Impact of Preoperative Medical Consultation on a Large Surgical Cohort
- In a very large retrospective cohort of surgical patients, preoperative medical consultation showed a small but statistically significant association with increased 30-day mortality, in both unadjusted and propensity-matched analysis.
- Preoperative medical consultation was associated with higher costs, longer hospitalization, and an increase in processes of care, including more than twice the rate of beta-blocker initiation, echocardiogram, stress testing, heart catheterization, and coronary revascularization.
- A large predictor of referral was the type of hospital and region within the province of Ontario in which surgery was performed.
Does referral of surgical patients for preoperative medical consultation alter the risk of adverse postoperative outcomes or use of processes of care?
From linked administrative health care databases covering residents of Ontario, Canada, a cohort of patients ≥40 years of age discharged from the hospital between April 1, 2005–March 31, 2018 after undergoing one of 14 intermediate-high risk surgical procedures was identified. Propensity-score matching was performed to adjust for characteristics among patients who were or were not referred for preoperative medical consultation within a 4-month period prior to surgery. The primary outcome was 30-day postoperative mortality, and secondary outcomes included 1-year mortality, stroke or myocardial infarction (MI) during hospitalization, use of postoperative mechanical ventilation, length of stay, and 30-day health care costs. Processes of care were identified including echocardiograms, stress tests, coronary revascularization, and new beta-blocker and statin prescriptions.
Multivariate logistic regression predicting probability of referral for preoperative medical consultation was performed using demographic factors (age, sex, neighborhood income quintile), surgical factors (year of surgery, hospital type, surgical procedure category, access to preoperative evaluation by an anesthesiologist), and presence of comorbid medical conditions. The model included hospital location in one of 14 local health integration networks (LHINs) within the province of Ontario as random intercept. For sensitivity analysis to determine possibility of influential unmeasured predictors outside of the measured covariates, health system behaviors including adherence to recommended health screening within 2 years prior to surgery were compared between the groups.
A total of 530,473 patients were included, of whom 186,299 (35.2%) underwent preoperative medical consultation. Patients receiving preoperative referral to medical consultation had higher 30-day mortality rates.
Among the unadjusted population, 30-day mortality occurred in 0.9% (1,666/186,299) referred for medical consultation, versus 0.6% (2,159/344,174) not referred (odds ratio [OR], 1.43; 95% confidence interval [CI], 1.34-1.53; p < 0.001). Propensity-score matching produced 179,809 pairs, comprising 67.8% of the total cohort.
Among the adjusted cohort, 30-day mortality occurred in 0.9% (1,534/179,809) referred for medical consultation, versus 0.7% (1,299/179,809) not referred (adjusted OR, 1.19; 95% CI, 1.11-1.29). In the adjusted analysis, the referral group had the following:
- Higher 1-year mortality (3.9% vs. 3.4%, OR, 1.15; 95% CI, 1.11-1.19)
- Higher inpatient stroke incidence (0.3% vs. 0.2%, OR, 1.21; 95% CI, 1.06-1.37)
- Greater use of in-hospital mechanical ventilation (2.2% vs. 1.7%, OR 1.38; 95% CI, 1.31-1.45)
- Higher 30-day postoperative emergency department visits (17.7% vs. 17.3%, OR, 1.07; 95% CI, 1.05-1.09)
- Longer hospital length of stay (difference, 0.4 days; 95% CI, 0.3-0.5 days)
- Higher 30-day postoperative costs (difference, $319; 95% CI, 229-959)
- No difference in hospital MI incidence (1.1% vs. 1.0%, OR, 1.05; 95% CI, 0.99-1.12)
- Higher utilization of preoperative processes of care including:
- Echocardiogram: 27.4% vs. 13.0% (OR, 2.64; 95% CI, 2.59-2.69)
- Stress testing: 19.0% vs. 8.9% (OR, 2.50; 95% CI, 2.43-2.56)
- Coronary angiogram: 2.6% vs. 1.0% (OR, 3.08; 95% CI, 2.89-3.28)
- Coronary revascularization: 0.8% vs. 0.4% (OR, 2.23; 95% CI, 2.03-2.44)
- New prescriptions for beta-blockers: 5.2% vs. 1.9% (OR, 2.96; 95% CI, 2.82-3.12)
- New prescriptions for statins: 4.8% vs. 3.0% (OR, 1.62; 95% CI, 1.59-1.70)
After adjustment, rates of patient adherence to recommended health screening did not differ significantly between groups. Rates of preoperative medical consultation referral remained stable over time but varied substantially among the 14 LHINs (ranging from 16.0–47.0%). Findings among categorical subgroups were consistent with those of the entire adjusted population.
The association between preoperative medical referral and 30-day mortality was stronger when patients were referred to medical subspecialists (OR, 1.46; 95% CI, 1.29-1.65) compared to general internists (OR, 1.15; 95% CI, 1.06-1.27). The authors did not make statistical adjustment for multiple comparisons.
In this very large cohort of patients ≥40 years of age undergoing intermediate-high risk noncardiac surgery, referral for preoperative medical consultation was not associated with reduction in major adverse postoperative outcomes, including 30-day postoperative mortality, and showed a small but statistically significant signal toward harm. Referral for medical consultation was associated with consistently greater use of preoperative processes of care.
Although the authors cite difficulty in inferring causality between medical consultation and adverse outcomes in this retrospective observational trial, they offer several possible explanations. The large impact of hospital and region on referral implies that the referral frequently reflects practice patterns, as opposed to recognition that an individual patient requires clinical optimization. Also raised is paucity of interventions conclusively proven to improve postoperative outcomes. Finally, possible association between preoperative consultation and surgical delay as a source of harm is mentioned.
Clinical Topics: Cardiac Surgery, Cardiovascular Care Team, Dyslipidemia, Invasive Cardiovascular Angiography and Intervention, Noninvasive Imaging, Prevention, Cardiac Surgery and Arrhythmias, Nonstatins, Novel Agents, Statins, Interventions and Imaging, Angiography, Echocardiography/Ultrasound, Nuclear Imaging
Keywords: Adrenergic beta-Antagonists, Anesthesiologists, Coronary Angiography, Critical Care, Echocardiography, Exercise Test, General Surgery, Hydroxymethylglutaryl-CoA Reductase Inhibitors, Length of Stay, Myocardial Infarction, Myocardial Ischemia, Myocardial Revascularization, Patient Care Team, Patient Discharge, Preoperative Care, Postoperative Care, Prescriptions, Secondary Prevention, Respiration, Artificial, Stroke
< Back to Listings