Appendix A

SSI SIRs (Superficial and Deep Tissue)

As mentioned in the Methods and Data Chapter of this report, the NHSN protocol defines two different definitions of SSIs, which are referred to as All and Complex. This appendix contains metrics using the All definition, which includes both superficial and deep tissue SSIs. For a more detailed description of the All case definition and adjustment models, please refer to the SIR Guide Supplement (Centers for Disease Control and Prevention 2017b).

State SIRs by Surgery Type

Table A.1 shows the number of reported SSIs, number of predicted infections and associated SIRs for each surgical procedure. Note that the number of surgeries performed is slightly different than those show in Table 12 in Chapter 4. As mentioned in the Methods and Data Chapter, surgeries were excluded from analysis if information on one of the risk factors was missing. Because the All and Complex models contain different risk factors and different risk factors have different completion rates, the number of procedures excluded differs also. SSIs that occurred among all seven monitored surgical procedures are lower than predicted when evaluated at the state level, and four are statistically lower than the 2015 national experience. The two surgical procedures with SIRs closest to similar national hospitals are knee replacements and hysterectomies (SIR = 0.95 and 0.93 respectively).

Table A1: State Superficial and Deep Tissue SSI SIRs for each of 7 Procedures | Pennsylvania 2016

Type of Surgical Procedure Number of Hospitals Number of Surgical Procedures Performed Percent of Surgeries performed at ACH Number of Infections Number of Predicted Infections Standardized Infection Ratio (SIR) 95% Confidence Interval
Cardiac 71 8,444 92.25% 55 93.69 0.59 (0.4, 0.8)
CBG 2 incisions 62 7,866 100% 95 142.08 0.67 (0.5, 0.8)
CBG 1 incision 55 1,449 100% 20 24.39 0.82 (0.5, 1.2)
Colon 156 16,229 97.85% 577 848.47 0.68 (0.6, 0.7)
Hip repl. 158 28,033 98.90% 281 302.37 0.93 (0.8, 1.0)
Abd. hyst. 139 12,806 99.59% 169 217.23 0.78 (0.7, 0.9)
Knee repl. 158 44,072 98.52% 238 250.40 0.95 (0.8, 1.1)

Hospital Type

Over 92 percent of each surgical procedure type was performed at an acute care hospital, and all CBGB and CBGC surgeries were only performed in those hospitals. The limited number of surgical procedures performed in critical access and children’s hospitals prohibits a large evaluation of hospital type and SSI-specific SIRs. However, five surgical type and hospital type combinations had large numbers and are presented in Table 13 below. All the other combinations had less than one predicted infection and zero infections. SIRs in CAHs for colon surgery and hip and knee replacement were all close to 2.0 (Table A2), although not statistically significant. These are strikingly higher than those same surgeries, which were predominately performed at acute care hospitals (Table A1, 0.68, 0.93 and 0.95). Cardiac surgery performed in the three children’s hospitals is at the predicted level (SIR = 1.02) compared to similar national hospitals, whereas that from acute care hospitals is substantially lower (SIR = 0.59). Colon surgery performed in children’s hospitals has a much lower SIR (SIR = 0.31) when compared to SIRs from all Pa. hospitals combined and other similar national hospitals.

Table A2: List of Several Hospital-Type Surgical-Procedure Combinations Commonly Performed and their SIRs | Pennsylvania 2016

Type of Surgical Procedure Hospital Type Number of Hospitals Number of Surgeries Number of Infections (SSI) Number of Predicted Infections Standardized Infection Ratio (SIR) 95% Confidence Interval
Colon Critical Access 7 63 6 3.06 1.96 (0.8, 4.1)
Hip repl. Critical Access 7 297 4 2.01 1.99 (0.6, 4.8)
Knee repl. Critical Access 7 653 6 2.81 2.14 (0.9, 4.4)
Cardiac Children’s 3 654 8 7.87 1.02 (0.5, 1.9)
Colon Children’s 3 286 4 13.00 0.31 (0.1, 0.7)

Hospital Type - Distribution

The distribution of 2016 and 2015-2016 SIRs for the seven surgery types is presented in Figure A.1. Because the SIR was calculated in those hospitals in which the predicted number of infections was one or more for either timeframe, 417 hospital procedure pairs are included. This includes only five critical access hospitals and five children’s hospital SIRs. The remaining SIRs are from acute care hospitals. The distribution of 2016 or 2015-2016 SIRs for colon surgery, hip and knee replacement, and hysterectomy SSIs look similar. These four have a substantial proportion of SIRS in the SIR category that includes 0.0, with an equal or slightly smaller number of SIRs located in the two categories just below SIR of 1.0. In SIR categories larger than 1.0, the number of hospitals with SIRs decreases proportionally, but all have SIR outlier values between two to 10 categories beyond the second highest cluster of elevated SIRs. Those SIRs range in value from 4.2 to 7.9. The distribution of cardiac procedures has a different shape because fewer hospitals perform these surgical procedures and have more than one predicted SSI. All three distributions have between 30-40 percent of SIRs of 0.0. The remaining SIR values were somewhat evenly spread out between the SIR category just above the one that contains 0.0 and the category that contains 2.0. All three have the largest SIRs in the category that contains 2.7.

Figure A.1: Distribution of Hospital Specific SIRs stratified by Surgical Procedure | Pennsylvania 2016

Table A3, below, shows the number of hospitals with statistically significant SSIs. More hospitals have SIRs that are significantly lower than predicted by the 2015 national experience than those with SIRs significantly more than the 2015 national experience.

*Table A.3: Number of Hospitals with Statistically Significant SSI SIRs | Pennsylvania 2015-2016 and 2016**

Surgical procedure SIR less than 1.0 SIR more than 1.0
Cardiac 5 0
Cardiac bypass graft 2 incisions 4 2
Colon 27 4
Hip replacement 3 6
Abdominal hysterectomy 3 3
Knee replacement 2 9

Hospital List

Below is the comprehensive list of all hospital-surgical procedure pairs in which one or more of the selected surgeries was performed during 2016 (N=794). The SIR was calculated for all hospitals and surgeries if the predicted number of infections was more than one in 2016 (N=417) or during years 2015 and 2016 (N=377). The SIR based upon two years of data was calculated for hospitals in which the predicted number of infections was less than one during 2016.

Despite the efforts to calculate a SIR for hospitals with less than one predicted infection in 2016 by using two years of data, a SIR could not be calculated for 235 (29.6 percent) of the 794 hospital-surgery pairs. In 79 (33 percent) of those pairs, one or more SSIs occurred even though less than one was predicted.

Table A4: List of 157 Hospitals with Superficial and Deep Tissue SSI SIRs and Associated Measures | Pennsylvania 2016 and 2015-2016

Hospitals printed in red have statistically significantly more SSIs than predicted. Hospitals printed in blue have statistically significantly fewer SSIs than predicted.