- This report provides a summary of S. aureus bacteraemia data for the first quarter of 2010 (January to March 2010) in 14 territorial NHS boards. This includes data on both meticillin resistant S. aureus (MRSA) and meticillin sensitive S. aureus (MSSA) bacteraemias; these data are compared to previous quarters.
- Altogether, 501 new S. aureus bacteraemia cases were reported to HPS during this period. This is a decrease of 9.9% compared with the same quarter last year (January to March 2009) when 556 cases were reported and the third lowest number of cases reported since the start of the mandatory S. aureus bacteraemia surveillance programme.
- In the last year, April 2009 to March 2010, 1983 episodes of S. aureus bacteraemia were reported to HPS. This represents a decrease of 10.9% on the previous year, April 2008 to March 2009, when 2226 episodes were reported.
- The overall rate of S. aureus bacteraemias for Scotland was 0.378 S. aureus bacteraemia cases per 1000 acute occupied bed days (AOBDs). This represents an increase of 3.8% compared to the previous quarter (from 0.365 to 0.378 per 1000 AOBDs). This small increase may be due to seasonal variations. This is the third lowest S. aureus bacteraemia rate reported since the start of the mandatory S. aureus bacteraemia surveillance programme.
- In comparison with the same quarter of 2009, the overall S. aureus bacteraemia rate for Scotland has decreased by 7.8%, from 0.411 to 0.378 per 1000 AOBDs.
- The rate of MRSA bacteraemias reported in the period January to March 2010 was the second lowest of any quarter reported since the start of this surveillance programme in January 2003 at 0.088 per 1000 AOBDs.
- During this quarter, the S. aureus, MRSA and MSSA bacteraemia rates for all NHS boards were within or below the 95% confidence limits of the funnel plots.
- The national surveillance programme reports on S. aureus bacteraemias occurring three months or longer before publication of this report. Therefore, this report does not replace the need for local monitoring of S. aureus bacteraemias by NHS boards themselves.
S. aureus is a gram positive bacterium which colonises the nasal cavity of about 30% of the healthy population. Although this colonisation is usually harmless, S. aureus may cause serious infections. These infections are commonly associated with healthcare interventions, often because of failures to implement infection prevention methods. As a result, both meticillin sensitive and meticillin resistant S. aureus (MSSA and MRSA) remain endemic in many UK hospitals, causing a range of infections. Amongst the most serious of these are bacteraemias.
The HPS S. aureus bacteraemia surveillance programme monitors the occurrence of S. aureus bacteraemias amongst all patients in Scotland. It includes S. aureus bacteraemias occurring in patients who have been in contact with the healthcare system (in both acute and non-acute hospitals, as well as in primary care settings) and those who have acquired S. aureus bacteraemias in the community, without any healthcare contacts.
The surveillance programme in Scotland includes data on both MRSA and MSSA bacteraemias. Many other countries restrict surveillance of S. aureus bacteraemias to those caused only by MRSA.
This quarterly report is concerned with the incidence of S. aureus bacteraemias in Scotland and within individual NHS boards. HPS publishes an annual report on S. aureus bacteraemias in Scotland.1 This annual report contains further analyses of S. aureus bacteraemia data, including S. aureus bacteraemia trends. Our experience has shown that quarterly reporting has been too frequent for meaningful conclusions about patterns of antibiotic resistance in S. aureus bacteraemias. HPS is now reporting annually on antimicrobial susceptibility.
The quarterly reports of S. aureus bacteraemia data produced by HPS are partly based on interim data for both bed occupancy and incident S. aureus bacteraemias. These data are subject to revision as finalised data become available. Therefore there may occasionally be minor numeric discrepancies between successive reports, reflecting the availability of such updated data.
Finally, it should be noted that the national S. aureus bacteraemia surveillance programme contains information on bacteraemias which occurred at least three months and possibly up to six months before this report was published. It is therefore imperative that NHS boards have their own local S. aureus bacteraemia surveillance systems to alert local teams to the need for possible action. This surveillance report does not replace local monitoring of S. aureus bacteraemias by NHS boards’ infection control teams.
2.1 Data sources
In Scotland, all (fourteen) NHS boards report all isolates of MSSA or of MRSA from blood cultures to HPS. These reports come from routine hospital laboratory data systems.
All NHS boards’ laboratories submit every first isolate of S. aureus from a patient’s blood to the Scottish MRSA Reference Laboratory (SMRSARL) for typing and antimicrobial susceptibility testing, under the European Antimicrobial Resistance Surveillance System (EARSS) programme.2 On rare occasions, NHS boards might not submit an S. aureus blood isolate to the SMRSARL, so typing and antimicrobial susceptibility will not be performed. This would lead to discrepancies between the numbers of S. aureus bacteraemias identified in local laboratories and those processed by SMRSARL.
Episodes of S. aureus bacteraemia are identified for this quarterly surveillance report from both the routine hospital laboratory systems and from the SMRSARL system. While most of these data are transmitted to HPS via the ECOSS electronic reporting system,3 some laboratories still have a direct reporting mechanism for their S. aureus bacteraemia isolates.
This report is based on data from all S. aureus bacteraemias reported to HPS between January 2010 and March 2010. These numerator data are derived from de-duplicated merged S. aureus bacteraemia case data supplied by both NHS boards’ laboratories and the SMRSARL.
Acute occupied bed days (AOBDs) statistics for individual NHS boards are provided by the Information Services Division (ISD) of NHS National Services Scotland (NSS).4 These AOBDs include some estimated bed occupancy data due to incomplete data availability at the time of this publication.
An episode of S. aureus bacteraemia (MRSA or MSSA) is defined as a person from whose blood MRSA or MSSA has been isolated and reported by a diagnostic microbiology laboratory, in the absence of a positive blood culture in the previous two weeks.5
One acute occupied bed day (AOBD) is a hospital acute bed which was occupied at midnight. An acute bed is any bed other than a care of the elderly long stay or psychiatric bed.5
Cases are allocated to individual NHS boards based on the location of the diagnostic laboratory where the specimen was tested.
2.3 Data analysis and reporting
S. aureus bacteraemia data are shown as simple numerator data in the run charts.
Calculation of rates
S. aureus bacteraemia data are also expressed as rates. These are calculated using a denominator of the number of acute occupied bed days (AOBDs) for the relevant quarter. These S. aureus bacteraemia rates are shown for each NHS board, enabling comparisons between NHS boards.
Interpretation of rates
MRSA bacteraemias are presented using Statistical Process Control (SPC) charts.6,7 However, MSSA bacteraemias (and hence all S. aureus bacteraemias) are not analysed in this way because the MSSA bacteraemia surveillance scheme has not operated for as long, so there are not currently sufficient data points to enable construction of valid MSSA SPC charts.
In interpreting these MRSA SPC charts, HPS has adopted the following eight criteria to assess whether a substantial change or unnatural variation in the S. aureus bacteraemia rate has occurred:
- A single value either above the upper control limit or below the lower control limit.
- Eight consecutive values on the same side of the central line (or mean).
- Any twelve out of fourteen consecutive values on the same side of the central line (or mean).
- Three consecutive values in either the top third (above upper warning limit) or bottom third (below lower warning limit) of the expected range.
- Five consecutive values in the top two-thirds (above upper highlight limit) or bottom two-thirds (below lower highlight limit) of the expected range.
- Thirteen consecutive values in the middle thirds (above the lower highlight limit and below the upper highlight limit) of the expected range.
- Eight consecutive values either increasing or decreasing.
- Any cyclic or periodic behaviour.
Comparison of NHS boards’ rates
Comparisons of NHS boards’ S. aureus bacteraemia, MRSA and MSSA rates are shown using funnel plots.8 These are scatterplots of the observed S. aureus bacteraemia events, measured as rates, plotted against their precision, indicated by the number of acute occupied bed days. The plots in this report show the upper and lower 95% confidence limits as curved lines. If an individual NHS board’s rate was outwith the 95% confidence limit, this would be regarded as an outlier, suggesting that that NHS board had an S. aureus bacteraemia rate which was significantly different to other NHS boards.
Analysis of trends
Analysis of trends in the rates of S. aureus bacteraemias was carried out within an over-dispersed Poisson regression model with the logarithm of the total occupied bed days as an offset. The regression included terms for NHS board, year and quarter. Hypothesis tests were carried out using F tests based upon the large sample approximation to the normal distribution. Model checking was performed using residual plots and these demonstrated that the model assumptions were reasonable. Trend analyses include the first eighteen quarters of data collection.
3.1 Surveillance data
During the period January to March 2010, HPS received reports of S. aureus bacteraemias from 23 diagnostic laboratories in 14 NHS boards as well as from SMRSARL.
The total number of S. aureus bacteraemia cases identified in Scotland was 501, of which 117 were MRSA bacteraemias (23.4%) and 384 were MSSA bacteraemias (76.6%) as shown in Table 1. The total numbers of S. aureus bacteraemias in Scotland increased by 5.5% compared with the previous quarter (October to December 2009) where 475 S. aureus bacteraemia cases were recorded. In that quarter 117 (24.6%) were due to MRSA and 358 to MSSA.
The overall S. aureus bacteraemia rate for Scotland during this quarter was 0.378 per 1000 AOBDs. This is 7.8% below the overall rate for Scotland for the corresponding quarter in the previous year. This represents the third lowest number and rate of S. aureus bacteraemias since the incorporation of MSSA bacteraemias into this surveillance system.
The most recent quarter, January to March 2010, has the second lowest rate of MRSA bacteraemias of any quarter reported to this surveillance programme since it began in January 2003.
The total number of S. aureus bacteraemia cases and the S. aureus bacteraemia rates for each NHS board during the current quarter, January to March 2010, are listed in Table 2, which also indicates the numbers and rates of S. aureus bacteraemia in the preceding year, January 2009 to December 2009.
Figure 1 shows the quarterly numbers of S. aureus bacteraemias in Scotland from April 2005 to March 2010, presented as a run chart.
3.2 Comparison of NHS boards’ rates
Figure 2a shows a funnel plot of S. aureus bacteraemia rates for individual NHS boards for this quarter, Figures 2b and 2c show similar funnel plots for MRSA and MSSA bacteraemias, respectively. They show that none of the S. aureus bacteraemia rates in NHS boards were above the 95% confidence limit.
3.3 Identification of trends
The overall quarterly rates for Scotland are shown in Figure 3. A Poisson regression analysis of trends in the S. aureus bacteraemia rate between April 2005 and March 2010 was carried out. This showed a significant year on year reduction in the rate of S. aureus bacteraemias for NHS Scotland. This was estimated to be 6.3% (95% CI: 3.8%, 8.6%). The rate of MRSA bacteraemias also showed a significant year on year reduction estimated at 15.3% (95% CI: 12.4%, 18.0%) whereas the rate of MSSA bacteraemias showed a smaller non significant decrease. This was estimated to be 1.6% (95% CI: -1.8%, 4.8%).
Figure 4 shows the MRSA bacteraemia rate per 1000 AOBDs
for Scotland as a Statistical Process Control (SPC) chart. This
chart has control limits set at three standard deviations (sd) from
the mean, with warning limits set at two sd from the mean and
highlight limits set at one sd from the mean. The centre line is
the mean of all results to date.
Figure 4 shows that the seven most recent quarterly MRSA bacteraemia rates were below the lower control limit and the ten most recent MRSA bacteraemia rates were below the centre line on the SPC chart. This indicates that a change has occurred in MRSA bacteraemia rates in Scotland.
Figure 5 shows run charts of the quarterly numbers of S. aureus bacteraemias (MRSA and MSSA) for each NHS board, plotted against time. From the run charts NHS Borders and NHS Western Isles have had high rates in the last quarter.
Figure 6 in the Appendix shows SPC charts of quarterly MRSA bacteraemia rates per 1000 AOBDs for each NHS board. The MRSA bacteraemia rate for NHS Borders has risen above the upper control limit indicating a substantial rise in rate during this quarter. Whereas the rate of MRSA bacteraemia within NHS Greater Glasgow and Clyde, NHS Lothian and NHS Tayside have all shown substantial decrease in recent quarters data.
This report covers S. aureus bacteraemia surveillance data for the period January to March 2010. Unlike the national surveillance schemes in many other countries, it includes both MRSA and MSSA bacteraemias.
A number of important caveats associated with the data in this report should be highlighted. These include the use of estimated denominator data, self reporting of numerator data by participating laboratories, the inclusion of some S. aureus bacteraemias which may have been due to contaminated cultures, the inclusion of some S. aureus bacteraemias which may have been community-acquired and only diagnosed on admission to hospital and the possibility that some patients with prolonged bacteraemic episodes may have been treated as multiple cases, by strict application of the surveillance case definition.
A total of 501 cases of S. aureus bacteraemia were identified during this period. This represents an increase of 3.8% compared with the total number from the previous quarter. This small increase may be due to seasonal variations.
The proportion of S. aureus bacteraemias which were meticillin resistant during the quarter January to March 2010 was 23.4%. The 2008 figures for the UK as published in the EARSS Annual Report 2008 show the proportion of S. aureus bacteraemia which were meticillin resistant to be 31%.1
Funnel plot analysis was used to examine variations in NHS board S. aureus bacteraemia rates. No NHS boards were above the 95% confidence intervals in the S. aureus bacteraemia funnel plots.
The national S. aureus bacteraemia rate per 1000 AOBDs during the period April 2005 to March 2010 showed a significant downward trend and this was estimated as a reduction of 6.3% (95% CI: 3.8%, 8.6%).
This report uses SPCs to distinguish between natural variation and real changes in the incidence of MRSA bacteraemias. Scottish MRSA bacteraemia rates have fallen in the past two years, confirmed by Figure 4 which shows that several of the criteria for significant change have now been achieved. In particular, the seven most recent Scottish rates were below the lower control limit and the last ten rates were below the centre line (mean). Some individual NHS boards have also met criteria for significant change.
The centre lines on the SPCs will be adjusted to reflect these significant improvements after the current HEAT target assessment period has ended, in March 2010. Lowering the centre line will be recognition of sustained improvements in infection control and the expectation that the NHS board will be able to sustain these improved results. This adjustment will be applied where there have been eight consecutive results below the centre line and the new centre line will be based on the eight results which were below the old centre line.5,6
Health Protection Scotland is grateful to all of the microbiologists throughout Scotland who provided the S. aureus bacteraemia data for this report and helped in its preparation, staff at the Scottish MRSA Reference Laboratory who provided the EARSS data and to the Information Services Division of National Services Scotland for providing the hospital activity data. Our thanks also go to Chris Robertson (Professor of Public Health Epidemiology, HPS and University of Strathclyde), Gwen Allardice (HPS and University of Strathclyde) and to Traini Stari (HPS and University of Strathclyde) for their statistical support.
1. The Annual Surveillance of Healthcare Associated Infections Report January - December 2009. Available at http://www.hps.scot.nhs.uk/haiic/sshaip/publicationsdetail.aspx?id=44385
2. European Antimicrobial Resistance Surveillance System. EARSS Annual Report 2008. Bilthoven: European Centre for Disease Prevention and Control (ECDC); 2009.
3. Information Paper for the Notifications Sub-Group, Public Health etc. (Scotland) Act 2008. National Systems Relating to Notifications. Available at http://www.scotland.gov.uk/Resource/Doc/924/0076150.doc
4. Information Services Division of National Services Scotland. Home page http://www.isdscotland.org/isd/CCC_FirstPage.jsp
5. Protocol for the Scottish mandatory surveillance programme for Staphylococcus aureus bacteraemia. Available at http://www.documents.hps.scot.nhs.uk/hai/sshaip/guidelines/s-aureus/s-aureus-bacteramia-protocol-v3-2007-09.pdf
6. Benneyan JC. Statistical quality control methods in infection control and hospital epidemiology, Part I: Introduction and basic theory. Infect Control Hosp Epidemiol. 1998; 19:194-214.
7. Benneyan JC. Statistical quality control methods in infection control and hospital epidemiology, Part II: Chart use, statistical properties, and research issues. Infect Control Hosp Epidemiol. 1998; 19(4):265-83.
8. Spiegelhalter D.J. Funnel plots for comparing institutional performance. Stat Med. 2005; 24(8):1185-202.