared to patients receiving care from an anesthesiologist covering
between 1-2 operations, 2-3 and 3-4 overlapping operations
demonstrated a higher risk of mortality/morbidity.
These findings highlight potential effects of anesthesiologist
responsibilities in perioperative team models and should be considered in
clinical coverage efforts.
It is important to balance potential efficiency with access benefits to
assess how much overlap may be appropriate.
Potential Next Steps
Quality of care exploration into staffing ratio to investigate the results we
have found in this study.
Investigate anesthesia team models compared to 1:0 (anesthesiologist
sitting their own cases).
Analyzing anesthesiology operation handovers.
