Materials and Methods

Hysterectomy cases between July 1, 2012, and July 1, 2014, in the Michigan Surgical Quality Collaborative (MSQC) were included. The MSQC is a statewide organization of 52 academic and community hospitals. Funded by Blue Cross and Blue Shield of Michigan/Blue Care Network, it includes patients from all insurance payers. At each participating hospital, a trained, dedicated nurse abstractor collects the patient characteristics, intraoperative processes of care, and 30 day postoperative outcomes from the general and vascular surgery and hysterectomy cases.

To reduce sampling error, a standardized data collection methodology that uses only the first 25 cases of an 8 day cycle (alternating on different days of the week for each cycle) is used. The standardized data collection methodology is routinely validated through scheduled site visits, conference calls, and internal audits. This study met the criteria for not regulated status by the University of Michigan Institutional Review Board–Medical (HUM00073978).

The collection of SSI data was consistent with definitions from the Centers for Disease Control and Prevention. An SSI had to occur within 30 days of the procedure and was classified as superficial or deep. A superficial SSI involved only skin and subcutaneous tissue of the incision. In this study, the term, deep SSI, refers to both deep and organ/space SSIs. They were combined because the fascia and muscle layers of the vaginal cuff are contiguous with the hysterectomy organ space.

Demographics, medical history, and complications were analyzed. Age was dichotomized at 50 years as an estimate for pre- and postmenopause. Values closest to the surgery date and within 90 days were collected for height, weight, and preoperative platelet count. Patients’ primary insurance was categorized as Medicare, Medicaid, or private. Women were considered to have Medicare if they had Medicare, Medicare with a supplemental plan (eg, Medigap), or Medicare Advantage (Blue Cross Blue Shield or Blue Care Network of Michigan).

Women were considered to have private insurance if they had Blue Cross Blue Shield of Michigan, Blue Care Network, health maintenance organization plans, and other private insurance plans.

Patients were considered to have Medicaid if they had Medicaid or a health maintenance organization Medicaid plan. A woman was considered uninsured if Medicaid pending was listed at discharge. The groups of patients having insurance listed as other or self-pay were considered separately. The diagnosis of gynecological cancer was considered present if the primary International Classification of Disease , ninth edition, code was 179–184, 198.6, 198.81, 198.82, or 233.0–233.9. Medical status was dichotomized using the American College of Surgeons (ASA) class such that patients with ASA classes 1 and 2 were placed in one group and ASA classes 3, 4, and 5 (severe systemic disease or worse) were placed in another.

Preoperative transfusion was a minimum of 1 U of whole blood/packed red blood cells during the 72 hours prior to surgery. Diabetes mellitus was considered present if the woman required oral hypoglycemic agents and/or insulin and not present if management was diet control. Hypertension was considered present when documented preoperatively or when antihypertensive treatment was used preoperatively. Tobacco use was indicated if a patient smoked cigarettes, cigars, or pipe, or chewed tobacco or used marijuana within the past year.

A history of deep venous thromboembolism (DVT) was noted if there was a history of DVT or pulmonary embolus (PE). Approach to hysterectomy was categorized as open, laparoscopic (including robotic-assisted cases), and vaginal (including laparoscopic-assisted cases). Postoperative transfusion was transfusion of whole red blood cells up to 72 hours after the surgery.

Complications within 30 days of surgery were readmission, reoperation, urinary tract infection, sepsis, myocardial infarction or stroke, venous thromboembolic event (DVE or PE), and death.

All cases from sites contributing at least 20 hysterectomies were included in the analysis. For bivariate analyses, χ ² , Fisher exact, parametric, and nonparametric (Wilcoxon-Mann-Whitney) tests were used. Multivariable hierarchical logistic regression models were developed to identify independent patient risk factors and hospital effects. Variables were excluded from model selection if not significant or not related in a clinically plausible manner.

Patients who were uninsured or self-pay or who were covered by Medicare and Medicaid, Worker’s Compensation, auto insurance, or government-sponsored plans such as Veterans Affairs or TriCare were excluded from model building because of the small sample sizes. Variables were considered using stepwise, backward, and forward selection. Multicollinearity was assessed using correlation matrices. Observed and model predicted rates were assigned by decile, quintile, and quartile to assess fit. Model fit was tested with Pearson’s residuals, with strong fit being close to 1 and calculation of the C statistic.

We investigated the impact of risk factor adjustment for deep or organ space SSI on hospital rank and quartile status. Deep and organ space SSIs were the focus of the analysis because these are the relevant events for the HAC reduction program reported by the Centers for Disease Control and Prevention to the Centers for Medicare and Medicaid Services.

Hospital rankings within MSQC were determined by calculating the adjusted rates and then ranking hospitals by the predicted to expected (P/E) ratio. Predicted-expected (vs observed to expected) ratios were used because a hierarchical logistic regression model taking into account hospital effects had been developed. The 95% confidence intervals were calculated for each hospital to identify outliers with respect to the mean adjusted MSQC rate and then to the reference line of one for the P/E ratio plot.

Results

There were 16,548 hysterectomies available for analysis in the MSQC. Hysterectomy cases were abstracted by 51 of 52 hospitals. Two sites with less than 10 cases were not included, leaving 49 hospitals available for ranking. Over the 2 year time period, case volume at the 49 institutions ranged from 66 to 793. The rate of all types of SSIs was 2.1% (n = 351). The rate of deep SSI, a subgroup of women having either deep or organ space SSI, was 1.0% (n = 167) and the rate of superficial SSI was 1.1% (n = 184).

Demographics, medical comorbidities, and perioperative variables for women undergoing hysterectomy were analyzed as a function of overall SSI and deep SSI ( Table 1 ). Women with SSI, compared with those without, more frequently had a body mass index (BMI) ≥ 30 kg/m ² and Medicaid insurance. Age and parity were similar. Women with SSI were more likely to have used tobacco and to have gynecological cancer, ASA class ≥ 3, diabetes mellitus, hypertension, a preoperative blood transfusion, platelet count ≥ 400,000, and an open surgical approach.

On average, women who had a SSI compared with those who did not had longer operative times (by 30 minutes), greater estimated blood loss (by 125 mL), and a length of stay almost 2 days longer. Women with SSI had higher rates of most complications, with readmission rates 16 times higher and reoperation rates 11 times than those without SSI.

The women with deep SSI were analyzed. Although risk factors were largely similar to those for SSI, there were some differences. Women with deep SSI, compared with those without, were younger and more often had Medicaid insurance. Parity, BMI ≥ 30 kg/m ² , and ASA class ≥ 3 were similar. Gynecological cancer and preoperative blood transfusion were more prevalent in those with deep SSI, but diabetes mellitus, hypertension, tobacco use, history of DVT, and platelets ≥ 400,000 were not.

Women with deep SSI more frequently had an open hysterectomy, longer surgical time (by 30 minutes), and greater estimated blood loss (by 100 mL). An increased length of stay was also observed among those with deep SSI, and the rates of readmission and reoperation were 24 times and 18 times greater than those who did not have a deep SSI, respectively.

Multivariable hierarchical logistic regression was used to develop models to identify independent risk factors for SSI and deep SSI ( Tables 2 and 3 ). Adjustment was made for age in both models. In the final model for overall SSI, risk factors included the following: BMI ≥ 30 kg/m ² , ASA class ≥ 3, Medicaid insurance, gynecological cancer, preoperative blood transfusion, longer surgical time, and open hysterectomy ( Table 2 ).

Pearson’s residuals for SSI were 0.87, indicating strong fit, and the C statistic was 0.762. In the final model for deep SSI, risk factors with an independent association were as follows: age < 50 years, gynecological cancer, longer surgical time, and open hysterectomy (vs laparoscopic) ( Table 3 ). Pearson’s residuals were 0.83 and the C statistic was 0.763.

A subgroup analysis was undertaken to determine the effect of surgical approach on the final model for deep SSI. The estimates and odds ratios when surgical approach is included and excluded from the model are provided in Table 4 . Age, surgical time, and gynecological cancer are significantly associated with deep SSI in both versions.

Table 4

Comparison of hierarchical logistic regression models with and without surgical approach

Factors	With surgical approach		Without surgical approach
	Estimate	Odds ratio (95% CI)	Estimate	Odds ratio (95% CI)
Age < 50 y	-0.52	0.59 (0.41, 0.86)	–0.50	0.61 (0.42-0.89)
Surgical time (per hour)	0.23	1.25 (1.10, 1.42)	0.20	1.22 (1.07, 1.39)
Gynecological cancer	0.59	1.80 (1.13, 2.86)	0.68	2.0 (1.25, 3.11)
Any transfusion	0.48	1.61 (0.86, 3.03)	0.74	2.1 (1.14, 3.9)
Vaginal hysterectomy (referent laparoscopic)	0.04	1.04 (0.60, 1.8)	Not included	Not included
Open hysterectomy (referent laparoscopic)	0.64	1.90 (1.32, 2.70)	Not included	Not included
Model fit
Pearson’s residuals	0.82		0.85
C statistic	0.763		0.747

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. New development of the yolk sac theory in diabetic embryopathy: molecular mechanism and link to structural birth defects
2. Long-term neurofunctional outcome, executive functioning, and behavioral adaptive skills following fetal myelomeningocele surgery
3. Surgical management of vulvovaginal agglutination due to lichen planus
4. Reply

Stay updated, free articles. Join our Telegram channel

Join