Association of Overlapping Surgery with Perioperative 10.18.2019

Association of Overlapping Surgery with Perioperative

Outcomes

Eric Sun, Michelle M. Mello, Chris A. Rishel, Michelle T.

Vaughn, Sachin Kheterpal, Leif Saager, Lee A. Fleisher,

Edward J. Damrose, Bassam Kadry, Anupam B. Jena

Disclaimer

• The work I am discussing precedes my appointment to the staff of the Council

of Economic Advisers

• I am speaking in my capacity as an assistant professor at Stanford University

• Nothing in this talk should be construed as representing the views of the

Federal Government

Disclosures

• Funding

• Sun: NIDA (K08DA042314)

• Jena: Office of the Director, NIH (DP50D017897)

• No COI to report

Motivation

• Overlapping surgery refers to a situation where 2+ procedures performed by the

same surgeon are scheduled so that the start of one procedure overlaps with

the end of the other

• Has received attention in the popular press and from policymakers due to

concerns about worse patient outcomes

• Are these concerns warranted?

Preview of Results

• Large retrospective analysis (>60,000 cases) of surgeries from eight centers

• Overall, overlapping surgery was not associated with worse outcomes

• In-hospital Mortality (1.9% overlapping vs. 1.6% nonoverlapping, p=0.21)

• Complications (12.8% overlapping vs. 11.8% nonoverlapping, p=0.08)

• Overlapping surgery was associated with increased length (204 vs. 173

minutes, p<0.001)

• Increases in mortality and complications seen for

• High Risk Patients (mortality 5.8 vs 4.7, p=0.03; complications 29.2% vs

27.0%, p=0.03)

• CABG (mortality 4.0% vs. 2.2%, p=0.009; complications 34.5% vs. 30.2%,

p=0.007)

Introduction

• Several studies have examined the association between overlapping surgeries

and perioperative outcomes (Morris, 2017)

• In general have found no association

• Limitations of prior studies

• Small sample size—potentially underpowered, particularly for mortality

• Focused on single procedure, single institution

• Used overly stringent definition of overlap (i.e., 1 or more minutes) which may

bias estimates downwards

• Study design-compared across surgeons

• May mask effects of overlapping surgery if surgeons who perform

overlapping surgery are more experienced

• One previous study found an association between concurrent surgery and

perioperative complications compared cases within surgeon (Ravi et al.,

2017)

Methods

• Benefits of MPOG Data

• Contain data that are crucial to studying this question

• Surgical times

• Surgeon identifiers

• Data on outcomes

• Mortality, ICD codesoutcomes

• Data for risk adjustment (ICD codes)

• Large number of cases

• Multiple Institutions

Methods

• MPOG Challenges

• Data not consistently reported across institutions

• Ultimately, only 8 institutions consistently reported data elements necessary

for this study

• Difficulty identifying surgery

• Initial plan was to include all surgeries and adjust for surgical CPT code

• However, surgical was very inconsistently reported (although this has

changed)

• In response, we made two changes to the plan

• Restricted to 7 surgeries (TKA, THA, crani, CABG, L/C/T-spine surgery)

• Identified these surgeries by parsing the free text descriptors (i.e. “L

TKA”)

Methods

• Variables taken from MPOG

• Surgery Start/Stop times

• Masked Surgeon ID

• Masked Institution ID

• Free text description of surgery

• In-hospital mortality

• Diagnosis Codes

• Patient Age

• Patient Sex

• ASA Score

Sample

• Inclusion criteria:

• Age 18-90

• Surgeries: TKA, THA, spine surgery (lumbar/thoracic/cervical), CABG,

craniotomy

• January 1,2010 to May 31, 2018

• 8 institutions reporting diagnosis codes >70% of cases

Sample

• Initial sample 87,560 cases

• Exclusion criteria:

• Missing data

• Diagnosis codes (n=7,023)

• Mortality (n=1,070)

• Surgery duration (n=4,728)

• Sex (n=30)

• ASA score (missing or=6; n=1,994)

• >1 surgery type (n=4,167)

• Surgeon-procedure combos with <20 observations (n=2,118)

• Final sample 66,430 (8,224 overlapping)

• 207 surgeons

• 373 surgeon-procedure pairs

Empirical Approach

• Overlapping surgery defined as a case with 60+ minutes of overlap with surgeon’s

other cases

• Or entirety of case if total duration <=60 minutes

• Outcomes

• In-hospital mortality

• In-hospital complications

• SSI, UTI, PNA, sepsis, DVT/PE, MI, stroke

• Surgery Length

• Multivariable regression utilizing surgeon-procedure fixed effects

• Our analysis explicitly compares outcomes between a given surgeon’s overlapping

and non-overlapping surgeries (of the same type)

• e.g., Dr. Smith’s overlapping vs non-overlapping TKAs and overlapping vs. non-

overlapping THAs

• Implicitly also adjusts for hospital-level factors

• e.g. OR staff

Sample Characteristics

• Generally speaking, few differences between overlapping and non-overlapping

patients

• Age, sex, ASA score

• Where there were differences, overlapping patients were generally sicker

• 7 comorbidities more common in overlapping cases

• 2 more common in nonoverlapping cases

Overlapping surgery and mortality

Overlapping Surgery and Complications

Overlapping Surgery and Surgery Times

Conclusions

• Overlapping surgery is generally safe

• However caution and further study warranted for some high-risk groups

• Is associated with longer cases (~30 minutes)

• Potential cost considerations ($30-$60/min)

Acknowledgements

• Funding

• Sun: NIDA (K08DA042314)

• Jena: Office of the Director, NIH (DP50D017897)

• Department of Anesthesiology, Perioperative and Pain Medicine at Stanford

University

• MPOG PCRC