Objective

National validation of claims-based surveillance for surgical-site infections (SSIs) following colon surgery and abdominal hysterectomy.

Design

Retrospective cohort study.

Setting

US hospitals selected for data validation by Centers for Medicare & Medicaid Services (CMS).

Participants

The study included 550 hospitals performing colon surgery and 458 hospitals performing abdominal hysterectomy in federal fiscal year 2013.

Methods

We requested 1,200 medical records from hospitals selected for validation as part of the CMS Hospital Inpatient Quality Reporting program. For colon surgery, we sampled 60% with a billing code suggestive of SSI during their index admission and/or readmission within 30 days and 40% who were readmitted without one of these codes. For abdominal hysterectomy, we included all patients with an SSI code during their index admission, all patients readmitted within 30 days, and a sample of those with a prolonged surgical admission (length of stay > 7 days). We calculated sensitivity and positive predictive value for the different groups.

Results

We identified 142 colon-surgery SSIs (46 superficial SSIs and 96 deep and organ-space SSIs) and 127 abdominal-hysterectomy SSIs (58 superficial SSIs and 69 deep and organ-space SSIs). Extrapolating to the full CMS data validation cohort, we estimated an SSI rate of 8.3% for colon surgery and 3.0% for abdominal hysterectomy. Our colon-surgery surveillance codes identified 93% of SSIs, with 1 SSI identified for every 2.6 patients reviewed. Our abdominal-hysterectomy surveillance codes identified 73% of SSIs, with 1 SSI identified for every 1.6 patients reviewed.

Conclusions

Using claims to target record review for SSI validation performed well in a national sample.

Background: Recent efforts to study quality improvement (QI) efforts to improve the effectiveness and efficiency of healthcare have raised important questions about ethical boundaries for waiving informed consent. Confusion exists because similar projects can be undertaken for research or QI purposes, a distinction currently used to define Institutional Review Board oversight. However, patients are not aware of such distinctions. We sought to evaluate patients' views of waiving consent for non-invasive projects to improve healthcare quality and delivery. Methods: We developed a 32-item measure of patient thresholds for waiving consent for different types of QI interventions, including those involving changes to: 1) the hospital environment; 2) hospital policies or procedures; 3) objects used by patients; 4) medications or devices; and 5) use of patient information. In a sample of 200 hospitalized patients, we tested and confirmed the reliability and validity of subscales representing each of the 5 intervention types. Results: For each of the five consent threshold scales, all items in each scale had substantial item-to-total correlations with the other items in that scale taken together. All five internal consistency reliability coefficients exceeded.70 for the total sample. Means for all 5 scales indicated general patient support for waiving consent across all categories of interventions studied. However, patients were significantly less comfortable foregoing consent for interventions involving medications or devices, or sharing of patient information. Conclusion: We developed and tested a survey instrument to contribute to the understanding of patient preferences for consent in QI assessment activities. Measures were found to be reliable and valid. Findings indicated general patient support for waiving consent across all minimal risk categories studied. However, more work is needed to reassure and protect patients during minimal risk studies involving medications or devices, or the sharing of patient information.

Background

Surgical site infections (SSIs) following vascular surgery have high morbidity and costs, and are increasingly tracked as hospital quality measures.

Objective

To assess the ability of Medicare claims to identify US hospitals with high SSI rates after vascular surgery.

Research design

Using claims from fee-for-service Medicare enrollees of age 65 years and older who underwent vascular surgery from 2005 to 2008, we derived hospital rankings using previously validated codes suggestive of SSI, with individual-level adjustment for age, sex, and comorbidities. We then obtained medical records for validation of SSI from hospitals ranked in the best and worst deciles of performance, and used logistic regression to calculate the risk-adjusted odds of developing an SSI in worst-decile versus best-decile hospitals.

Results

Among 203,023 Medicare patients who underwent vascular surgery at 2512 US hospitals, a patient undergoing surgery in a hospital ranked in the worst-performing decile based on claims had 2.5 times higher odds of developing a chart-confirmed SSI relative to a patient with the same age, sex, and comorbidities in a hospital ranked in the best-performing decile (95% confidence interval, 2.0-3.1). SSI confirmation among patients with claims suggesting infection was similar across deciles, and we found similar findings in analyses limited to deep and organ/space SSIs. We report on diagnosis codes with high sensitivity for identifying deep and organ/space SSI, with one-to-one mapping to ICD-10-CM codes.

Conclusions

Claims-based surveillance offers a standardized and objective methodology that can be used to improve SSI surveillance and to validate hospitals' publicly reported data.

Objective

To assess the ability of Medicare claims to identify US hospitals with high rates of surgical site infection (SSI) after hip arthroplasty.

Design

Retrospective cohort study.

Setting

Acute care US hospitals.

Participants

Fee-for-service Medicare patients 65 years of age and older who underwent hip arthroplasty in US hospitals from 2005 through 2007.

Methods

Hospital rankings were derived from claims codes suggestive of SSI, adjusted for age, sex, and comorbidities, while using generalized linear mixed models to account for hospital volume. Medical records were obtained for validation of infection on a random sample of patients from hospitals ranked in the best and worst deciles of performance. We then calculated the risk-adjusted odds of developing a chart-confirmed SSI after hip arthroplasty in hospitals ranked by claims into worst- versus best-performing deciles.

Results

Among 524,892 eligible Medicare patients who underwent hip arthroplasty at 3,296 US hospitals, a patient who underwent surgery in a hospital ranked in the worst-performing decile based on claims-based evidence of SSI had 2.9-fold higher odds of developing a chart-confirmed SSI relative to a patient with the same age, sex, and comorbidities in a hospital ranked in the best-performing decile (95% confidence interval, 2.2-3.7).

Conclusions

Medicare claims successfully distinguished between hospitals with high and low SSI rates following hip arthroplasty. These claims can identify potential outlier hospitals that merit further evaluation. This strategy can also be used to validate the completeness of public reporting of SSI.

We assessed 4045 ambulatory surgery patients for surgical site infection (SSI) using claims-based triggers for medical chart review. Of 98 patients flagged by codes suggestive of SSI, 35 had confirmed SSIs. SSI rates ranged from 0 to 3.2% for common procedures. Claims may be useful for SSI surveillance following ambulatory surgery.

Timely identification of outbreaks of hospital-associated infections is needed to implement control measures and minimize impact. Survey results from 33 hospitals indicated that most hospitals lacked a formal cluster definition and all targeted a very limited group of prespecified pathogens. Standardized, statistically based outbreak detection could greatly improve current practice.

Whether targeted or universal decolonization strategies for the control of methicillin-resistant Staphylococcus aureus (MRSA) select for resistance to decolonizing agents is unresolved. The REDUCE-MRSA trial (ClinicalTrials registration no. NCT00980980) provided an opportunity to investigate this question. REDUCE-MRSA was a 3-arm, cluster-randomized trial of either screening and isolation without decolonization, targeted decolonization with chlorhexidine and mupirocin, or universal decolonization without screening to prevent MRSA infection in intensive-care unit (ICU) patients. Isolates from the baseline and intervention periods were collected and tested for susceptibility to chlorhexidine gluconate (CHG) by microtiter dilution; mupirocin susceptibility was tested by Etest. The presence of the qacA or qacB gene was determined by PCR and DNA sequence analysis. A total of 3,173 isolates were analyzed; 2 were nonsusceptible to CHG (MICs, 8 μg/ml), and 5/814 (0.6%) carried qacA or qacB At baseline, 7.1% of MRSA isolates expressed low-level mupirocin resistance, and 7.5% expressed high-level mupirocin resistance. In a mixed-effects generalized logistic regression model, the odds of mupirocin resistance among clinical MRSA isolates or MRSA isolates acquired in an ICU in intervention versus baseline periods did not differ across arms, although estimates were imprecise due to small numbers. Reduced susceptibility to chlorhexidine and carriage of qacA or qacB were rare among MRSA isolates in the REDUCE-MRSA trial. The odds of mupirocin resistance were no different in the intervention versus baseline periods across arms, but the confidence limits were broad, and the results should be interpreted with caution.