Discussion Questions

• ■
  

  What was the study objective?

  
  
• ■
  

  Can you summarize the study design?

  
  
• ■
  

  Who comprised the study population?

  
  
• ■
  

  Can you describe the statistical methodology?

  
  
• ■
  

  What are the advantages and limitations of the validation strategy?

  
  
• ■
  

  How do these findings affect individual patient management?

  
  
• ■
  

  How do risk-assessment models improve patient care?

  
  
• ■
  

  What are the strengths and weaknesses of this study?

Introduction

Of the more than 600,000 hysterectomies performed yearly in the United States, some 20-25% are completed exclusively by the vaginal route, an approach associated with a shorter stay and faster recovery than occurs with abdominal surgery; when performed by experienced surgeons, the methods have comparable complication rates. Interest in national systems to improve surgical quality is growing. The National Surgical Quality Improvement Program has reduced postoperative morbidity and mortality by 30-45%. Reliable metrics for assessment of risk-adjusted outcomes after common procedures are essential to such programs. This month, Journal Club members discuss a new model for predicting perioperative complications in patients undergoing vaginal hysterectomy.

Israel Zighelboim, MD and George A. Macones, MD, MSCE, Associate Editor

Study Design

Zighelboim: What was the study objective, and what was the rationale behind this type of study?

Ms Zhang: Almost a quarter (23.3%) of women surveyed for the Behavioral Risk Factor Surveillance System said they had had hysterectomies, according to the Centers for Disease Control and Prevention. Patients undergoing hysterectomy for benign conditions are generally healthy, and the procedure is done primarily to improve their quality of life. It is thus very important to identify factors that could predispose these patients to perioperative complications. While intra- and postoperative complications of hysterectomies have been reviewed in several studies, most of the work has focused on abdominal and laparoscopic approaches. In the 1980s, 26% of all hysterectomies in the United States were performed vaginally. This rate increased significantly in the early 1990s and has since leveled off at around 30%. Therefore, investigation of quality improvement methods in vaginal hysterectomies is both timely and clinically relevant.

Zighelboim: Can you summarize the study design?

Ms Zhang: This retrospective cohort study examined a group of subjects over time to identify any correlation between various factors and a specific outcome. In this case, the cohort was a group of women undergoing vaginal hysterectomy for benign conditions. Records from the initial surgical consultations and preanesthesia medical examination were used to identify comorbidities and other perioperative risk factors. Results from standard perioperative laboratory tests were also analyzed.

The outcome of interest was the occurrence of complications within 9 weeks of the index surgery. These complications included hospital readmission, repeat surgery, unplanned intensive care admission, or intervention for medical problems such as infection, symptomatic anemia, and fluid overload. The authors chose to exclude urinary tract infection (UTI) and antibiotic treatment for UTI. After data collection, multivariate logistic regression models were used to identify factors most significantly associated with the selected perioperative complications. These were fitted into a risk-adjusted model for predicting the outcome of interest. Finally, the model’s validity was assessed by applying it to a randomly-generated sample of patients who had a vaginal hysterectomy.

Zighelboim: Who comprised the study population?

Anand: The study population consisted of patients who underwent vaginal hysterectomy for benign indications with or without additional vaginal gynecologic procedures. Data from 712 women treated at the Mayo Clinic between the years 2004-2005 were studied; details are presented in Table 1 of the article. Patients were predominantly privately insured, perimenopausal, and Caucasian, and generally, they had a low prevalence of comorbidities. Nearly one-half of the patients had a hysterectomy for uterine prolapse and/or menorrhagia, and one-quarter of patients underwent surgical correction for urinary incontinence. Few had prior surgery for prolapse or incontinence. Approximately one-half of the patients underwent vaginal hysterectomy with or without salpingectomy or oophorectomy; the remaining patients also underwent reconstructive vaginal pelvic surgery. Patients who had additional nongynecologic surgical procedures were excluded.