Chapter 4 Results

In Pennsylvania, 314 hospitals completed a survey to the healthcare-associated infections (HAI) reporting database, NHSN. For reporting purposes, the definition of hospitals includes large acute care hospitals, critical access hospitals that typically have no more than 25 beds, long-term acute care facilities where patients typically stay over 10 days and require hospital level care, and specialty hospitals such as women’s, children’s and psychiatric. This also includes inpatient rehabilitation hospitals that consist of both free-standing buildings and wards within large acute care hospitals.

Table 1 shows the number of hospitals used in the analysis stratified by the hospital type that were open for every month of 2016 for each HAI and corresponding SUR metric. One hospital that submitted central line-associated blood stream infection (CLABSI) data did not report catheter-associated urinary tract infection (CAUTI) data.

Ten hospitals indicated that they would report days in which a patient had a urinary catheter, but zero days were reported. These acute care hospitals were included in the urinary catheter standardized utilization ration (SUR) analyses and excluded from CAUTI analyses. The sum of predicted device days from these ten hospitals was 2,087.8 and represented 0.14 percent of all predicted acute care hospital urinary catheter days.

Likewise, eleven hospitals elected to report days of central line use, but zero days were reported. These hospitals were included in the SUR analysis and excluded from the CLABSI analysis. They were a mix of acute care, critical access and psychiatric hospitals.

Table 1: Number of Hospitals Used to Calculate HAI SIRs and SURs Stratified by Hospital Type | Pennsylvania 2016

4.1 CAUTI’s

4.1.1 SIR

In 2016, 1,658 CAUTIs were reported from 285 hospitals. This is 35 fewer CAUTIs than predicted using national baseline data (SIR: 0.98; 95% CI:0.933-1.027). Statistically, this was neither better nor worse than the 2015 national experience.

4.1.1.1 Hospital Type

As seen below in Table 2, the clear majority of CAUTIs occurred in ACHs (1,427 of 1,658). Overall, the ACHs performed 6 percent better than hospitals from the 2015 national experience (SIR 0.94, 95% CI: 0.9-1.0). On the other hand, CAHs, LTACs and IRFs had 91 percent, 17 percent and 50 percent respectively, more infections than predicted by the 2015 national experience. The SIR from children’s (SIR 0.83 95% CI 0.5-1.3) and psychiatric hospitals (SIR 1.37 95% CI 0.4-3.3) indicated a similar experience to the 2015 national experience.

Table 2: Summary of CAUTI Measures Stratified by Hospital Type | Pennsylvania, 2016

Type of Hospital Number of Hospitals Number of Urinary Catheter Days Number of CAUTIs Number of Predicted Infections Standardized Infection Ratio(SIR) 95% Confidence Interval
Acute care 157 1,288,709 1,427 1,513.32 0.94 (0.9, 1.0)
Critical access 15 9,648 12 6.27 1.91 (1.0, 3.3)
Children’s 5 14,906 17 20.57 0.83 (0.5, 1.3)
Long-term acute care 22 64,805 97 82.98 1.17 (1.0, 1.4)
Psychiatric 8 4,691 4 2.93 1.37 (0.4, 3.3)
Inpatient rehabilitation 78 45,881 101 67.12 1.50 (1.2, 1.8)

4.1.1.2 Hospital Type — Distribution

Figure 1 displays the distribution of hospital specific SIRs (N=204) grouped by hospital type. This includes 204 hospitals in which a SIR calculated with 2016 data and 43 hospitals with 2015 and 2016 data combined. Despite the inclusion of two years of surveillance data, 81 (28 percent) hospitals had less than one predicted infection and relatively few urinary catheter days. CAUTIs occurred in 18 (22 percent) of those hospitals and SIRs were not calculated due to methodological reasons.

Figure 1: Distribution of CAUTI SIRs by Hospital Type | Pennsylvania 2015-2016 and 2016

4.1.1.3 Hospital List

Table 3 lists the 285 hospitals and relevant CAUTI data, as well as county location of hospital and year(s) from which the data were based. The SIRs can be compared between hospitals with one and two years of data. Thirty-two hospitals (15 percent) had SIRs of zero, meaning that no CLABSIs occurred and more than one was predicted. The corresponding predicted CAUTIs ranged between one and 6.22.

Eleven hospitals (identified by blue font) had statistically significant SIRs that were less than 1.0, meaning that they had fewer CAUTIs than in the 2015 national experience. Three of these hospitals had no infections, although the predicted number of infections ranged between 4.4 and 6.2. Among the remaining eight hospitals, where at least one CAUTI occurred, the range of predicted CAUTIs was 5.2 to 26.7.

Twenty-one hospitals (identified by red font) had statistically significant SIRs that were more than 1.0, meaning that they had more CAUTIs than in the 2015 national experience. Seventeen of these hospitals had less than five predicted CAUTIs.

Table 3: List of Hospitals with CAUTI SIRs and Associated Measures | Pennsylvania 2015-2016 and 2016

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

4.1.2 Urinary Catheter SUR

Urinary catheter SUR was calculated for 295 hospitals. Ten hospitals reported no urinary catheter days, 41 hospitals reported between 19 and 195 urinary catheter days, 72 reported between 202 and 986 urinary catheter days, 86 reported between 1,001 and 4,745 urinary catheter days, 40 hospitals reported between 5,138 and 9,866 urinary catheter days, and the remaining 46 reported between 10,010 and 58,469 urinary catheter days.

4.1.2.1 Hospital Type

Table 4 lists summary statistics regarding urinary catheter use in different types of hospitals. Psychiatric hospitals in Pennsylvania, as a group, use urinary catheters 2.77 times more often than hospitals from the 2015 national experience (95% CI: 2.7-2.8). SURs for LTAC hospitals (SUR 0.8 95% CI: 0.8-0.8), IRFs (SUR 0.8 95% CI: 0.8-0.8), children’s hospitals (SUR 0.9 95% CI: 0.8-0.9), and ACHs (SUR 0.9 95% CI: 0.9-0.9) use urinary catheters less frequently in Pennsylvania than those from the 2015 national experience. CAHs use urinary catheters as often as hospitals from the 2015 national experience (SUR 1.0 95% CI: 1.0-1.0). When data from 40 hospitals with between 6 and 195 predicted urinary catheter days was combined, the SUR was 2.26, meaning that these hospitals used urinary catheters more than twice those in the 2015 national experience. Hospitals with more than 216 predicted urinary catheter days on average had SURs less than those from the 2015 national experience.

Table 4: Urinary Catheters SURs by Hospital Type | Pennsylvania, 2016

Type of Hospital Number of Hospitals Number of Predicted Device Days Number of Urinary Catheter Days Standardized Utilization Ratio (SUR) 95% Confidence Interval
Acute care 162 1,441,688.6 1,288,709 0.9 (0.9, 0.9)
Critical access 15 9,748.3 9,645 1.0 (1.0, 1.0)
Children’s 6 17,479.1 14,906 0.9 (0.8, 0.9)
Long-term acute care 22 84,481.0 64,805 0.8 (0.8, 0.8)
Psychiatric 12 1,694.6 4,691 2.8 (2.7, 2.8)
Inpatient rehabilitation 78 53,806.3 45,881 0.9 (0.8, 0.9)

4.1.2.2 Distribution by Hospital Type

Figure 2 shows that the hospital specific urinary catheter SUR values were quite dispersed and ranged from 0 to 19.5. Two hundred ninety-four hospitals had a calculated SUR. Three psychiatric hospitals and one ACH had SURs over 10 (SUR: 19.52, 15.11,12.9, and 12.75), meaning that urinary catheters were used more than 10 times that of similar hospitals sampled from the 2015 national experience. The distribution of SURs in psychiatric hospitals is unique because 33 percent (N=4) had no urinary catheter use while 33 percent (N=4) used urinary catheters between 4.62 and 19.52 times more than the 2015 national experience.

Figure 2: Distribution of Urinary Catheter SURs Stratified by Type of Hospital | Pennsylvania 2016

4.1.2.3 Facility List

Examination of the SURs in Table 5 shows that seven of ten hospitals with the highest ten SURs had fewer than 200 predicted urinary catheter days. The other three had between 235 and 1,952. Of the 10 hospitals that reported no urinary catheter days, eight had fewer than 200 predicted urinary catheter days.

Table 5: List of Urinary Catheter SURs by Hospital Name | Pennsylvania 2016

Hospitals printed in red have statistically significantly more urinary catheter use than predicted. Hospitals printed in blue have statistically significantly less urinary catheter use than predicted.

4.2 CLABSIs

4.2.1 SIR

4.2.1.1 Hospital Type

As seen in Table 6, Pennsylvania patients in an inpatient rehabilitation hospital were 73 percent (SIR: 1.73, 95% CI: 1.1-2.5) more likely to have a CLABSI than those in the 2015 national experience. All other types of hospitals had SIRs with 95% CIs that included 1.0. This means that they had about the same number of CLABSIs as the 2015 national experience.

Table 6: Summary of CLABSI SIRs Stratified by Hospital Type | Pennsylvania 2016

Hospital Type # of Hospitals # of Central Line Days # of CLABSIs # of Predicted Infections SIR 95% CI
Acute care 152 1,223,704 1,121 1,221.57 0.92 (0.9, 1.0)
Critical access 14 5,463 3 1.52 1.98 (0.5, 5.4)
Children’s 6 107,664 169 145.41 1.16 (1.0, 1.3)
Long-term acute care 22 100,421 121 111.18 1.09 (0.9, 1.3)
Psychiatric 1 4 0 0.00 0.00 (NA, NA)
Inpatient rehabilitation 78 46,196 25 14.46 1.73 (1.1, 2.5)

4.2.1.2 Hospital Type - Distribution

Figure 3 displays the distribution of hospital specific CLABSI SIRs (N=156) grouped by hospital type. This includes 132 hospitals in which a SIR calculated with 2016 data and 24 hospitals with 2015 and 2016 data combined. Despite the inclusion of two years of surveillance data, 117 (43 percent) hospitals had less than one predicted infection and relatively few central line days. CLABSIs occurred in 28 (24 percent) of those hospitals and SIRs were not calculated due to methodologic reasons. Hospital specific SIRs were not calculated for psychiatric or critical access hospitals because the number of predicted CLABSIs was so low.

Figure 3: Distribution of CLABSI SIRs Stratified by Hospital Type | Pennsylvania 2015-2016 or 2016

4.2.1.3 Hospital List

Table 7 lists CLABSI data, as well as county location of hospital and year(s) from which the data were based from 273 hospitals. The SIRs can be compared between hospitals that have one year and two years of data.

Twenty-six hospitals had SIRs of zero even though the number of predicted CLABSIs ranged between 1.02 and 8.45. Twenty-one hospitals (identified by blue font) had statistically significant SIRs that were less than 1.0, meaning that they had fewer CLABSIs than in the 2015 national experience. Eight hospitals (identified by red font) had statistically significant SIRs that were more than 1.0, meaning that they had more CLABSIs than in the 2015 national experience.

Table 7: List of Hospitals with CLABSI SIR and Associated Measures |Pennsylvania 2015-2016 and 2016

Hospitals printed in red have statistically significantly more CLABSIs than predicted. Hospitals printed in blue have statistically significantly less CLABSIs than predicted.

4.2.2 Central Line SUR

Central line catheters were most frequently used in acute care hospitals (1,223,704 catheter days). Central lines were rarely used in psychiatric and critical access hospitals because those patients either typically have medical conditions that don’t need monitoring with a central line or are transferred to an acute care hospital after placement of a central line. Eleven hospitals reported no central line days, 52 reported between one and 197 central line days, 62 reported between 203 and 958 central line days, 42 reported between 5,089 and 9,788 central line days, and 32 reported between 10,327 and 84,640 central line days.

The SUR compares the number of days in which a patient had a central line inserted and the predicted number of days for similar hospitals. The state SUR was 0.90 which is 10 percent less use than predicted by the national experience. As a group, hospitals with predicted central line days between 2.1 and 192.7 had a SUR of 1.0, meaning that they used central line catheters as often as similar hospitals did in 2015.

4.2.2.1 Hospital Type

Table 8 lists summary statistics regarding central line use stratified by hospital type. Acute care hospitals use central line catheters less frequently than predicted (SUR: 0.89, 95% CI: 0.9-0.9), whereas critical access hospitals use it more often than predicted (SUR:1.17, 95% CI:1.1-1.2) by the 2015 national experience. Children’s hospitals and long-term acute care hospitals use central line catheters about as often as predicted by the 2015 national experience (SUR: 0.99, 95% CI: 1.0-1.0 and SUR= 1.02, 95% CI: 1.0-1.0, respectively). Psychiatric hospitals and inpatient rehabilitation hospitals use central line catheters much less than predicted (SUR = 0.01, 95% CI: 0.0-0.0 and 0.70, 95% CI: 0.7-0.7, respectively) by the 2015 national experience. Eleven hospitals reported no days of central line catheters and the predicted days of use ranged from two to 322.

Table 8: Central Line Catheters SURs by Hospital Type | Pennsylvania, 2016

Number of Facilities Number of Predicted Device Days Number of Central Line Catheter Days Standardized Utilization Ratio (SUR) 95% Confidence Interval
Acute care 159 1,358,341.32 1,208,614 0.89 (0.9, 0.9)
Critical access 15 4,679.49 5,463 1.17 (1.1, 1.2)
Children’s 6 83,931.94 83,065 0.99 (1.0, 1.0)
Long-term acute care 22 98,165.38 100,421 1.02 (1.0, 1.0)
Psychiatric 4 372.64 4 0.01 (0.0, 0.0)
Inpatient rehabilitation 78 66,284.03 46,196 0.70 (0.7, 0.7)

```

4.2.2.2 Distribution

The central line catheter SURs among 283 hospitals ranged from zero to 15.22. The second highest SUR was 3.4. Seventeen hospitals had central line SURs between zero and .09, and the predicted number of central lines among those hospitals ranged between 2.1 and 671. The facility level distributions of central line SURs are shown in Figure 4. Psychiatric hospitals were excluded because the number of device days was essentially zero.

Figure 4: Distribution of Central Line SUR by Hospital Type | Pennsylvania 2015-2016 and 2016

4.2.2.3 Facility List

The number of predicted central line catheter days in hospitals ranged from two to 68,554, and the observed use ranged from zero to 84,640 (Table 9). Examination of the hospitals with the highest 10 SURs shows that two had fewer than 200 predicted central line catheter days. The others had between 283 and 3,030. Of the 11 hospitals that reported no central line catheter days, 10 had less than 200 predicted urinary catheter days.

Table 9: List Hospitals with Central Line Catheters SURs |Pennsylvania 2016

Hospitals printed in red have statistically significantly more central line use than predicted. Hospitals printed in blue have statistically significantly less central line use than predicted.

4.3 SSIs

4.3.1 Description of Surgical Procedures

This 2016 report includes hospital acquired infections that occurred following seven types of surgery: three types of cardiac surgery, hip and knee replacements, abdominal hysterectomy, and colon surgery. Prior to applying the exclusion criteria, a total of 119,695 procedures were performed and followed for signs of a surgical site infection (SSI).

Of the 251 hospitals, 168 performed at least one type of surgery and 83 performed none. Table 10 shows that 55 hospitals perform all seven surgeries while 10 hospitals only performed knee and hip replacement surgeries. Table 11 shows the list of hospitals and the types of surgeries performed there in 2016.

Table 10: Number of Hospitals that Perform Combinations of the Seven Surgeries | Pennsylvania 2016

Number of Hospitals Colon Hip Replacement Knee Replacement Hysterectomy Cardiac Surgery CBG Surgery 2 incisions CBG Surgery 1 incision
63 X X X X
55 X X X X X X X
10 X X
9 X X X
7 X X X X X
6 X X X X X X
4 X X
3 X
3 X X X
2 X X
2 X X
1 X X X X
1 X X X
1 X X
1 X

Table 11: List of Hospitals that Perform Selected Surgeries | Pennsylvania 2016

4.3.2 SIRs

As mentioned in the Methods and Data Chapter of this report, the NHSN protocol sets forth two different methods to evaluate SSIs, which are referred to as All and Complex. Results presented in this section contain metrics using the Complex definition, which includes only deep tissue SSIs. (SSI results that use the All case definition are presented in Appendix A.) Superficial infections are not counted as cases in the Complex model. For a more detailed description of the Complex case definition and adjustment models, please refer to the SIR Guide Supplement (Centers for Disease Control and Prevention 2017b).

4.3.2.1 State SSI SIRs

Table 12 shows the number of surgeries performed and included in the analysis, number of subsequent SSIs, number of predicted infections and associated SIRs for each surgical procedure. Knee replacement surgeries are the most commonly performed surgeries reported in NHSN with over 44,000 performed during 2016.

SSI SIRs for all seven monitored surgical procedures range between 1.04 and 0.82. The state-wide SSI SIR for colon surgeries is statistically significantly lower than the 2015 national experience (SIR: 0.82, 95% CI: 0.7 - 0.9). SIRs for the other six surgical procedures are close to 1.0 and are not statistically different than the 2015 national experience. Coronary artery bypass grafting with both chest and donor site incisions (CBGB) and coronary artery bypass grafting with chest incision only (CBGC) surgeries were only performed in acute care hospitals.

Table 12: 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 8,413 92.74% 30 33.95 0.88 (0.6, 1.2)
CBG 2 incisions 62 7,938 100% 53 58.87 0.90 (0.7, 1.2)
CBG 1 incision 55 1,449 100% 8 9.94 0.81 (0.4, 1.5)
Colon 156 16,236 97.85% 350 424.33 0.82 (0.7, 0.9)
Hip repl. 158 28,172 98.91% 164 177.19 0.93 (0.8, 1.1)
Abd. hyst. 139 12,807 99.59% 98 94.22 1.04 (0.8, 1.3)
Knee repl. 158 44,602 98.53% 149 146.11 1.02 (0.9, 1.2)
  • As defined in the complex statistical model that includes only deep tissue SSIs

4.3.2.2 Hospital Type

Because nearly all surgical procedures were performed in acute care hospitals, calculation of SSI SIRs stratified by hospital type is very limited. Only five surgical type and hospital type pairs had more than one predicted SSI, and these are presented in Table 13 below.

Knee and colon surgeries that were performed in critical access hospitals had higher SIRs than the 2015 national experience, although they were not statistically significant. Cardiac and colon surgery performed at children’s hospitals had lower SIRs than the 2015 national experience, but they were not statistically significant. Hip replacement surgery performed at critical access hospitals had the same experience as the nation in 2015.

Table 13: List of several hospital-type surgical-procedure combinations commonly performed and the 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 Crit. acc. 7 63 2 1.27 1.58 (0.3, 5.2)
Hip repl. Crit. acc. 7 297 1 1.01 0.99 (0.0, 4.9)
Knee repl. Crit. acc. 7 653 3 1.32 2.27 (0.6, 6.2)
Cardiac Children’s 3 611 3 4.38 0.69 (0.2, 1.9)
Colon Children’s 3 286 3 5.94 0.51 (0.1, 1.4)

4.3.2.3 Hospital Type - Distribution

The distribution of 2016 and 2015-2016 SIRs for the seven surgery types is presented in Figure 5. Because the SIR was calculated in those hospitals in which the predicted number of infections was one or more for either timeframe, 401 hospital-procedure pairs are included. This includes only five hospital surgery pairs from children’s hospitals. 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 surgical procedures have a substantial number of SIRS equal to 0.0, with the second highest number of hospitals falling in the SIR category of 0.6-0.8. In SIR categories larger than 1.0, the number of hospitals with SIRs decreases proportionally, but all have SIR values above 2.0.

The distributions of cardiac procedures have a different shape because fewer hospitals perform these surgical procedures and have more than one predicted SSI.

Figure 5: Distribution of Specific SIRs Stratified by Surgical Procedure and Hospital Type | Pennsylvania 2015-2016 and 2016

Table 14, below, shows the number of hospitals with statistically significant SSIs. More hospitals have SIRs that are significantly higher than predicted by the 2015 national experience than those with SIRs significantly less than the 2015 national experience.

Table 14: List 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 bypass graft 2 incisions 0 3
Colon 5 4
Hip replacement 0 4
Abdominal hysterectomy 2 3
Knee replacement 0 10

4.3.2.4 Hospital List

Below is the comprehensive list of all hospital-surgical procedure pairs in which one or more of the surgeries was performed during 2016 (N=799). The SIR was calculated for 257 hospitals surgery pairs when the predicted number of infections was more than one in 2016. An additional 144 SIRs could be calculated when 2015 and 2016 data was combined. The remaining 398 hospital surgery pairs (49.8 percent of all pairs) still had less than one predicted infection and SIRs could not be calculated. In 83 (21 percent) of those pairs, one or more SSIs occurred even though less than one was predicted.

Table 15: List of Hospitals with 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.