Surgical site infections (SSI) occur 4–25% of patients after major surgical procedures and are estimated to cause over 8,000 deaths annually (Klevens et al., 2007). The occurrence of an SSI is increasingly being recognized to be a preventable harm and, as a result, SSI rates are being used as a surrogate measure healthcare quality. SSIs are also costly. The associated increased length of hospital stay, wound care supplies, and home care needs are estimated to cost $10.07 billion annually (Scott, 2009). Despite many efforts to reduce these events and associated harm and costs, sustained reductions in SSI rates have not been achieved.
One strategy that has been highly effective in improving quality on a large scale is public reporting (Fung, Lim, Mattke, Damberg, & Shekelle, 2008; Jha, Orav, Li, & Epstein, 2007). Benchmarking outcomes has been demonstrated to have a global improvement effect with surgical outcomes (Hall et al., 2009) and public reporting has been shown to improve cardiac surgery outcomes in New York State. In the 4 years after implementation of public reporting, the State of New York saw mortality from coronary artery bypass surgery decrease by 41% (Hannan, Kilburn, Racz, Shields, & Chassin, 1994). However, despite the evidence supporting the dramatic benefits of benchmarking and public reporting in surgical outcomes, there is no national standard for SSI. We designed a study to review current state legislation on SSI monitoring and public reporting.
- Top of page
Twenty-one states (42%) have legislation regarding monitoring and reporting of SSIs. All 21 states have legislation that requires monitoring of SSIs. Twenty (40%) of the states require SSI reporting and one state (Arkansas) has legislation encouraging voluntary reporting. The District of Columbia does not have any monitoring regulations. The remaining 29 states (58%) had no laws specifying SSI monitoring and reporting.
Data are disseminated in 15 states by Annual Reports, 18 states use their State Agency Websites, and 12 states use both. Additionally, three states also provide quarterly or semiannual reports (see Table 1).
Table 1. Data Collection, Reporting, and Availability by State of Surgical Site Infections
| || ||Data Reporting Mechanism|| |
| ||Data Collection||Annual||Posting to|| ||Is Data Currently|
|State||Entity||report||state website||Other||Publicly Available?|
|AL||NHSN||✓|| || ||No|
|CA||NHSN|| ||✓|| ||No|
|MA||NHSN|| ||✓|| ||Yes|
|NH||NHSN|| ||✓|| ||No|
|NJ||NHSN|| ||✓|| ||No|
|OH||NHSN|| ||✓|| ||Yes|
|OK||NHSN||✓|| || ||No|
|OR||NHSN||✓|| ||Quarterly report||Yes|
|RI||State agency||✓||✓|| ||No|
|VT||NHSN|| ||✓|| ||Yes|
Of the eight states with SSI public data currently available, a total of 10 procedures are reported. Seven (88%) states report coronary artery bypass graft procedures, six (75%) states report on hip replacement procedures, and six (75%) report on knee replacement procedures. Four states (50%) report hysterectomy, two states (25%) report colon surgery, and two states (25%) report herniorrhaphy. One state (13%) reports on breast surgery procedures, one state (13%) reported on gallbladder procedures, one state (13%) reports on cesarean section procedures, and one state (14%) reports on spinal fusion (see Table 2).
Table 2. Procedures Publicly Reported by States of Surgical Site Infections
|Procedure (Number of States Reporting, %)||SC||MO||CO||MA||NY||OH||VT||OR|
|Coronary artery bypass graft (n = 7, 88%)||✓||✓||✓||✓||✓||✓|| ||✓|
|Hip replacement (n = 6, 75%)||✓||✓||✓||✓||✓|| ||✓|| |
|Knee replacement (n = 6, 75%)||✓|| ||✓||✓|| ||✓||✓||✓|
|Hysterectomy (n = 4, 50%)||✓||✓|| ||✓|| || ||✓|| |
|Herniorrhaphy (n = 2, 25%)|| ||✓||✓|| || || || || |
|Colon surgery (n = 2, 25%)||✓|| || || ||✓|| || || |
|Breast surgery (n = 1, 13%)|| ||✓|| || || || || || |
|Gallbladder (n = 1, 13%)||✓|| || || || || || || |
|Cesarean section (n = 1, 13%)|| || || || || ||✓|| || |
|Spinal fusion (n = 1, 13%)||✓|| || || || || || || |
Among the eight states with currently available data, a range of two to seven procedures are reported. South Carolina reports seven procedures, Missouri reports five procedures, Colorado and Massachusetts report four procedures, New York, Ohio, and Vermont report three procedures, and Oregon reports two procedures. The average time lag between collection and publication was 6 months, with a range of 2–11 months.
Of the 21 states with legislation, 20 (95%) report using the Center for Disease Control's National Healthcare Safety Network (NHSN) and one state reports to its own state agency. Public SSI data reports are required in all 21 states with SSI legislation; eight (38%) of these states currently have public data available (see Figure 1).
- Top of page
This study describes the variation in state reporting of SSI rates. We found legislation in 21 of 50 states requiring reporting of SSI rates and currently public data available for eight states. The type of data available was varied with states reporting anywhere from two to seven procedures. Interestingly, colon surgery that has the highest rates of SSIs nationally was only reported by two states. And gallbladder surgery, which is among the most common surgical procedures a patient undergoes, was only reported by one state. The motivation to monitor and report certain procedures over others is unclear, and further highlights the variability between states. Without the same quality and type of data, it is difficult for consumers, payers, or regulators to compare infections within or across states, potentially making inaccurate inferences about the quality of care (Pronovost, Miller, & Wachter, 2007).
While the most striking variation in reporting is in whether or not a state reports at all, even within the reporting variation exists. In order for the consumer to interpret publicly reported SSI rates, it is imperative that the data be collected and reported in a standardized manner. For example, for colorectal SSIs, in-hospital monitoring versus monitoring with 30-day follow-up results in nearly a 40% discrepancy in SSI rates as many infections are diagnosed postdischarge (Smith et al., 2004). Furthermore, certain patient factors not amenable to change are associated with significantly increased risk of SSIs (Wick, Vogel, Church, Remzi, & Fazio, 2009). Thus, hospital patient mix will impact SSI rates and risk adjustment will be needed to make meaningful comparisons. Two of the most prevalent programs to calculate risk adjusted SSI rates are the CDC-NHSN and the American College of Surgeons National Surgical Quality Improvement Program (NSQIP). Although they use the same definitions for SSIs, the risk-adjustment techniques are markedly different and can consequently result in different expected rates for SSIs. In addition, numerous private sector groups have developed programs (e.g., ArborMix, Ann Arbor, Michigan) that will provide yet another set of metrics to follow surgical quality. Thus, as we move to mandated public reporting, it will be important to use a common program or at least ensure common parameters, inputs, and definitions that are used.
Measuring SSI rates is labor intensive but benchmarking these data represent one common-sense strategy to quality improvement in healthcare (Edwards et al., 2009). In New York State, legislating mandatory public reporting of cardiac surgery mortality rates resulted in an overall reduction in mortality rates. There was continued improvement with each subsequent year of public reporting. One explanation is that the public was more empowered with the data to make informed decisions, pressuring hospitals to focus on improving quality. Similarly, participation in benchmarking of hospital data within a large collaborative has been shown to improve surgical outcomes (Hall et al., 2009). Despite the dramatic impact of these monitoring practices, hospitals remain reluctant to publicly disclose their outcome data when it is not required. Legislation requiring reporting would motivate hospitals to re-evaluate their practices in SSI prevention and address internal outliers.
While we believe national standards would best facilitate improved reporting practices, other alternatives exist. At the very least, agreed upon definitions for a surgical site infection would be helpful so that the term is being used with universal understanding as it relates to policy and quality improvement. This could be facilitated, in part, by soliciting input from physician groups and societies. Payment incentives from payers with appropriate definitions and expectations may also be helpful to motivate accurate and timely reporting. Finally, if these efforts cannot be coordinated at a national level, multistate collaboratives may be a useful stepping stone.
Our study highlights the need for the Federal Government to set the rules for how hospitals define, monitor, and report SSIs. Though SSI process measures have received tremendous attention from the Centers of Medicare and Medicaid Services (CMS), the public, and the media, studies have failed to demonstrate that adherence to these measures reduce SSIs (Hawn et al., 2008; Nguyen, Yegiyants, Kaloostian, Abbas, & Difronzo, 2008; Stulberg et al., 2010). In fact, focusing on process measures rather than outcomes may divert important resources that could improve quality (Makary & Ibrahim, 2010). Accordingly, CMS recently announced that hospitals will have to begin reporting SSIs rates for select procedures beginning in January 2012 to impact payment for fiscal year 2014 (Hanshaw, Olson, & Gillespie, 2010). It is critical to standardize the reporting process before SSIs are incorporated into the Medicare payment scheme.
This study has some important limitations. First, we searched only state health department websites, thus data may be publicly available that we did not consider (such as on individual hospital websites). We recognize that not all publicly available SSI data are present in this manuscript; however, a component to consumer-friendly data is easy access. Nevertheless, the few number of states that have publicly available data further advance our thesis that data are not readily accessible. Second, we conducted our review during a single month, and we do not know what legislation might be forthcoming. It is likely that states are preparing for SSI reporting in light of recent CMS announcements (Hanshaw, Olson, & Gillespie, 2010), however national guidelines and definitions for collection and reporting are not in place. Third, we did not validate the accuracy of the web reporting. Yet, the intent of making SSI data publicly available is so consumers may use it when selecting a provider or assessing improvement efforts within a hospital, thus we assumed accuracy of the data.
We applaud efforts to promote SSI public reporting. We believe that there is potential to mitigate this form of preventable harm through these standardized efforts. At minimum it represents a novel strategy to decrease SSI in the context of many other previously unsuccessful efforts. Yet these policy provisions do not go far enough. We encourage CMS to build a rigorous process to validate SSI data or for CMS to require participation in the ACS-NSQIP or the CDC-NHSN so that appropriate definitions and risk adjustment can be used for monitoring. Only when SSI data are uniformly collected, reported, clinically validated, and made transparent will such a choice be possible. SSI can provide a model for monitoring and reducing other types of preventable harm.