surveillance of surgical site infections in nhs …...surveillance of surgical site infections in...

Surveillance of surgical site infections in NHS hospitals in England 2015/16

Surveillance of surgical site infections in NHS hospitals in England, 2015/16

2

About Public Health England

Public Health England exists to protect and improve the nation’s health and wellbeing,

and reduce health inequalities. We do this through world-class science, knowledge

and intelligence, advocacy, partnerships and the delivery of specialist public health

services. We are an executive agency of the Department of Health, and are a distinct

delivery organisation with operational autonomy to advise and support government,

local authorities and the NHS in a professionally independent manner.

Public Health England

133-155 Waterloo Road

Wellington House

London SE1 8UG

Tel: 020 7654 8000

http://www.gov.uk/phe

Twitter: @PHE_uk

Facebook: www.facebook.com/PublicHealthEngland

Suggested citation: Public Health England. Surveillance of surgical site infections in

NHS hospitals in England, 2015/16. London: Public Health England, December 2016.

Available from: www.gov.uk/phe

Prepared by: S Elgohari with the assistance of C Wloch, A Saei, A Charlett, P

Harrington and T Lamagni. We extend thanks to Professor A Johnson, head of the

Department of Healthcare Associated Infections and Antimicrobial Resistance for

comments. We are also grateful to the administrative staff at PHE’s Surgical Site

Infection Surveillance Service, staff within the PHE Software Development Unit and the

considerable contribution made by NHS trusts in England who have devoted time and

effort in collecting these data. Finally, many thanks to hospitals who shared their

experience of the SSI surveillance for inclusion in this report.

© Crown copyright 2016

You may re-use this information (excluding logos) free of charge in any format or medium,

under the terms of the Open Government Licence v3.0. To view this licence, visit OGL or

email [email protected]. Where we have identified any third party copyright

information you will need to obtain permission from the copyright holders concerned. Any

enquiries regarding this publication should be sent to: [email protected]

Published: December 2016

PHE gateway number: 2016506


https://twitter.com/PHE_uk

http://www.facebook.com/PublicHealthEngland


https://www.nationalarchives.gov.uk/doc/open-government-licence/version/2/

mailto:[email protected]

mailto:[email protected]


3

Contents

Main points 4

Section 1. Introduction and surveillance methods 5

1.1 Introduction 5

1.2 Data collection and feedback 5 1.3 Definitions 6 1.4 Participation in international surveillance 6 1.5 Analyses presented in this report 7

Section 2. Overview 8

2.1 Hospital participation and surgical volumes 8 2.2 Continuous surveillance 8 2.3 Data quality 9

Section 3. Cumulative incidence (%) of SSI 14

3.1 Incidence (%) of inpatient and readmission SSI by surgical category 14 3.2 Risk factors for SSI 18

Section 4. Trends in the incidence of SSI 21

4.1 Orthopaedic categories 21

4.2 Non-orthopaedic categories 22

Section 5. Characteristics of surgical site infections 26

5.1 Type of SSI 26

5.2 Causative micro-organisms 28

Section 6. Variation in SSI incidence between NHS hospitals 35

6.1 Box and whisker plots 35

6.2 Funnel plots 36

Section 7. Hospital perspectives 38

Section 8. Conclusions 40

8.1 Summary of findings in 2015/16 40 8.2 Variation in the SSI rate between participating hospitals 41

8.3 Optional post-discharge surveillance 41 8.4 Microbial aetiology 42

8.5 Future directions 43

Section 9. Glossary 44

Section 10. References 46


4

Main points in 2015/16, data on 136,872 procedures and 1,632 surgical site infections (SSIs)

detected during the inpatient stay or on readmission following the initial operation were

collected by 198 NHS hospitals and seven independent sector NHS treatment centres

for 17 surgical categories under surveillance

data completion for age, patient sex and duration of operation was high in 2015/16

(≥95% of submitted records) in the majority of surgical categories. Data completion for

ASA score (standardised pre-operative health score) and BMI continued to vary between

categories. BMI data completion of ≥50% was achieved in nine surgical categories in

2015/16 compared to six in 2014/15

the cumulative SSI incidence (data from April 2011 to March 2016) ranged from 9.8% in

large bowel surgery to <1% in hip and knee prosthesis depending on the inherent wound

contamination

a significant decrease in the SSI incidence occurred for repair of neck of femur from

1.7% in 2008/9 to 1.0% in 2015/16

the SSI incidence remained low for hip and knee prosthesis (<1%) with decreasing

trends observed for hip prosthesis in the last three successive years

nine NHS trusts were identified as high outliers for the mandatory orthopaedic categories

in 2015/16 with an SSI incidence higher than expected. Eight NHS trusts and an

additional NHS treatment centre were identified as low outliers. All 18 providers have

been contacted and asked to investigate possible reasons

trends in SSIs for the other categories (from 2008/9 to 2015/16) identified a significant

increase for spinal surgery from 1.4% in 2008/9 to 1.8% in 2015/16

a significantly decreasing SSI trend occurred in bile duct/liver/pancreatic surgery, from

9.6% in 2008/9 to 5.2% in 2015/16. In gastric and large bowel surgery decreasing trends

were identified in recent years (3.1% and 8.5%, respectively, in 2015/16)

Enterobacteriaceae continued to increase, accounting for 28% of SSIs in 2015/16, the

highest to date. Staphylococcus aureus accounted for 13% of inpatient-detected SSIs in

2015/16 whilst the methicillin-resistant form (MRSA) accounted for 4%, both decreasing

since 2006/7. Small, steady decreases in methicillin-susceptible S. aureus (MSSA) were

observed since 2013/14, reaching 9% of SSIs in 2015/16

S. aureus predominated in orthopaedic and spinal surgery accounting for ≥35% of SSI

cases. Coagulase-negative staphylococci and Enterobacteriaceae predominated in

coronary artery bypass graft and large bowel surgery SSIs, respectively

in primary total hip prosthesis, the proportion of SSIs due to S. aureus was highest in the

uncemented (46%) compared to the cemented (37%) and hybrid fixation (31%) groups


5

Section 1. Introduction and surveillance

methods

1.1 Introduction

This report is a summary of data on surgical site infections (SSIs) collected by NHS

hospitals and independent sector (IS) NHS treatment centres in England participating in

one of 17 surgical categories of surveillance between April 2004 and March 2015. The

results include orthopaedic data submitted by hospitals following the mandatory

surveillance requirement introduced by the Department of Health in April 2004 [1]. This

requires all NHS trusts undertaking orthopaedic surgical procedures to carry out a

minimum of three months’ surveillance in each financial year in at least one of four

categories (hip prosthesis, knee prosthesis, repair of neck of femur or reduction of long

bone fracture). Trusts with very small volumes are exempt from the mandatory

surveillance but are expected to undertake surveillance in a category that reflects the

largest component of their surgical activity.

In response to the Competitions and Market Authority (CMA) order in 2014, all

independent sector hospitals have been required to provide specified patient outcome

data including SSIs from September 2016 and to publish results from April 2017 [2].

1.2 Data collection and feedback

To participate in PHE’s national SSI surveillance programme, hospitals must first attend

training at the national co-ordinating centre in London. Once trained, all participating

hospitals are required to follow the surveillance protocol outlining the case definitions

and follow-up methods [3]. Surveillance data are collected prospectively on all eligible

patients in a given surgical category on a quarterly basis (three months). Patients are

followed-up to identify SSIs for 30 days after surgery for non-implant procedures and

one year for prosthetic implant procedures. A set of demographic and operation-related

data are collected for each eligible procedure and submitted via a secure web-based

application.

Since July 2008, hospitals are required to have systems in place to identify patients who

are included in the surveillance and later readmitted to hospital with an SSI. SSIs

identified on readmission are assigned to the hospital where the original operation took

place. Other post-discharge surveillance methods were introduced in 2008 but remain

optional. They are: a) systematic review of patients attending outpatient clinics or at

home by clinical staff trained to apply the case definitions and b) wound healing

questionnaires completed by patients 30 days after the operation [3]. Data derived from


6

these optional methods are not currently included in the national benchmarks or used

for outlier assessment.

After each completed quarter, participating hospitals can download automated

confidential reports accessed securely from the SSI web application for dissemination

within their trust. These reports provide the hospitals’ crude SSI incidence and the

corresponding national benchmark from the same surgical category. A range of reports

are available from the web application providing stratified rates for key risk factors and

data quality indicators.

PHE analyses the submitted data at quarterly intervals to identify hospitals whose SSI

incidence falls above the 90th or below the 10th percentiles nationally for a given

surgical category. PHE alerts these hospitals of their outlier status and encourages

them to investigate possible reasons. Additional support is provided to hospitals who

request advice or in-depth epidemiological analysis to assist with investigating

persistent SSI problems. On-site visits are offered by PHE as a vehicle for sharing in-

depth analyses and further surveillance advice to multi-disciplinary teams.

1.3 Definitions

SSIs are defined according to a standard set of clinical criteria for infections that affect the

superficial tissues (skin and subcutaneous layer) of the incision and those that affect the

deeper tissues (deep incisional or organ-space). These are based on the definitions

established by the US Centers for Disease Control and Prevention (CDC) [4] with a minor

modification, involving the requirement for pus cells in addition to a positive culture from wound

samples (for all SSI types) and the need for at least two symptoms to accompany a clinical

diagnosis (superficial SSIs only).

A range of risk factors are captured through the surveillance including the ASA score. This is

the patient’s pre-operative physical status on a scale from one to five, a classification system

developed by the American Society of Anesthesiologists (ASA), with higher scores indicating

severe systemic disease. This along with the category-specific T-time (duration of surgery

exceeding the >75th percentile rounded to the nearest hour) and the degree of wound

contamination constitute the three components of the CDC National Healthcare Safety Network

(NHSN) Risk Index [5]. There are four wound contamination classes (clean, clean-

contaminated, contaminated and dirty) each with accepted standard definitions [3;4;6]

1.4 Participation in international surveillance

PHE shares anonymised SSI surveillance data with the European Centre for Disease

Prevention and Control (ECDC) HAI-Net on an annual basis using ECDC’s protocol,

also based on CDC definitions [6]. As data are anonymised, they cannot be traced back

to individual patients, surgeons or named hospitals. ECDC collates SSI data from other


7

European member states and publishes comparative analyses including trends. These

provide an opportunity to examine variation in the SSI incidence between European

countries and to improve understanding of how these infections may be prevented.

Inter-country variation, however, may be affected by differences in surveillance

methodology and/or risk factors. A recent comparison of SSI data between England and

Norway identified differences in the SSI incidence for selected surgical categories due

to differences in surveillance methodology and risk factors [7].

1.5 Analyses presented in this report

Data on surgical procedures carried out between April 2004 and March 2016 collected

by NHS hospitals and independent sector NHS treatment centres were extracted on 17

October 2016 for this report. The SSIs included in this report are based on cases

detected during the inpatient stay or on readmission to the same or other hospitals

following the initial operation. Data from 2004 were used for showing secular trends in

inpatient-detected SSI incidence for illustrative purposes. SSI trends from July 2008

were evaluated as this was when readmission surveillance became a requirement for all

participating hospitals (readmission surveillance was optional prior to July 2008).

For benchmarking purposes, cumulative five-year data were used (April 2011 to March

2016).

The cumulative SSI incidence (%) presented in this report was based on SSIs detected

during hospital stay combined with SSIs identified on readmission following the initial

operation. Where appropriate, inpatient SSIs were analysed separately for meaningful

interpretation. To take into account the variation in the length of follow-up, the incidence

density (ID) was calculated using the number of cases in the numerator and the total

number of days of patient follow-up (from inpatient surveillance) in the denominator

giving the number of SSIs per 1,000 patient days of follow-up. The exact binomial 95%

confidence interval (CI) for proportions and Poisson CI for the ID are given in this report.

Orthopaedic data from participating hospitals are aggregated to trust level for public

reporting purposes. Funnel plots were constructed for each orthopaedic category using

the inpatient/readmission cumulative SSI incidence per NHS trust in 2015/16. Each plot

identifies trusts with estimates falling within the expected variation and outliers falling

beyond the upper or lower 95% control limits (significantly higher or lower than expected

respectively). Independent sector NHS treatment centres are also included in these

plots. An additional supplement to this report presents the cumulative SSI estimate by

NHS trust in 2015/16 for the mandatory orthopaedic surveillance:

https://www.gov.uk/government/publications/surgical-site-infections-ssi-surveillance-nhs-

hospitals-in-england

https://www.gov.uk/government/publications/surgical-site-infections-ssi-surveillance-nhs-hospitals-in-england



8

Section 2. Overview

2.1 Hospital participation and surgical volumes

Overall, a total of 198 NHS hospitals representing 142 NHS trusts and an additional

seven IS NHS treatment centres participated in 2015/16, contributing data on 136,872

procedures and 1,632 inpatient/readmission SSIs for 17 surgical categories. This

compares with 130,321 procedures and 1,533 inpatient/readmission SSIs in 2014/15

collected by 200 NHS hospitals representing 143 trusts and an additional eight IS NHS

treatment centres.

Figure 1 shows trends in hospital participation and surgical volume in the national SSI

surveillance scheme split by the mandatory orthopaedic and the voluntary non-

orthopaedic modules. In 2015/16, 184 NHS hospitals representing 138 NHS trusts

participated in the mandatory orthopaedic categories along with seven IS NHS

treatment centres, the combined total (n=191) being one less than in 2014/15 (n=192).

A marginal increase (1%) in surgical volume submitted was seen in 2015/16

(n=103,838) compared to 2014/15 (n=102,501). Two eligible NHS trusts did not

participate in the orthopaedic surveillance in 2015/16. Of the remaining 138 NHS trusts,

85% (n=118) fulfilled more than the minimum requirement (one surveillance category for

a minimum of three months) for the mandatory surveillance.

For the voluntary non-orthopaedic categories, total surgical volume increased by 19%

between 2014/15 (n=27,820) and 2015/16 (n=33,034). The number of participating

hospitals also increased slightly from 70 to 72 hospitals.

2.2 Continuous surveillance

After excluding five of 17 surgical categories that had a small number of participating

hospitals in 2015/16 (<10), the remaining categories showed considerable variation in

the proportion of hospitals undertaking continuous surveillance. This was highest in

knee prosthesis and spinal surgery (60% each), followed by hip prosthesis (59%),

reduction of long bone fracture (47%), small bowel and cardiac (non-CABG) surgery

(46% each), large bowel surgery and repair of neck of femur (45% each). Coronary

artery bypass graft (CABG) and breast surgery had the lowest proportions (40% and

23%, respectively).

Among the orthopaedic categories only knee prosthesis showed an increase in

continuous surveillance in 2015/16 compared to 2014/15 (56%). Hip prosthesis

remained comparable to 2014/15 (also 59%). In the non-orthopaedic categories, trends


9

in continuous surveillance varied between years although for spinal surgery continuous

surveillance in 2014/15 (60%) was sustained in 2015/16. Figure 1: Trends in the number of NHS hospitals participating in the SSI Surveillance Service and procedures submitted, England

-30

20

70

120

170

220

0

20,000

40,000

60,000

80,000

100,000

120,000

Nu

um

be

r o

f h

os

pit

als

Nu

mb

er

of

op

era

tio

ns

operations - non-orthopaedic data operations - orthopaedic data

hospitals - non-orthopaedic data hospitals - orthopaedic data

2.3 Data quality

Data completion for key SSI risk factors is essential for risk stratification purposes,

enabling hospitals to investigate factors underpinning unusual deviations from the SSI

national benchmark or from previous local trends. Table 1 shows the proportion of

records submitted with completed fields for key data items by surgical category during

2015/16.

Nationally, data completion for age, patient sex, date of admission and procedure code

remained at 100% for records submitted in 2015/16 across all surgical categories. For

the risk factors used in the NHSN risk index (see section 1), data completion was very

high (≥99%) for wound class and duration of surgery across most surgical categories

although for ASA score there was more variation, ranging between 63% for cardiac

(non-CABG) surgery and 97% for gastric, hip prosthesis and knee prosthesis.


10

Data completeness was considerably lower for height and weight used to calculate body

mass index (BMI), ranging from 21% in reduction of long bone fracture to 80% for

CABG surgery. In 2015/16, there were nine categories where BMI data completion was

50% or more of submitted records (bile duct/liver/pancreatic, breast, cholecystectomy,

cardiac (non-CABG), CABG, cranial, hip prosthesis, knee prosthesis and large bowel

surgery). In 2015/16, BMI data completion increased in 12 surgical categories

compared to 2014/15, with the biggest increase in cranial surgery (increasing from 35%

in 2014/15 to 61% in 2015/16) followed by gastric surgery (from 9% in 2013/14 to 33%

in 2015/16). BMI data completion for knee prosthesis was 60% in 2015/16 and in

2014/15. For the remaining four surgical categories, BMI data completion decreased

compared to 2014/15. Six surgical categories had BMI data completion of ≥50% in

2015/16 and in 2014/15: cardiac non-CABG, cholecystectomy, CABG, hip prosthesis,

knee prosthesis and large bowel surgery. Table 1: Percentage of submitted surveillance records with complete data for key data items, NHS hospitals in England, 2015/16

* patients aged ≥16 years

Optional post-discharge surveillance methods are available to participating hospitals

and these are: patient post-discharge questionnaires (PDQs) filled in at 30 days after

the operation and outpatient review clinics. These two surveillance methods are

particularly important for surgical categories with short length of hospital stay. Of these

two methods, PDQ surveillance is the most commonly used by hospitals and offers a

comprehensive system of monitoring to identify infections that occur after discharge and

N = 136,872

Total No.

operations Age

ASA

score

Duration of

operation

Wound

class BMI*

Patient

sex

Date of

admission

OPCS

code

Abdominal hysterectomy 589 100 91 100 86 41 100 100 99

Bile duct, liver or pancreatic surgery 479 100 96 100 98 53 100 100 100

Breast surgery 5,337 100 95 100 100 51 100 100 100

Cardiac (non-CABG) 3,143 100 63 100 100 79 100 100 100

Cholecystectomy 943 100 97 100 100 52 100 100 100

Coronary artery bypas graft 5,116 100 75 100 97 80 100 100 100

Cranial surgery 1,531 100 87 99 100 61 100 100 100

Gastric 653 100 97 100 100 33 100 100 100

Hip replacement 40,125 100 97 100 100 55 100 100 100

Knee replacement 42,243 100 97 100 100 60 100 100 100

Large bowel surgery 4,607 100 96 100 100 58 100 100 100

Limb amputation 492 100 81 100 83 22 100 100 100

Reduction long bone fracture 2,273 100 94 100 100 21 100 100 100

Repair of neck of femur 19,197 100 95 100 100 22 100 100 100

Small bowel surgery 1,004 100 95 100 100 45 100 100 100

Spinal surgery 7,938 100 94 95 99 30 100 100 100

Vascular surgery 1,202 100 95 100 100 44 100 100 100


11

are treated in the community. Overall, the proportion of procedures where PDQs were

given to patients increased from 39% in 2010/11 to 50% in 2015/16 (Figure 2 showing

proportions of PDQ given and PDQ completed based on number of operations). The

patient response rate based on PDQs completed increased from 69% of questionnaires

given in 2010/11 to 77% in 2015/16. PDQ use was highest at >70% for abdominal

hysterectomy, bile duct liver and pancreatic, cholecystectomy, cranial and breast

surgery and lowest at <40% for cardiac surgery, coronary artery bypass graft (CABG),

repair of neck of femur (RNoF) and vascular surgery. Although lower, collection of data

from outpatient review surveillance has increased, from 21% of procedures in 2010/11

to 26% in 2015/16 (not shown). Figure 2: Trends in the proportion of procedures with post-discharge questionnaire given and completed, NHS hospitals in England, April 2010 to March 2016

2.4 Patient and operation-related characteristics

Patient and operation-related characteristics derived from surveillance records for

patients undergoing surgery in 2015/16 are shown in Tables 2 and 3.

The median age varied by surgical category (Table 2) being lowest in abdominal

hysterectomy (50 years) and highest in repair of neck of femur (85 years). Duration of

operation was highest in both cardiac categories (CABG and cardiac non-CABG), each

0

10

20

30

40

50

60

70

% o

f o

pe

rati

on

s

PDQ given (%)

PDQ completed (%)


12

having median duration in excess of 200 minutes. The categories of surgery with the

highest proportion of patients with an ASA score ≥3 were in CABG and cardiac

categories (>90% each) and in limb amputation surgery (78%). Small bowel, limb

amputation and large bowel surgery had the highest proportion of procedures with a

wound classified as contaminated or dirty (47%, 25% and 16%, respectively).

Of nine surgical categories with BMI data completion of 50% or more (see Table 1),

knee prosthesis and cholecystectomy had the highest proportion of patients (55% and

45% respectively) classed as obese (BMI ≥ 30kg/m2). Five of these nine categories also

had BMI data completion of 50% or more in 2014/15 and within these, the proportion of

patients that were obese in 2015/16 remained similar to the previous year.

The proportion of operations performed on an emergency basis (defined as procedures

that are immediate, unplanned and life-saving or those that are performed immediately

after resuscitation) was most common for vascular and large bowel surgery (6% each).

Although repair of neck of femur patients are usually admitted on an emergency basis,

the proportion classified as emergency procedures based on the PHE surveillance

definitions was 1.2%.

The proportion of patients receiving surgical antimicrobial prophylaxis was high (≥95%)

in 11 surgical categories including the orthopaedic categories, reflecting the current

recommendations aimed at defined surgical groups [8;9]. The lowest proportion

observed was in cholecystectomy (61%).

The proportion of patients where discontinuation of inpatient surveillance was due to

death was highest in repair of neck of femur and limb amputation surgery (6% and 4%,

respectively). These should be interpreted with caution as these data only capture

deaths within the inpatient stay.

Table 3 shows primary indication for surgery in patients undergoing hip prosthesis, knee

prosthesis and repair of neck of femur. In hip and knee prosthesis, osteoarthritis

accounted for the majority of procedures (82% and 92%, respectively) and the second

most common indication was revision (10% and 6%, respectively). For repair of neck of

femur, the most common primary indication was trauma (99%).

Surveillance of Surgical Site Infections in NHS hospitals in England 2015/16

13

Table 2: Patient and operation-related characteristics by surgical category, NHS hospitals, England, 2015/16

*patients aged ≥16 years

Table 3: Primary indication for surgery, orthopaedic procedures, NHS hospitals, England, 2015/16

Abdominal hysterectomy 50 (44-60) 101 (77-134) 3 (2-4) 13.3 0.0 42.4 N/A 0.5 N/A 96.9 0.0 0.0

Bile duct, liver or pancreatic 62 (51-72 170 (82-306) 6 (1-10) 29.2 2.8 37.1 46.6 0.4 1.3 96.8 1.9 1.3

Breast surgery 58 (48-68) 60 (40-90) 1 (1-1) 9.1 0.1 31.1 1.7 0.6 1.6 72.5 12.5 0.1

Cardiac surgery 68 (52-76) 235 (185-290) 8 (6-12) 90.1 0.3 29.3 62.4 1.9 3.4 98.4 89.7 3.0

Cholecystectomy 52 (37-65) 60 (40-80) 1 (1-1) 10.1 0.7 45.2 28.3 0.1 N/A 60.7 0.1 0.3

Coronary artery bypass graft 69 (61-75) 230 (190-270) 6 (5-9) 94.3 0.0 34.4 80.2 1.1 0.2 99.1 84.4 1.8

Cranial surgery 59 (46-71 104 (60-175) 5 (3-10) 43.4 1.6 25.5 53.6 2.8 0.0 98.8 48.5 2.4

Gastric surgery 58 (44-69) 145 (80-265) 4 (2-10) 28.2 5.2 49.8 51.5 1.2 2.6 87.1 3.7 1.1

Hip replacement 71 (62-78) 84 (65-105) 4 (3-6) 22.8 0.1 38.5 39.6 0.1 9.7 99.1 100 0.2

Knee replacement 70 (63-77) 78 (62-98) 4 (3-5) 21.2 0.0 55.3 43.0 0.0 5.7 99.4 100 0.1

Large bowel surgery 69 (57-78) 173 (125-235) 8 (5-13) 39.9 16.4 25.0 51.1 6.0 N/A 96.2 2.5 2.8

Limb amputation 68 (57-76) 49 (28-72) 8 (1-19) 78.2 24.5 28.2 71.1 3.9 N/A 84.5 1.0 4.1

Reduction of long bone fracture 64 (42-80) 88 (60-120) 5 (1-13) 39.5 2.4 18.7 38.5 1.5 N/A 99.4 98.8 2.4

Repair of neck of femur 85 (78-90) 65 (51-87) 12 (7-20) 71.7 0.0 10.5 30.3 1.2 0.3 96.5 100 5.9

Small bowel surgery 63 (45-73) 120 (73-184) 8 (5-15) 42.2 46.5 18.1 50.8 4.7 0.7 93.0 2.7 2.9

Spinal surgery 54 (39-69) 126 (85-188) 3 (1-7) 21.9 0.3 33.7 47.3 0.2 35.7 98.7 49.8 0.4

Vascular surgery 73 (65-79) 175 (122-231) 4 (2-8) 72.0 0.4 24.5 73.3 6.1 0.1 92.7 59.7 2.5

Discontinuation of

inpatient surveillance

due to death (%)

Antibiotic

prophylaxis

(%)

Implant

(%)

Male

(%)

Emergency

(%)

Revision

(%)

BMI ≥ 30

(%)*

Median length

hospital stay

in days (IQR)

Median age

in years

(IQR)

Median duration

operation in

minutes (IQR)

ASA ≥ 3

(%)

Contaminated/

dirty incision

(%)

Hip replacement 39,417 32,324 (82.0) 251 (0.6) 292 (0.7) 1,883 (4.8) 755 (1.9) 423 (1.1) 368 (0.9) 2,876 (7.3) 245 (0.6) 3,912 (9.9)

Knee replacement 41,665 38,202 (91.7) 318 (0.8) 27 (0.1) 81 (0.2) 629 (1.5) 86 (0.2) 384 (0.9) 1,746 (4.2) 192 (0.5) 2,408 (5.8)

Repair of neck of femur 19,125 60 (0.3) 2 (<0.1) 0 (0.0) 18,963 (99.2) 47 (0.2) 16 (0.1) 8 (<0.1) 21 (0.1) 8 (<0.1) 53 (0.3)

No.

operations

Osteoarthritis

No. (%)

Inflammatory joint

disease No. (%)

Avascular

necrosis No. (%)

Trauma

No. (%)

Other

No. (%)Unknown

No. (%)

All

No. (%)

Fracture

No. (%)

Infection

No. (%)

Other

No. (%)

Revision


14

Section 3. Cumulative incidence (%) of SSI

3.1 Incidence (%) of inpatient and readmission SSI by surgical category

Table 4 shows the volume of surgical procedures, the number of SSI cases and the

cumulative SSI incidence (%) by surgical category for surgery undertaken between April

2011 and March 2016. Five-year aggregated data were used to give a robust measure

of the national average for benchmarking purposes. Over this period, data for a total of

611,700 procedures across 17 surgical categories were submitted by 218 participating

hospitals representing 147 NHS trusts. An additional nine NHS treatment centres

participated in the surveillance.

The cumulative incidence (%) of SSI varied by surgical category (Figure 3; Table 4)

depending largely on the inherent likelihood of microbial contamination at the operative

site associated with that type of surgery. The highest incidence (risk) was observed in

large bowel surgery at 9.8% over this five-year period. The lowest incidence was

observed in hip and knee prosthesis surgery (0.6% each).

SSIs detected through readmission accounted for a relatively high proportion of

inpatient and readmission cases in surgical categories with a relatively short length of

hospital stay such as breast surgery (89%), knee prosthesis (72%), abdominal

hysterectomy (68%), and hip prosthesis (64%). In categories with longer length of

hospital stay (gastro-intestinal, repair of neck of femur, cardiac and limb amputation

surgeries), readmission SSIs accounted for ≤35% of SSIs detected.

To take into account the variation in the length of patient follow-up during hospital stay,

the incidence density was estimated. This metric takes into account the length of

inpatient follow-up, defining SSI incidence per 1,000 days of inpatient follow-up. Using

this metric, the variation between surgical categories was less pronounced when

compared to the inpatient-based cumulative incidence. However, large bowel surgery

remains as having the highest SSI incidence density (7.8) and hip and knee surgery as

the lowest (<0.5) (Table 4).

Table 5 and Table 6 present the five-year benchmarks (April 2011 to March 2016) for

three orthopaedic categories stratified by primary and revision procedures. This analysis

excluded procedures with three-character OPCS codes (as it was not possible to

distinguish between primary and revision procedures based on standard definitions

[10]). Records with missing OPCS codes were also excluded. The total excluded was

1.4% of hip procedures (2,678/193,589); 1.7% of knee procedures (3,338/200,435) and

4.5% of repair of neck of femur procedures (4,204/92,532). The SSI incidence for

revision procedures was higher compared to primary surgery for all three procedures.

Repair of neck of femur had the highest cumulative incidence of SSI for both the


15

primary and revision subsets (1.2% and 4.7%, respectively). For repair of neck of femur,

time to onset of infection appeared more rapid for revision compared to primary

procedures (9 days and 15 days, respectively). However, the reverse was observed for

knee prosthesis, where time to onset of infection was longer for revision compared to

primary procedures (21 days and 18 days respectively). For hip prosthesis, time to

onset of SSI differed by a marginal amount. Figure 3: Cumulative SSI incidence (%) with 95% CIs (inpatient and readmission) by surgical category, NHS hospitals in England, April 2011 to March 2016


16

Table 4: Cumulative SSI incidence (%) by surgical category, NHS hospitals in England, April 2011 to March 2016

Surgical Site Infections

No. operations

No. hospitals

No. Inpatient

No. inpatient & readmission

Inpatient & readmisson % 95% CIs

Median time to infection

(days)*

Incidence density/1,000

inpatient

days† 95% CI

Abdominal hysterectomy 3,249 20 13 41 1.3% 0.9% - 1.7% 10 1.1 0.6 - 1.9

Bile duct, liver or pancreatic surgery 2,546 7 125 142 5.6% 4.7% - 6.5% 8 4.8 4.0 - 5.7

Breast 15,656 34 14 130 0.8% 0.7% - 1.0% 14 0.6 0.3 - 1.0

Cholecystectomy 1,761 7 37 53 3.0% 2.3% - 3.9% 7 4.6 3.3 - 6.4

Coronary artery bypass graft 28,335 20 731 1,081 3.8% 3.6% - 4.0% 12 2.7 2.6 - 3.0

Cardiac (non-CABG) 13,851 17 113 172 1.2% 1.1% - 1.4% 14 0.7 0.6 - 0.8

Cranial 7,159 8 45 108 1.5% 1.2%- 1.8% 19 0.7 0.5 - 1.0

Gastric 1,517 8 32 37 2.4% 1.7% - 3.3% 8 2.5 1.7 - 3.6

Hip prosthesis 193,589 192 450 1,248 0.6% 0.6% - 0.7% 17 0.4 0.4 - 0.4

Knee prosthesis 200,435 183 325 1,169 0.6% 0.6% - 0.6% 18 0.3 0.3 - 0.4

Large bowel 19,228 52 1,599 1,890 9.8% 9.4% - 10.3% 8 7.8 7.4 - 8.2

Limb amputation 2,036 15 42 60 2.9% 2.3% - 3.8% 13 1.6 1.1 - 2.1

Reduction of long bone fracture 13,420 34 77 148 1.1% 0.9% - 1.3% 17 0.7 0.5 - 0.8

Repair of neck of femur 92,532 120 829 1,129 1.2% 1.2% - 1.3% 15 0.5 0.5 - 0.6

Small bowel 4,163 20 266 300 7.2% 6.4% - 8.0% 8 5.3 4.7 - 6.0

Spinal 35,648 26 216 482 1.4% 1.2% - 1.5% 14 1.0 0.9 - 1.2

Vascular 6,137 21 109 172 2.8% 2.4% - 3.2% 12 2.1 1.7 - 2.6

Total 641,262 227 5,023 8,362

*six records excluded due to missing operation data or hospital discharge date or negative values for follow-up time


17

Table 5: Cumulative SSI incidence (%) in primary procedures, orthopaedic categories, NHS hospitals in England, April 2011 to March 2016

Table 6: Cumulative SSI incidence (%) in revision procedures, orthopaedic categories, NHS hospitals in England, April 2011 to March 2016

No.

operations

No.

hospitals

No.

Inpatient

No. inpatient &

readmission

Inpatient &

readmisson % 95% CIs

Median time

to infection

(days)

Incidence

density/1,000

inpatient days 95% CIs

Hip prosthesis 169,249 192 284 927 0.5% 0.5% - 0.6% 17 0.3 0.3 - 0.4

Knee prosthesis 184,948 181 242 964 0.5% 0.5% - 0.6% 18 0.3 0.2 - 0.3

Repair of neck of femur 88,031 120 769 1,055 1.2% 1.1% - 1.3% 15 0.5 0.5 - 0.6


No.

operations

No.

hospitals

No.

Inpatient

No. inpatient &

readmission

Inpatient &

readmisson % 95% CIs

Median time

to infection

(days)

Incidence

density/1,000

inpatient days 95% CIs

Hip prosthesis 21,662 181 155 299 1.4% 1.2% - 1.5% 16 0.8 0.6 - 0.9

Knee prosthesis 12,149 173 68 179 1.5% 1.3% - 1.7% 21 0.7 0.5 - 0.9

Repair of neck of femur 297 72 14 14 4.7% 2.6% - 7.8% 9 2.2 1.2 - 3.7



18

3.2 Risk factors for SSI

Figure 4 shows the cumulative incidence of SSI for selected risk factors by surgical

category based on data for 2015/16. The selected risk factors presented are age, ASA

score, ‘T-time’ (duration of operation in minutes dichotomised into two groups – see

Section 1.3), and BMI. The BMI analysis was restricted to surgical categories where

BMI data completion was ≥50% of submitted records.

Differences in the SSI incidence between the youngest (<45 years) and oldest patients

(≥65 years) were observed in some surgical categories in 2015/16. The SSI incidence

was highest among patients aged ≥65 years undergoing bile duct/liver/pancreatic,

cholecystectomy, CABG, limb amputation and spinal surgery. Conversely, the SSI

incidence was highest in the youngest patients undergoing cranial, knee prosthesis or

large bowel surgery.

Patients with an ASA score of ≥3 typically had an increased SSI incidence except in

cardiac (non-CABG), cholecystectomy and gastric surgery where the reverse pattern

was observed. Duration of operation greater than the NHSN 75th percentile for that

category had an increased risk of SSI compared to procedures below this threshold

across all surgical categories with the exception of cholecystectomy and cranial surgery

where the opposite was observed. Excess incidence for longer duration of procedures

was most marked in breast and large bowel surgery.

The risk of SSI increased among patients who were obese (BMI ≥30kg/m2) in the

majority of surgical categories examined with the exception of bile duct/liver/pancreatic

and cholecystectomy surgery, where the reverse pattern was observed. For knee

prosthesis, the SSI risk was equivalent in both groups.

Table 7 shows the SSI incidence by primary indication for hip prosthesis, knee

prosthesis and repair of neck of femur in 2015/16. For hip prosthesis, the risk of SSI

was highest for surgery undertaken for revision due to a previous infection (3.8%). In

knee prosthesis, the SSI incidence was highest for avascular necrosis (3.7%) although

this was based on a very small sample size (n=27). The next highest incidence within

the knee category was for revision due to a previous infection (2.3%). For repair of neck

of femur, small sample sizes for the majority of indications affected the precision of

estimates although aggregated data for revision (all reasons) carried the highest SSI

risk at 7.5%.


19

Figure 4: Stratification of the inpatient/readmission SSI incidence, selected risk factors, by surgical category, NHS

hospitals, England 2015/16

0%

2%

4%

6%

8%

10%

12%

14%

SS

I in

cid

en

ce (

%)

<45y 45-64y ≥65y

0%

2%

4%

6%

8%

10%

12%

14%ASA <3 ASA ≥3

0%

2%

4%

6%

8%

10%

12%

14%<T time ≥T-time

0%

2%

4%

6%

8%

10%

12%

14%

SS

I ic

iden

ce (

%)

BMI <30 BMI ≥30

The BMI analysis is based on surgical categories where the BMI data completion was ≥50%


20

Table 7: SSI incidence by primary indication for surgery in orthopaedic categories, NHS hospitals, England, 2015/16

Hip prosthesis Knee prosthesis Repair of neck of femur

Given primary Indication No.

operations No. SSI % SSI No.

operations No. SSI % SSI No.

operations No. SSI % SSI

Primary procedures

Osteoarthritis 32,324 152 0.5% 38,202 167 0.4% 60 0 0.0%

Inflammatory joint disease 251 0 0.0% 318 5 1.6% 2 0 0.0%

Avascular necrosis 292 2 0.7% 27 1 3.7% 0 0 -

Fracture 1,883 19 1.0% 81 1 1.2% 18,963 192 1.0%

Other 755 10 1.3% 629 6 1.0% 47 1 2.1%

All primary procedures 35,505 183 0.5% 39,257 180 0.5% 19,072 193 1.0%

Revision procedures

Infection 368 14 3.8% 384 9 2.3% 8 1 12.5%

Fracture 423 3 0.7% 86 1 1.2% 16 2 12.5%

Other 2,876 26 0.9% 1,746 21 1.2% 21 1 4.8%

Unknown 245 2 0.8% 192 3 1.6% 8 0 0.0%

All revision procedures 3,912 45 1.2% 2,408 34 1.4% 53 4 7.5%


21

Section 4. Trends in the incidence of SSI

4.1 Orthopaedic categories

Figure 5 shows trends in the crude cumulative incidence of SSI in the mandatory

orthopaedic categories between 2004/5 and 2015/16. Three types of trends are shown:

inpatient-detected, readmission-detected and combined inpatient and readmission SSIs.

To estimate the average annual change in SSI, a univariable logistic regression model

was used. Trends based on inpatient and readmission SSIs during the eight-year period

from July 2008 to March 2016 were evaluated. Since some categories exhibited

variation in the annual crude incidence, the independent-year effects model (with

2008/9 as the baseline year) was compared against a linear trends model using a

likelihood ratio test (LRT). The test determined whether model fit improved or not

compared to the linear model. The odds ratio (OR) with 95% CIs are computed from

each model and reported as appropriate.

For hip prosthesis, the observed crude incidence peaked in 2008/9 at 0.9% reducing to

0.6% in 2015/16. Although over this eight-year period, no clear trends were evident

initially (from 2008/9 to 2012/13), small yet steady decreases were discernible in the

most recent three years (from 2013/14 to 2015/16). Comparison of the two models

indicated that there was evidence of non-linearity in the SSI trends (LRT p=0.0078) with

a significant decrease compared to the baseline in the last three successive years

(2013/14, 2014/15 and 2014/15), of which 2015/16 exhibited the largest decrease (OR:

0.64, 95% CI: 0.53 – 0.78, p <0.001).

For knee prosthesis the observed incidence increased after 2008/9 peaking in 2011/12

(0.7%) but returned to baseline levels (0.5%) by 2013/14 and remained at this level in

2015/16. As in hip prosthesis, there was evidence of non-linearity in the SSI trends

(LRT p=0.0004). No evidence of change in the odds of SSI in the last three years was

found compared to the baseline in 2008/9.

For repair of neck of femur, a downward trend in the crude SSI incidence was observed

from 1.7% in 2008/9 to 1.0% in 2015/16. The incidence in 2015/16 was similar to that

observed in 2014/15. There was no evidence against linearity (LRT p=0.3969) in the

SSI trends with an average decrease of 8% per year identified in the linear model (OR:

0.92; 95% CI: 0.90 – 0.94; p<0.001).

For reduction of long bone fracture, the incidence peaked in 2009/10 at 1.5% with a

subsequent decline until 2014/15 at which point the crude risk increased to 1.4%. The

incidence subsequently decreased to 1.2% in 2015/16. Although there was no evidence


22

against linearity (p=0.3806), the overall average annual decrease was not significant in

this category (0.97; 95% CI: 0.92 – 1.03; p=0.363).

Figure 5: Trends in the annual cumulative incidence of SSI (%) in orthopaedic surveillance categories, with lower and upper 95% CIs, NHS hospitals in England

4.2 Non-orthopaedic categories

Figures 6-8 show the trends in the crude inpatient, readmission and the combined

inpatient and readmission SSI incidence for the voluntary non-orthopaedic categories.

Surgical categories introduced after July 2008 are presented last.

For most of the 13 voluntary surgical categories, the observed crude SSI incidence

varied substantially between years. The linear and year-effects models were compared

to evaluate trends based on inpatient readmission SSIs from July 2008 to March 2016.

Of 13 categories examined, only two categories showed linear trends: bile duct, liver or

pancreatic and spinal surgery.

Inpatient/Readmission Inpatient Readmission


23

Bile duct, liver or pancreatic surgery showed evidence of a decreasing trend in the

crude incidence, from 9.6% in 2008/9 to 5.2% in 2015/16, despite small fluctuations in

the intervening period (Figure 6). Modelled data showed there was no evidence against

linearity (although borderline) (LRT p=0.0588) indicating that model fit was improved by

the linear model. The average annual decrease in the odds of SSI was 6% (OR: 0.94;

95% CI: 0.88 – 1.00; p=0.048).

Spinal surgery showed evidence of an increasing trend in the crude incidence albeit

small from 1.4% in 2008/9 to 1.8% in 2015/16 with small fluctuations in the intervening

period (Figure 8). There was no evidence against linearity (LRT p=0.1844) indicating

that model fit was improved by the linear model. The average annual increase in the

odds of SSI was 11% (OR: 1.11; 95% CI: 1.06 – 1.15; p<0.001).

In the remaining 11 surgical categories, there was evidence for non-linearity in the

trends. However, seven of these surgical categories did not show evidence of change in

the trends over time in either model: abdominal hysterectomy, breast surgery, cardiac

(non-CABG), cranial surgery, limb amputation, small bowel surgery and vascular

surgery (Figures 6-8).

The remaining four surgical categories showed evidence of a decrease in some years

compared to the baseline 2008/9. Of these four categories, only gastric and large bowel

surgery (Figure 6) showed a significant decrease in 2015/16 compared to 2008/9. In

gastric surgery a significant decrease in the last four successive years was observed; in

particular, the crude SSI incidence was 3.1% in 2015/16 compared to 8.6% in 2008/9

(OR: 0.66; 95% CI: 0.14 – 0.78; p=0.012). In large bowel surgery, a decrease in the last

three successive years was observed of which 2015/16 exhibited the largest decrease;

the crude SSI incidence was 8.5% in 2015/16 and 11.8% in 2008/9 (OR: 0.69; 95% CI:

0.58 – 0.8; p<0.001).


24

Figure 6: Trends in the annual cumulative incidence of SSI (%) in gastro-intestinal surveillance categories with lower and upper 95% CIs, NHS hospitals in England



25

Figure 7: Trends in the annual cumulative incidence of SSI (%) in cardiac and vascular surveillance categories with lower and upper 95% CIs, NHS hospitals in England

Figure 8: Trends in the annual cumulative incidence of SSI (%) in other voluntary surveillance categories with lower and upper 95% CIs, NHS hospitals in England



26

Section 5. Characteristics of surgical site

infections

5.1 Type of SSI

Figures 9 and 10 show the distribution of SSI type for inpatient cases in 2015/16

compared to inpatient and re-admission cases combined by surgical category.

Information on SSI type was available for all 1,632 SSIs detected through inpatient or

readmission surveillance in 2015/16 (all surgical categories). SSI type data from seven

surgical categories with ≥45 inpatient and readmission SSIs were analysed (n=1,390).

In 2015/16, the proportion of inpatient-detected SSIs that were superficial incisional

infections varied by surgical category, ranging from 32.8% in spinal surgery to 68.8% in

knee prosthesis. The observed proportions were affected by differences in the length of

post-operative hospital stay between surgical categories.

As readmission-detected SSIs will involve more serious wound complications, inclusion

of these cases tended to increase the proportion of deep SSIs particularly in surgical

categories where the median length of hospital stay was relatively short. For example,

in hip and knee prosthesis surgery, the proportion of inpatient and readmission-detected

cases that were deep increased to 45.7% and 41.1%, respectively, compared with the

proportions based on inpatient-detected cases alone (34.8% and 22.9%, respectively).

A similar pattern was observed for the organ-space SSIs although inclusion of

readmission cases did not alter the proportion that were organ-space for hip prosthesis

(23% for inpatient-detected alone or those including readmission-detected cases).

In most other voluntary categories included in this analysis, the median length of

hospital stay was longer than that for hip or knee prosthesis. However, typically, the

proportion of inpatient and readmission SSIs that were deep or organ-space was

broadly similar to the corresponding proportion for inpatient-detected SSIs.


27

Figure 9: Distribution of SSI type in inpatient cases compared to combined inpatient and readmission cases, orthopaedic surgery, NHS hospitals in England, 2015/16

Figure 10: Distribution of SSI type in inpatient cases compared to combined cases, non-orthopaedic surgery, NHS hospitals in England, 2015/16

31.8

38.9

37.9

68.8

31.3

42.0

52.0

47.2

41.1

22.9

45.7

34.8

16.2

13.9

21.0

8.3

23.0

23.2

0 20 40 60 80 100

Inpat/Readm (n=198)

Inpatient (n=144)

Inpat/Readm (n=214)

Inpatient (n=48)

Inpat/Readm (n=230)

Inpatient (n=69)

% of total SSI type

Superficial Deep Organ-space

Hip

prosthesis

Knee

prosthesis

Repair of neck

of femur

27.5

32.8

60.5

62.5

55.2

56.9

58.3

59.8

66.9

59.0

27.6

26.6

24.3

24.5

34.5

30.5

5.6

8.2

11.8

10.9

20.5

18.6

7.2

9.8

0 20 40 60 80 100

Inpat/Readm (n=142)

Inpatient (n=61)

Inpat/Readm (n=76)

Inpatient (n=64)

Inpat/Readm (n=391)

Inpatient (n=318)

Inpat/Readm (n=139)

Inpatient (n=82)

% of total SSI type

Coronary artery bypass graft

Large bowel

Spinal

Small bowel


28

5.2 Causative micro-organisms

Figure 11 shows trends in organisms reported as causing inpatient SSIs from 2004/5 to

2015/16 based on data from all surgical categories except breast, cardiac (non-CABG),

cranial and spinal surgery as these modules were introduced more recently (April 2010).

Only inpatient SSIs were analysed to remove the effect of readmission surveillance

introduced in July 2008. Overall, of 11,547 inpatient SSIs detected over this 12-year

period, 67% (7,751/11,547) had data on microbial aetiology reported as causing SSI.

Staphylococcus aureus accounted for 13% of inpatient cases in 2015/16, with no

change from 2014/15. This follows a decreasing trend since 2006/7 where S. aureus

accounted for 39% of inpatient cases. The decreases in S. aureus has been driven by

decreasing trends in methicillin-resistant S. aureus (MRSA) first appearing in 2006/7

(25% of inpatient cases), reaching 4% by 2015/16. Although this proportion was slightly

higher than that observed in 2014/15 (3%) these proportions were based on a small

number of MRSA cases (n=20 2015/16 and n=15 in 2014/15).

Methicillin-susceptible S. aureus (MSSA) did not exhibit any discernible trends until

small decreases emerged over the last three successive years: 2013/14 (12%), 2014/15

(10%) and 2015/16 (9%). Coagulase-negative staphylococci (CoNS) exhibited marginal

increases. Enterobacteriaceae accounted for 28% of cases in 2015/16, the highest to

date since their increases emerged in 2008/9.

Table 8 shows the distribution of organisms reported to cause inpatient and readmission

SSIs for all SSI types by surgical category based on aggregate three-year data from

April 2013 to March 2016. Categories with organism data on ≥200 SSI cases were

analysed to maximise statistical precision. Repair of neck of femur and spinal surgery

had the highest proportion of SSI cases with microbiology data at 81% each (503/619

and 280/345, respectively). Large bowel surgery had the lowest proportion of SSI cases

with microbiology data at 52% (549/1,066). Polymicrobial SSIs (cases with more than

one organism reported as causing SSI) were most frequent in CABG surgery at 39%

(157/405) of cases and lowest in knee prosthesis surgery at 22% (101/450).

Among monomicrobial SSIs (one organism reported as causing SSI), those reported as

having MRSA aetiology was highest in repair of neck of femur, accounting for 13%

(44/330). This may reflect the emergency nature of hip fracture admissions, which may

prevent pre-admission screening and eradication of MRSA. The lowest MRSA

proportion was in spinal surgery at 1% (3/205). S. aureus aetiology was highest in repair

of neck of femur accounting for 44% (146/330) of monomicrobial SSIs followed by knee

prosthesis at 42% (147/349). SSIs reported as having Enterobacteriaceae aetiology

were highest in large bowel surgery at 53% (187/350) followed by CABG surgery at

24% (272/248). Organisms reported as ‘coliforms’ accounted for 23% (111/482) of all

monomicrobial Enterobacteriaceae SSIs for the studied categories in 2015/16 hence an


29

analysis of the leading species could not be undertaken. CoNS were highest in CABG

surgery at 29% (64/248). Mycobacteria were not reported as causing SSIs in any of the

17 surgical categories under survey in this period (April 2013 to March 2016).

Among polymicrobial infections, Gram-positive and Gram-negative combinations were

the most frequent type in all surgical categories examined (range: 41% - 50%).

Table 9 shows the distribution of organisms causing deep/organ-space SSIs detected

during inpatient stay or on readmission for the same selected categories and time

period (April 2013 to March 2016). The distribution of monomicrobial organisms was

broadly similar to that for all SSI types with two exceptions. In CABG surgery, the

proportion of deep/organ-space SSIs reported to be caused by MSSA was elevated at

30% (33/111) compared to 22% based on all SSI types. In large bowel surgery, the

proportion due to Enterobacteriaceae was elevated at 61% (62/101) compared to 53%

for all SSI types. In the polymicrobial subset, the distribution of organisms was broadly

similar to the corresponding data for all SSI types except that the proportion of purely

Gram-positive combinations was elevated in all categories except in large bowel

surgery which remained comparable. The burden of certain organisms identified in this

subset of serious infections warrants closer vigilance.

Figure 12 shows the distribution of organisms (counts) by SSI anatomical site collected

only for organ-space SSIs. Data for selected surgical categories are presented. This

analysis was expanded to five years from April 2011 to March 2016 to maximise

statistical precision for this very small subset of infections. Reporting of SSI site was

100% in CABG, hip prosthesis and spinal surgery and 99.6% in knee prosthesis.

Organism data were available for 78%, 87%, 84% and 85% of infections with SSI site

data in these categories, respectively. The final samples available for analysis for these

categories were n=66, n=261, n=32 and n=206, respectively, although small in some.

Joint bursa was the most frequent organ-space SSI in hip and knee prosthesis, (n=246

and n=203, respectively). Among joint bursa SSIs with monomicrobial aetiology,

staphylococci followed by Enterobacteriaceae were the most common causes although

for hip prosthesis, Enterobacteriaceae were relatively more frequent compared to those

in knee prosthesis. Osteomyelitis was an infrequent organ-site SSI in these orthopaedic

categories but predominantly caused by staphylococci (S. aureus and CoNs combined).

In CABG surgery, mediastinum was the most frequent organ-space SSI overall (n=39).

Among mediastinum SSIs with monomicrobial aetiology, staphylococci were the

predominant causes followed by Enterobacteriaceae. Among endocardial SSIs with

monomicrobial aetiology, the predominant causes were staphylococci, particularly

CoNS. Spinal abscess (without meningitis) was the most frequent organ-space SSI

(n=16) in spinal surgery. Among spinal abscess SSIs with monomicrobial aetiology,

staphylococci followed by Enterobacteriaceae were the predominant causes.


30

Figure 11: Trends in micro-organisms reported as causing inpatient SSIs, proportions with lower and upper 95% CIs, all surgical categories*, NHS hospitals, England

*excludes breast, cardiac (non-CABG), cranial, and spinal surgery

*excludes breast, cardiac (non-CABG), cranial and spinal surgery

0

5

10

15

20

25

30

35

40

45

50

% o

f in

pati

en

t S

SI c

as

es

Enterobacteriaceae

S. aureus

MRSA

MSSA

CoNS

Enterococcus

Pseudomonas spp.

spp.


31

Table 8: Distribution of micro-organisms reported as causing SSI (inpatient/readmission), all SSI types, by surgical category, NHS hospitals, England, April 2013 to March 2016

*a combination that include 'anaerobic bacilli' 'anaerobic cocci' ,‘other bacteria’ or fungi

Reported causative organism No. % No. % No. % No. % No. % No. %

Monomicrobial Methicillin-sensitive S. aureus 121 29.7% 133 38.1% 102 30.9% 54 21.8% 19 5.4% 78 38.0%

Methicillin-resistant S. aureus 20 4.9% 14 4.0% 44 13.3% 11 4.4% 10 2.9% 3 1.5%

Coagulase-negative staphylococci 100 24.6% 99 28.4% 55 16.7% 73 29.4% 6 1.7% 55 26.8%

Enterobacteriaceae 79 19.4% 39 11.2% 75 22.7% 64 25.8% 187 53.4% 38 18.5%

Pseudomonas spp. 15 3.7% 8 2.3% 12 3.6% 22 8.9% 32 9.1% 11 5.4%

Streptococcus spp. 23 5.7% 24 6.9% 8 2.4% 3 1.2% 18 5.1% 4 2.0%

Enterococcus spp. 29 7.1% 12 3.4% 19 5.8% 5 2.0% 17 4.9% 4 2.0%

Other bacteria 19 4.7% 19 5.4% 10 3.0% 15 6.0% 47 13.4% 12 5.9%

Fungi including Candida spp. 1 0.2% 1 0.3% 5 1.5% 1 0.4% 14 4.0% 0 0.0%

Total monomicrobial 407 100% 349 100% 330 100% 248 100% 350 100% 205 100%

Polymicrobial Gram-positive combinations only 70 40.5% 37 36.6% 39 22.5% 32 20.4% 6 3.0% 37 49.3%

Gram-negative combinations only 26 15.0% 12 11.9% 31 17.9% 36 22.9% 38 19.1% 9 12.0%

Gram-positive and Gram-negative combinations 71 41.0% 46 45.5% 87 50.3% 73 46.5% 108 54.3% 26 34.7%

Other combinations* 6 3.5% 6 5.9% 16 9.2% 16 10.2% 47 23.6% 3 4.0%

Total polymicrobial 173 100% 101 100% 173 100% 157 100% 199 100% 75 100%

Total cases with microbiology data (% of total SSI cases) 580 79.3% 450 68.3% 503 81.3% 405 65.1% 549 51.5% 280 81.2%

151 20.7% 209 31.7% 116 18.7% 217 34.9% 517 48.5% 65 65.0%

731 100% 659 100% 619 100% 622 100% 1,066 100% 345 146%Total cases

SpinalHip Prosthesis Knee Prosthesis CABG Large bowelRepair of neck of femur

Total cases with unknown microbial aetiology (% of total SSI cases)


32

Table 9: Distribution of micro-organisms reported as causing deep/organ-space SSI (inpatient/readmission), by surgical category, NHS hospitals, England, April 2013 to March 2016

*a combination that include 'anaerobic bacilli' 'anaerobic cocci' ,‘other bacteria’ or fungi

Reported causative organism No. % No. % No. % No. % No. % No. %

Monomicrobial Methicillin-sensitive S. aureus 93 30.1% 103 38.4% 59 27.8% 33 29.7% 3 3.0% 49 34.0%

Methicillin-resistant S. aureus 15 4.9% 10 3.7% 25 11.8% 3 2.7% 2 2.0% 3 2.1%

Coagulase-negative staphylococci 79 25.6% 74 27.6% 39 18.4% 29 26.1% 3 3.0% 38 26.4%

Enterobacteriaceae 61 19.7% 29 10.8% 52 24.5% 30 27.0% 62 61.4% 31 21.5%

Pseudomonas spp. 8 2.6% 6 2.2% 6 2.8% 8 7.2% 4 4.0% 8 5.6%

Streptococcus spp. 20 6.5% 20 7.5% 6 2.8% 1 0.9% 5 5.0% 3 2.1%

Enterococcus spp. 19 6.1% 9 3.4% 15 7.1% 2 1.8% 4 4.0% 4 2.8%

Other bacteria 13 4.2% 16 6.0% 8 3.8% 5 4.5% 15 14.9% 8 5.6%

Fungi including Candida spp. 1 0.3% 1 0.4% 2 0.9% 0 0.0% 3 3.0% 0 0.0%

Total monomicrobial 309 100% 268 100% 212 100% 111 100% 101 100% 144 100%

Polymicrobial Gram-positive combinations only 58 42.3% 32 41.6% 32 23.9% 21 21.6% 2 2.2% 29 55.8%

Gram-negative combinations only 18 13.1% 7 9.1% 25 18.7% 26 26.8% 16 17.2% 6 11.5%

Gram-positive and Gram-negative combinations 56 40.9% 34 44.2% 68 50.7% 36 37.1% 48 51.6% 14 26.9%

Other combinations* 5 3.6% 4 5.2% 9 6.7% 14 14.4% 27 29.0% 3 5.8%

Total polymicrobial 137 100% 77 100% 134 100% 97 100% 93 100% 52 100%

Total cases with microbiology data (% of total SSI cases) 446 87.3% 345 85.4% 346 89.4% 208 80.6% 194 43.5% 196 87.5%

65 12.7% 59 14.6% 41 10.6% 50 19.4% 252 56.5% 28 12.5%

511 100% 404 100% 387 100% 258 100% 446 100% 224 100%

Large bowel Spinal

Total cases with unknown microbial aetiology (% of total SSI cases)

Total cases

Hip Prosthesis Knee Prosthesis Repair of neck of femur CABG


33

Figure 12: Distribution of organisms by site of organ-space SSIs (inpatient/readmission), NHS hospitals, England,

April 2011 to March 2016

0

5

10

15

20

MS

SA

MR

SA

Co

NS

Ente

robacte

riaceae

Pse

udo

mon

as s

pp.

Str

epto

co

ccus s

pp

.

Ente

roco

ccus s

pp.

Oth

er

bacte

ria

Fun

gi in

clu

din

g C

and

ida s

pp.

Gra

m+

ive

co

mbin

atio

ns

Gra

m-i

ve

co

mb

ina

tio

ns

Gra

m+

ive

and

Gra

m-ive

Oth

er

co

mbin

atio

ns*

Coronary artery bypass graftArterial/venous

Bone (osteomyelitis)

Endocardium

Mediastinum

Pericardiuam/Myocardium

0

10

20

30

40

50

60

MS

SA

MR

SA

Co

NS

Ente

robacte

riaceae

Pse

udo

mon

as s

pp.

Str

epto

co

ccus s

pp

.

Ente

roco

ccus s

pp.

Oth

er

bacte

ria

Fun

gi in

clu

din

g C

and

ida s

pp.

Gra

m+

ive

co

mbin

atio

ns

Gra

m-i

ve

co

mb

ina

tio

ns

Gra

m+

ive

and

Gra

m-ive

Oth

er

co

mbin

atio

ns*

No

of

org

an

-sp

ace S

SIs

(in

pati

en

t/re

ad

mis

sio

n

Hip prosthesis Bone (osteomyelitis)

Joint/bursa

0

10

20

30

40

50

60

MS

SA

MR

SA

Co

NS

Ente

robacte

riaceae

Pse

udo

mon

as s

pp.

Str

epto

co

ccus s

pp

.

Ente

roco

ccus s

pp.

Oth

er

bacte

ria

Fun

gi in

clu

din

g C

and

ida s

pp.

Gra

m+

ive

co

mbin

atio

ns

Gra

m-i

ve

co

mb

ina

tio

ns

Gra

m+

ive

and

Gra

m-ive

Oth

er

co

mbin

atio

ns*

Knee prosthesis Bone (osteomyelitis)

Joint/bursa

0

5

10

15

20

MS

SA

MR

SA

Co

NS

Ente

robacte

riaceae

Pse

udo

mon

as s

pp.

Str

epto

co

ccus s

pp

.

Ente

roco

ccus s

pp.

Oth

er

bacte

ria

Fun

gi in

clu

din

g C

and

ida s

pp.

Gra

m+

ive

co

mbin

atio

ns

Gra

m-i

ve

co

mb

ina

tio

ns

Gra

m+

ive

and

Gra

m-ive

Oth

er

co

mbin

atio

ns*

No

of

org

an

-sap

ce S

SIs

(in

pati

en

t/re

ad

mis

sio

n) Spinal surgery Bone (osteomyelitis)

Joint/bursa

Meningitis

Spinal abscess (without meningitis)

Vertebral disc space

PolymicrobialMonomicrobialPolymicrobialMonomicrobial


34

Figure 13 shows the distribution of organisms in primary total hip prosthesis infections (inpatient and readmission) by fixation type. The analysis was based on infections of monomicrobial aetiology. The data were expanded to five years from April 2011 to March 2016 to maximise statistical precision given that this is a small subset of infections. Procedures under the ‘Other’ group (OPCS codes W39, W54 and W58) were excluded as these do not distinguish between fixation types. The analysis showed that of three fixation groups, the uncemented group had the highest proportion of infections due to S. aureus at 46% (89/192) compared to 37% (92/251) for the cemented group

and 31% (17/55) for the hybrid group. The cemented group had the highest proportion due to Enterobacteriaceae (20%) compared to the other two fixation groups.

Figure 13: Distribution of SSI organisms (inpatient/readmission) in primary total hip prosthesis by fixation group, NHS hospitals, England, April 2011 to March 2016

33.5

3.2

26.7

20.3

5.2

5.6

3.2

2.4

0.0

40.1

6.3

21.4

15.1

1.6

4.7

5.7

5.2

0.0

29.1

1.8

27.3

16.4

3.6

9.1

10.9

0.0

1.8

0

5

10

15

20

25

30

35

40

45

% o

f in

pa

tie

nt/

rea

dm

iso

n S

SIs

Cemented (n=251)

Uncemented (n=192)

Hybrid (n=55)


35

Section 6. Variation in SSI incidence

between NHS hospitals

6.1 Box and whisker plots

Figure 14 is a box and whisker plot showing the distribution of individual hospital SSI

estimates (%) against five percentiles (10th, 25th, 50th, 75th and 90th) by surgical

category. Each percentile represents a value below which a proportion of the total

observations lie. The box plots show inter-hospital variation within each surgical

category as well as inter-category variation. SSI estimates that lie at extreme ends

(above the 90th percentile or below the 10th percentile) may be indicative of a problem

(either excess infections or failure to detect SSIs, respectively).

Figure 14: Distribution of the cumulative SSI incidence (%) at hospital level* by surgical category, NHS hospitals in England, April 2011 to March 2016

* hospitals with <95 operations in hip, knee or abdominal hysterectomy surgery are excluded; in the remaining categories, hospitals with <45

operations are excluded. For categories with <10 hospital participants, the distribution without a box plot is presented only.

hospital


36

6.2 Funnel plots

In 2015/16, data on 97,174 orthopaedic procedures and 657 inpatient/readmission SSIs

was submitted by 184 NHS hospitals representing 138 NHS trusts. An additional seven

independent sector NHS treatment centres contributed data on 6,664 orthopaedic

procedures and 13 SSIs. Two NHS trusts performing orthopaedic surgery did not

contribute orthopaedic data in 2015/6.

Figure 15 shows funnel plots displaying variation in the trust SSI incidence in 2015/16

by orthopaedic category. The cumulative incidence of SSI per 100 operations is plotted

against the number of procedures for each participating NHS site. The upper and lower

95% control limits (dashed black lines) define the ‘limits’ of expected variation. Trusts

lying outside these limits are considered to be outliers. Outlier status should be used as

a trigger for further investigation as the results are indicative not confirmatory. The

margin of error associated with 95% control limits is 5%. The 99% control limits (solid

red lines) are also presented for information. The funnel plots do not adjust for case-mix

so the results should be interpreted with caution.

The funnel plots identified 18 providers that were statistical outliers in 2015/16 with SSI

estimates falling beyond the 95% CLs. Of these, nine NHS trusts were high outliers and

nine (eight NHS trusts and one NHS treatment centre) were identified as low outliers.

Overall, four of 18 outliers were outliers in more than one orthopaedic surgical category.

Seven of these NHS trusts and one NHS treatment centre were also outliers in the

previous year.

Although a low outlier may reflect a high standard of infection control practice and

surgical skill, other factors should be considered to eliminate methodological bias. For

example failure to implement a systematic readmission alert system may miss SSI

cases. All high and low outlier trusts have been contacted to make them aware of their

outlier status and encouraged to review their practices including their surveillance

methodology.

The inpatient/readmission SSI incidence for individual NHS trusts (and participating

NHS treatment centres), are published separately as part of this report:

https://www.gov.uk/government/publications/surgical-site-infections-ssi-surveillance-nhs-

hospitals-in-england




37

Figure 15: Cumulative incidence of inpatient/readmission SSI plotted against the number of operations by NHS trust and mandatory orthopaedic surgical category, England, 2015/16


38

Section 7. Hospital perspectives

In September 2016, PHE emailed all participating hospitals offering them an opportunity

to submit a short piece describing their experience of SSI surveillance including how

they used surveillance to change practice. This chapter presents a selection of

examples contributed by hospitals where they describe their own experiences with SSI

surveillance and initiatives undertaken to prevent or reduce the SSI incidence at local

level.

Royal Devon & Exeter NHS Foundation Trust

PHE commentary

The trust has described using a

range of surveillance activities to

change clinical practice and

improve its SSI incidence

particularly for spinal surgery.

These include continuous

surveillance and timely feedback of

data. Ongoing review of local

trends has been found to be

particularly useful as is a multi-

disciplinary approach to developing

prevention strategies.

At the RD&E hospital we have committed to continuous voluntary surveillance in the hip, knee and spinal categories for the last 8 years. The short bursts of surveillance that we perform, such as breast surgery, have been useful drivers for effecting change. We have been able to demonstrate the impact of these short bursts of surveillance in driving infection rates down. As a trust, however, we find that the true value of surveillance lies in doing it continuously.

Continuous voluntary surveillance provides real time information on incidence, hotspots and trends of infection over periods of time which can be used to reduce, prevent and manage infection before, during and after the operation. Whilst benchmarks are a valuable resource for comparison, the real value for us is being able to compare ourselves against ourselves. This is coupled with continuous review and feedback. Recent experience of a sudden rise in the spinal SSI rate allowed us to implement various SSI care bundle improvement strategies, based on multi-disciplinary action planning. This was supplemented with additional in-depth PHE data analysis. Ongoing surveillance and timely feedback has led to us being able to submit our first quarter of surveillance data with no spinal SSI in 2 years. Melanie Thornton, infection surveillance specialist nurse


39

PHE commentary This trust targeted readmission SSIs as an area for quality improvement. The trust then developed a novel, low-cost technique to reduce the incidence of readmission SSI cases using a photo at discharge scheme. The benefits reported include a decrease in the incidence of readmission SSIs and a strengthened multi-disciplinary team, working across the acute and community sectors.

Our hospital used Public Health England (PHE) analysis to determine that the average date of onset is day 10 for superficial SSI (interquartile range: 7-15 days), and day 14 for deep incisional SSI (interquartile range: 9-26 days) for our coronary artery bypass graft (CABG) patient group. These patients are typically discharged between day 5-7, suggesting many wound concerns arise post-discharge.

provides patients and their families with important infection prevention advice, and a photo of their wound to monitor their wound healing. Patient feedback on the initiative suggests patients find the photo reassuring and useful. This approach also improves the link between acute and community care, as district nurses or GPs can use PaD to assess whether the wound is improving or deteriorating. PaD is associated with a fourfold reduction in readmissions for SSI and associated savings between £5-15K per readmission for incisional SSI prevented [11]. Staff engagement has been crucial in the success of the program; the Harefield Hospital nursing staff and discharge team ensure >95% of cardiac surgical patients receive PaD at discharge. Melissa Rochon, clinical nurse specialist in surveillance

In 2014, the surveillance team implemented a quality improvement approach to reduce the risk of readmission for SSI. The Photo at Discharge (PaD) scheme

Royal Brompton & Harefield NHS Foundation

Trust


40

Section 8. Conclusions

8.1 Summary of findings in 2015/16

Overall, a total of 198 NHS hospitals representing 142 NHS trusts and an additional

seven NHS treatment centres participated in SSI surveillance in 2015/16 (all surgical

categories). There was a 1% and 19% increase in the number of procedures submitted

for orthopaedic and non-orthopaedic categories, respectively, compared to 2014/15.

The increased focus on non-orthopaedic surgery is an encouraging development and in

line with new NICE quality standards for HCAI surveillance, emphasising the importance

of assessing and tackling infections in non-mandatory categories [12].

Data quality was high for most of the key data items in 2015/16 although further

improvements in the completion of ASA score and BMI data would enhance the

interpretative value of the data. In 2015/16, BMI data completion of 50% or more of

submitted records was observed in nine surgical categories. The proportion of patients

that were obese was high in a number of surgical categories surveyed. If obesity levels

in the population continue to rise this will increase the propensity for diseases such as

osteoarthritis or coronary heart disease, increasing the need for surgery. This will

increasingly contribute to the total burden of SSI given the impact of elevated BMI on

risk of patients developing an SSI [13;14].

Trends in the SSI incidence showed a variable picture between surgical categories. The

incidence of SSI following hip and knee prosthesis remained low in 2015/16 (<1%) and

comparable with other European estimates [15]. Despite the low incidence observed in

England, it was encouraging to see decreasing trends in hip prosthesis in the last three

successive years. It remains to be seen whether this trend will continue or remain at this

level. The decreasing trend in SSI rates following repair of neck of femur continued in

2015/16 whereas no discernible trends in SSI could be determined for patients

undergoing reduction of long bone fracture. Among the voluntary-based surgical

categories, spanning gastrointestinal, cardiac, vascular, obstetrics and gynaecology,

neurology and other general surgery procedures, inter-year variations in the SSI

incidence were observed. Of these categories, an increasing trend was identified for

spinal surgery and a decreasing trend for bile duct, liver or pancreatic surgery. In gastric

and large bowel surgery decreasing trends were identified only over the most recent

successive years. It remains to be seen whether these trends will continue. Further

investigations are needed to explain the trends seen in these categories.


41

8.2 Variation in the SSI rate between participating hospitals

Analysis of data continues to show inter-hospital variation in the SSI incidence for some

surgical categories. This suggests that more can be done to reduce this variation by

implementing a range of interventions described by some hospitals earlier in this report.

Variation in the SSI incidence between hospitals or within hospitals provides a trigger

for investigating possible causes and identifying ways of reducing infection rates. PHE

undertakes quarterly outlier detection to ensure hospitals are aware of their elevated

infection rates and works in partnership with high outlier hospitals to assist them with

further investigations. Hospitals identified as low outliers are also alerted so that they

can investigate their case ascertainment methods. The use of reports available from the

web-based application allows hospitals to risk-stratify elevated SSI rates to determine

whether differences in case-mix or operation-related factors can explain some of the

variation. Auditing clinical practice against NICE’s quality standards for SSI, comprising

a simple evidence-based checklist of pre, peri-operative and post-operative

interventions, provides an opportunity for hospitals to optimise clinical practices as a

means of reducing or preventing SSIs [16]. The quality standards, based on the NICE

SSI prevention guidelines, cover glucose control, normothermia, skin asepsis and

surgical antibiotic prophylaxis (choice, timing of agent, frequency and dose) [8]. The

guidelines on surgical antimicrobial prophylaxis are an additional resource with similar

recommendations based on a review of the evidence [9]. New WHO guidelines on the

prevention of SSI are also available providing a global standardised approach to pre,

peri and post-operative measures whilst addressing previous gaps in knowledge

[16;17]. The experience shared by the hospitals in Chapter 7 provides an opportunity for

other hospitals to learn what interventions have worked for others in preventing SSIs.

8.3 Optional post-discharge surveillance

In addition to readmission surveillance, which all hospitals are required to undertake,

optional post-discharge surveillance methods using patient post-discharge

questionnaires (PDQ) or surveillance in outpatient settings can provide hospitals with

valuable additional post-discharge data on SSIs for internal comparison of hospital

trends, providing consistent response rates are achieved by hospitals over time. Post-

discharge surveillance is particularly important for surgical categories with a short length

of hospital stay.

The proportion of procedures followed up using optional post-discharge surveillance

methods has been increasing steadily over the last few years, although there remains

some variation in completion rates between hospitals. Such variation has a significant

effect on the overall SSI incidence [18], precluding its use for national benchmarking.


42

8.4 Microbial aetiology

A markedly decreasing trend in S. aureus as a reported cause of SSI has been

occurring since 2006/7. These trends have continued into 2015/16 although the

proportion of SSIs caused by S. aureus in 2015 was 13%, similar to that observed in

2014/15. The decreasing trend in S. aureus SSIs is likely to be explained by the marked

reductions in MRSA SSIs, also occurring since 2006/7. In 2015/16, MRSA accounted

for 4% of SSIs and, although slightly higher than that observed in 2014/15 (3%), this

effect may have been due to random error resulting from decreasing sample sizes over

time for this pathogen although it remains to be seen whether this trend will persist. The

reduction in MRSA is likely to reflect the impact of infection control initiatives directed at

controlling MRSA of which pre-admission screening and decolonisation of carriers

represents an important component [19].

MSSA decreased over the last three successive years reaching 9% of cases in

2015/16. It remains to be seen whether the trend will also persist. Nevertheless, dual

reductions in the methicillin resistant and susceptible forms of S. aureus are

encouraging as these contribute to reductions in overall S. aureus. Enterobacteriaceae

as reported causes of SSIs had been increasing since 2008/9, accounting for 28% of

SSIs by 2015/16, the highest to date.

The trends in MRSA and Enterobacteriaceae SSIs have been modelled and found to be

significant, with the increase in Enterobacteriaceae raising particular concern in the

context of increased antibiotic resistance in these pathogens [20]. Initiatives to reduce

Gram-negative infections are now a major focus of national activity [21].

At surgical-category level, data for 2015/16 showed that among monomicrobial

infections, S. aureus was the predominant organism in orthopaedic surgery and spinal

surgery, accounting for between 35% and 44% of SSI cases. Coagulase–negative

staphylococci were the predominant organisms in CABG surgery and these may reflect

the use of implants. Enterobacteriaceae were the predominant organisms in large bowel

surgery, reflecting the expected organisms at the operative site.

Among polymicrobial infections, Gram-positive and Gram-negative combinations were

the most frequent type of infections in all surgical categories examined. Although this

combination affected a small set of patients, this may have implications for choice of

surgical antimicrobial prophylaxis at local level.

The organism distribution by SSI site collected only for organ-space SSIs identified

staphylococci and Enterobacteriaceae as generally the most frequent pathogens.

However, data quality needs further improvement in terms of collection of organism data

in all specialities examined to enhance interpretive value in this subset of infections.


43

The organism distribution among primary total hip prosthesis infections by fixation group

is a novel analysis and provides further useful insights. The burden of S. aureus was

found to be highest in the uncemented group compared to the cemented and hybrid

fixation groups. The cemented group had the highest proportion due to

Enterobacteriaceae compared to the other two fixation groups.

Due to recent concern over severe Mycobacterium chimaera infections associated with

heater-cooler units in cardiothoracic surgery [22-24], guidance has been published by

PHE to assist hospitals undertaking cardiac surgery in managing this risk [24]. Hospitals

are also asked to report possible M. chimaera infections or other members of the

Mycobacteria avium complex (MAC) isolated from clinically significant specimens in

patients with a history of cardiopulmonary bypass surgery or extracorporeal membrane

oxygenation to their local PHE Health Protection Teams [25]. M. chimaera is a slow

growing atypical mycobacterium and most infections associated with heater coolers

have arisen more than one year after surgery, beyond the follow-up period within the

SSISS programme [24]. However, monitoring SSI levels in cardiac patients, especially

those undergoing valve repair or replacement, will provide a means to assess risk in

these patients from a range of pathogens that could be acquired during surgery. 8.5 Future directions

The SSI web-based application is currently undergoing development to enhance the

user experience of data by improving navigation and streamlining the in-built quality

assurance process. The new web-application will be launched during 2017 with

upgraded functionality and response speed.

PDQs are the most commonly used of the optional post-discharge methods and offer a

comprehensive way of collecting data. However, before PDQ data can be included in a

benchmark an effective means to improve uptake by hospitals and completion by

patients is needed. To support continuous improvement in SSI surveillance, including

reducing the hospital resources needed to run the surveillance, PHE are in the process

of developing an electronic PDQ system as an extension to the existing SSISS system.

This should enhance the ease of data collection, both for patients and healthcare

workers, as well as improving the quality and timeliness of data collection by boosting

PDQ return rates. In turn, these more consistent data could be used to derive a national

benchmark.


44

Section 9. Glossary

ASA score

Patient’s pre-operative physical status scored by the anaesthetist according to the

American Society of Anesthesiologitsts’ classification of physical status. There are five

ASA scores from A1 denoting normally healthy patient to A5 denoting moribund patient

with little chance of survival.

Wound Class

This describes the degree of wound contamination at the time of the operation, based

on an intenational standard classification system. The classification ranges from W1

denoting a clean uninfected wound outside the respiratory, alimentary, genital or urinary

tract to W4 denoting dirty or infected wounds and include operations in which acute

inflammation with pus is encountered or in which perforated viscera are found.

Cumulative incidence

The total number of SSIs as a proportion of the total number of patients undergoing a

procedure in the same category of surgery per 100 procedures (%).

Incidence density

The total number of SSIs (identified through inpatient surveillance) divided by the total

number of days of inpatient follow-up expressed as the number of SSIs per 1,000 days

of patient follow-up.

Confidence interval and control limits

The 95% confidence interval (exact binomial for proportions and exact Poisson for the

incidence density) provides a guide to the precision of the estimate based on the

number of procedures (or days of follow-up) surveyed from participating hospitals. The

calculation does not take into account patient clustering within hospitals. Given the level

of confidence, the interval indicates that the ‘true’ incidence’ could lie anywhere

between the lower and higher confidence limits. Control limits (CL) used in funnel plots

are equivalent to exact binomial confidence limits at 95% (warning) and 99% (action).

These define the limits of expected variation. The probability of a cumulative incidence

lying above the 95% CL by chance alone is less than 5% (there is a 1 in 20 chance that

the observed estimate is due to chance alone in a defined period); if above the 99%

CLs, this is less than 1%. The 99% control limits also define the limits of expected

variation and are computed based on the same ‘sample size; the confidence range is

wider but this is offset with a lower margin of error (1%). The 99% control limits are

presented for information since the conventional level of significance is 5%.


45

P-value

This measures the probability of an estimate (e.g. a risk) being as or more extreme than

that observed in a study population occurring by chance alone if the null hypothesis is

true (the assumption that there is no difference between different groups). The

significance level for the P-value is conventionally set at ≤0.05. This means that we are

prepared to accept no more than a 5% chance of being wrong in claiming that our

observed result is true. If the p-value around an estimate is >0.05, this indicates that the

data do not provide sufficient evidence to reject the null hypothesis.

Odds ratio

This is the odds or likelihood of an event, for example, an infection in an ‘exposed’

group of subjects compared to another group (‘non-exposed’). An odds ratio (OR) of 1

indicates that there is no difference in the odds of developing an SSI between the two

groups. An odds ratio of >1 indicates that the odds or likelihood of developing an SSI is

greater in the ‘exposed’ group compared to the ‘non-exposed’. An OR of <1 indicates

that the odds or likelihood of developing an SSI is higher in the ‘non-exposed’ group.


46

Section 10. References

[1] Chief Medical Officer. Surveillance of Healthcare Associated Infections PL CMO 2003(4).

Department of Health. 2003. Available at

http://webarchive.nationalarchives.gov.uk/20130107105354/http://www.dh.gov.uk/en/Publicationsa

ndstatistics/Lettersandcirculars/Professionalletters/Chiefmedicalofficerletters/DH_4003782

[2] Competition and Markets Authority(2016). CMA publishes final report on private healthcare

remittal. Available at:

https://www.gov.uk/government/news/cma-publishes-final-report-on-private-healthcare-remittal

[3] Public Health England. Protocol for the surveillance of surgical site infection. Version 6 June 2013.

Public Health England. 2013. Available at

https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/364412/Protocol_for

_surveillance_of_surgical_site_infection_June_2013.pdf

[4] Horan TC, Gaynes RP, Martone WJ et al. CDC definitions of nosocomial surgical site infections,

1992: a modification of CDC definitions of surgical wound infections. Am J Infect Control 1992;

20:271-274.

[5] Culver DH, Horan TC, Gaynes RP et al. Surgical wound infection rates by wound class, operative

procedure, and patient risk index. National Nosocomial Infections Surveillance System. Am J Med

1991; 91:152S-157S.

[6] European Centre for Diseased Prevention and Control. Surveillance of surgical site infections in

European hospitals - HAISSI Protocol. Version 1.02. Stockholm: ECDC; 2012.

[7] Meijerink H, Lamagni T, Eriksen HM et al. Is valid to compare surgical site infections rates between

countries? Insights from a study of English and Norwegian surveillance systems Infect Control

Hosp Epidemiol 2016; doi 10.1017/ice.2016.253

[8] National Collaborating Centre for Women's and Children's Health. Surgical site infection -

prevention and treatment of surgical site infection. Clinical Guideline. NICE. 2008. Available at

http://www.nice.org.uk/nicemedia/pdf/CG74FullGuideline.pdf

[9] Scottish Intercollegiate Guidelines Network. Antibiotic prophylaxis in surgery: A National Clinical

Guideline. Scottish Intercollegiate Guidelines Network. 2008. Available at

http://www.sign.ac.uk/pdf/sign104.pdf

[10] National Joint Registry. OPCS codes relevant to proecdures recorded in the NJR Version 3.

December 2013. Available at:

http://www.njrcentre.org.uk/

[11] Rochon M, Makhecha S, Morais C et al. Quality Improvement approach to reducing readmission

for surgical site infection. Wounds UK 2016; 12(2): 26-31

[12] National Institute for Health and Care Excellence. Healthcare-associated infections Quality

Standard, QS113. 2016. London, National Institute for Health and Care Excellence. Available at:

https://www.nice.org.uk/guidance/QS113/chapter/Quality-statement-1-Surveillance

[13] Thelwall S, Harrington P, Sheridan E et al. Impact of obesity on the risk of wound infection

following surgery: results from a nationwide prospective multicentre cohort study in England. Clin

Microbiol Infect 2015 21(11):1008.e1-1008.e8.

http://www.nice.org.uk/nicemedia/pdf/CG74FullGuideline.pdf

http://www.sign.ac.uk/pdf/sign104.pdf


47

[14] Lamagni, T, Elgohari, S, Harrington P. Trends in surgical site infections following orthopaedic

surgery in England. Curr Opin Infect Dis 2015 28(2):125-32.

[15] European Centre for Disease Prevention and Control. Surveillance of surgical site infections in

Europe, 2010-2011. European Centre for Disease Prevention and Control 2013. Available at:

http://ecdc.europa.eu/en/publications/Publications/SSI-in-europe-2010-2011.pdf

[16] National Institute for Health and Care Excellence. Surgical Site Infection Quality Standards, QS49.

2013. London, National Institute for Health and Care Excellence. Available at:

https://www.nice.org.uk/guidance/qs49

[16] Allegranzi B, Bischoff P, de Jonge S et al. New WHO recommendations on preoperative measures

for surgical site infection prevention: an evidence-based global perspective. Lancet Infec Dis 2016;

published online Nov 2

http://www.thelancet.com/pdfs/journals/laninf/PIIS1473-3099(16)30398-X.pdf

[17] Allegranzi B, Zayed B, Bischoff P et al. New WHO recommendations onintraoperative and

ostoperative measures for surgical site infection prevention: an evidence-based global perspective.

Lancet Infec Dis 2016; published online Nov 2

http://www.thelancet.com/pdfs/journals/laninf/PIIS1473-3099(16)30402-9.pdf

[18] Wilson J, Wloch C, Saei A et al. Inter-hospital comparison of rates of surgical site infection

following caesarean section delivery: evaluation of a multicentre surveillance study. J Hosp Infect

2013; 84:44-51

[19] Department of Health. Screening for Meticillin-resistant Staphylococcus aureus (MRSA)

colonisation - a strategy for NHS trusts: a summary of best practice. 2006. London: Department of

Health.

[20] Elgohari S, Wilson J, Saei A, et al. Impact of national policies on the microbial aetiology of surgical

site infections: analysis of trends between 2000 and 2013 using multi-centre prospective cohort

data Epidemiol Infect [Accepted November 2016]

[21] Public Health England (2016). Health Protection Report 10(39): news (11 November). Extenstion of

surveillance and infection control measures for Gram-negative infections. Available at:

https://www.gov.uk/government/publications/health-protection-report-volume-10-2016

[22] Streulens MJ, Plachouras D Mycobacterium chimaera infection associated with heater-cooler units

(HCU): closing another loophole in patient safety. Euro Surveill 2016 Nov 17; 21(46)

[23] Perkins KM, Lawsin A, Hasan NA et al. Notes from the field: Mycobacterium chimaera

contamination of heater-cooler devices used in cardiac surgery - Unites States. MMWR Morb

Mortal Wkly Rep 2016; 65:1117–1118

[24] Chand M, Lamagni T, Kranzer K et al. Insidious risk of severe Mycobacteraium chimaera infection in cardiac

surgery patients. Clin Infect Dis 2016; doi:10.1093/cid/ciw754

[25] Public Health England (2015). Mycobacterial infections associated with heater cooler units.

Guidance and updates concerning the risk of mycobacterial infections associted wih heater cooler

units used in cardopthoracic surgery. Available at:

https://www.gov.uk/government/collections/mycobacterial-infections-associated-with-heater-cooler-

units

surveillance of surgical site infections in nhs …...surveillance of surgical site infections in...

Documents