Maternal outcomes of cesarean sections: do generalists' patients have different outcomes than specialists' patients?

To compare maternal outcomes of cesarean sections performed by GPs with the outcomes of those performed by specialists.

Retrospective, comorbidity-adjusted study.

Mostly small isolated rural hospitals in Ontario, British Columbia, Alberta, and Saskatchewan compared with all levels of specialist obstetric programs offered in Canada.

Fifteen GPs with less than 1 year of surgical training who performed cesarean sections.

Using data from the Canadian Institute for Health Information's Discharge Abstracts Database for the years 1990 to 2001, we matched each of 1448 cesarean section cases managed by these GPs to 3 cases managed by specialists and looked for comorbidity. In total, we analyzed the outcomes of 5792 cesarean sections.

Composites of major morbidity possibly attributable to surgery:death, sepsis, cardiac arrest, shock, hypotension, ileus or bowel obstruction,major puerperal infection, septic or fat embolism, postpartum hemorrhage requiring hysterectomy, need for cardiopulmonary resuscitation, or another operation; and all major morbidity: major surgical morbidity, acute coronary syndrome, endocarditis, pulmonary edema, cerebrovascular disorder, pneumothorax, respiratory failure, amniotic fluid embolism, complications of anesthesia, deep vein thrombosis, pulmonary embolism, acute renal failure, and need for mechanical ventilation.

The rate of all major morbidity was higher among GPs' patients than among specialists' patients (3.1% vs 1.9%, odds ratio [OR] 1.6, 95% confidence interval [CI] 1.1 to 2.3, P = .009) as was the rate of major surgical morbidity (2.5% vs 1.6%, OR 1.6, 95% CI 1.1 to 2.4, P = .024). Differences in major morbidity variables were not significant if major postpartum infection was excluded (all major morbidity 1.5% vs 1.1%, major surgical morbidity 1.0% vs 0.8%). Secondary outcomes included rate of transfer to acute care institutions (6.0% vs 1.5%, OR 4.6, 95% CI 3.6 to 6.5, P < .001), mean length of hospital stay (5.2 vs 4.9 days, P= .006), need for blood transfusion (5.9% vs 7.0%, OR 0.76, 95% CI 0.5 to 1.1, P = .11) and frequency of surgical error (0.8% vs 0.7%, OR 1.1, 95% CI 0.6 to 2.3, P = .72).

Although major morbidity was higher among GPs' patients, differences were entirely attributable to the rate of postpartum infection. Infection rates in both groups were far below expected rates. The observation that blood transfusion and surgical error rates were similar suggests that surgical technique was not the cause of differences between groups. We conclude that these GPs with a mean of 4 months' training subsequently performed cesarean sections with an acceptable degree of safety compared with specialists.