MRSA screening guidance

August 2014

Implementation of modified

admission MRSA screening

guidance for NHS (2014)

Department of Health expert advisory committee on

Antimicrobial Resistance and Healthcare Associated

Infection (ARHAI)

You may re-use the text of this document (not including logos) free of charge in any format

or medium, under the terms of the Open Government Licence. To view this licence, visit

www.nationalarchives.gov.uk/doc/open-government-licence/

Published to gov.uk, in PDF format only.

www.gov.uk/dh

Implementation of modified

admission MRSA screening

guidance for NHS (2014)

Department of Health expert advisory committee

on Antimicrobial Resistance and Healthcare

Associated Infection (ARHAI)

Prepared by ARHAI MRSA Screening Implementation Group

Disclaimer

The MRSA Screening Implementation Group, comprising members from key

professional groups, was established to determine how to give implementation guidance

on MRSA screening to NHS England, in the context of recommendations from the NOW

study.

The Working Group members were: Mark Wilcox (Chair, ARHAI), Peter Cowling (BIA),

Brian Duerden, Carole Fry (DH), Susan Hopkins (PHE), Peter Jenks (HIS), Sally

Kingsland (NHS England), and Sally Palmer (IPS).

Contents

Contents ..................................................................................................................................... 4

Executive summary .................................................................................................................... 5

Introduction ................................................................................................................................ 6

Objective .................................................................................................................................... 8

High risk specialties/units ................................................................................................ 8

Interventions .............................................................................................................................. 9

Compliance and audit .............................................................................................................. 10

References ............................................................................................................................... 11

Appendix .................................................................................................................................. 12

Annex: Consultation ................................................................................................................. 16

Executive summary

The following guidance outlines a more focused, cost-effective approach to MRSA

screening. The recommendation for Trusts to move to focussed screening programmes

has been designed to promote a more efficient and effective method for identifying and

managing high risk MRSA positive patients. Importantly, focussed screening should be

adopted in line with local risk assessments to ensure that Trusts concentrate on reducing

negative patient outcomes for their own populations. Changes to current practice need to

be undertaken with a commitment to improved compliance with focussed screening, which

should be monitored and reported to Trust Boards and commissioners. Trusts will need to

regularly review (and where necessary improve) their compliance with national screening

guidance for each specialty, recent MRSA infection data, patient demographics and types

of services provided within individual organisations. Continued surveillance through both

local and national surveillance systems will be needed to monitor the levels of MRSA

infection. The outputs of the NOW study could be used to inform policy making decisions

should the prevalence of MRSA increase.

Introduction

In the last decade there have been very marked declines in morbidity and mortality related to

MRSA infection in England. Annual MRSA bacteraemia rates fell from 17.7 (April 2005-March

2006) to 3.2 cases per 100, 000 bed days (April 2011-March 2012).

Significant declines have

also been observed in surgical site infections (SSI) where MRSA was reported as the causative

micro-organism (from 27% in 2004-6 to only 4% in 2011/12).

2,3

The number of death

certificates in England and Wales mentioning MRSA infection has decreased each year since

2006, when the figure peaked at 1,652; in 2012, MRSA accounted for 292 mentions of MRSA

on death certificates (a 20% decrease on the previous year).

Until April 2009, national guidance in England recommended targeted screening of patients in

high risk specialties.

There were no randomised controlled trials, however, to provide

evidence on the most effective and cost-effective screening strategies. The Department of

Health (DH) in England introduced mandatory screening of all elective and emergency

admissions from April 2009 and December 2010, respectively. This decision was based on a

DH impact assessment that modelled the cost-effectiveness of different screening and

decolonisation strategies in preventing MRSA bacteraemias, wound infections and deaths. We

note that in other settings (e.g. Wales), where mandatory screening has not been implemented,

MRSA infection rates have fallen markedly.

The DH impact assessment committed to a review

of this policy with additional data; thus, the NOW study was commissioned in 2011. The study

report underwent peer review and its findings were endorsed by ARHAI.

A full report of the NOW study and a summary of its findings are available.

7,8,9

The study

showed that compliance with the current mandatory screening policy was poor (e.g. only 61%

were screened; about half of new positives were isolated when their result became known; and

about a quarter did not receive decolonisation therapy) (Appendix, Table 1). The prevalence of

MRSA in new admissions was low (1.5% overall), although this varied according to type of

admission (2.1% in emergency admissions, 0.9% elective admissions and 0.7% in day cases

admissions) (Appendix, Tables 2 &3). These observations mean that the numbers of patients

needed to be screened in order to identify one new positive were high in all admission types

(emergency n=102; elective n=180; and day case n=186).

The NOW study went on to model the effectiveness and cost-effectiveness of six different

screening strategies at a whole hospital level for the three categories of NHS Trust (acute,

teaching and specialty) at four different levels of MRSA prevalence (the current prevalence; and

twice, three and four times the current prevalence) and two levels of transmission (current and

twice the estimated transmission rate). Six screening and intervention strategies were

evaluated: 1) no screening (interventions applied to clinical cases only), 2) screening all

admissions (emergency and elective), 3) screening all admissions to high risk specialties, 4)

checklist activated screening of all admissions, 5) screening all admissions to high risk

specialties plus checklist activated screening of other admissions, and 6) screening all

admissions plus pre-emptive isolation of those known to be previously MRSA positive.

At current prevalence none of the screening strategies was likely to be cost-effective at

conventional NHS levels of ‘willingness to pay’ (less than £30,000 per QALY). Costs per QALY

for routine admission screening ranged from £86,000 - £170,000, and were consistently more

costly and less effective than alternatives for all hospital types. However, of the strategies

involving active screening, targeting high risk specialty patients was the optimal option; i.e.

mean incremental cost-effectiveness ratios (ICERs) of approximately £45,000 and £48,000 in

Acute, and Teaching Trusts, respectively (Appendix, Table 4) (see below for Specialist trust

results). As prevalence rose, the cost effectiveness of this strategy increased, and fell within

Implementation of modified admission MRSA screening guidance for NHS (2014)

the conventional NHS levels of ‘willingness to pay’ (Appendix, Table 4). The model results were

strongly influenced by the assessment of the risk of infection, which is greater in high risk

specialty patients, and has the largest impact on length of stay and mortality, the major

determinants of cost and health benefits. Importantly, lack of screening of MRSA in patients

admitted to low-risk specialties results in more transmission, but less proportionate risk of

serious infection and death than in high risk specialties. Approximately 60% of MRSA colonised

individuals will be detected by screening those known to be previously MRSA positive and high

risk specialty admissions. For acute and teaching trusts, the most cost effective of the

screening strategies is screening high risk specialties only, at both the current prevalence and

up to 4 times current prevalence (~5%). For specialist trusts, at current and twice current

prevalence, again the ‘no screening’ strategy has the greatest probability of being the most

cost-effective option. However, of the active screening strategies, screening high risk

specialities and performing check list activated screening of others is optimal. At higher

prevalence (>3%), high risk specialty screening alone becomes the most cost effective strategy

overall. However, the probability of cost-effectiveness for high risk specialty screening and high

risk specialty plus checklist activated screening cluster together closely. Thus, the most

pragmatic decision is that high risk screening for all trust types is the simplest and most cost

effective of the screening strategies (Appendix, Table 5).

A recent study examined MRSA control in an ICU setting in the US, (where all patients were

nursed in side-rooms), comparing screening and isolation with targeted or universal

decolonisation.

Universal decolonisation was found to be effective at reducing rates of MRSA

clinical isolates and any pathogen bloodstream infections in ICUs. However, universal

decolonisation has not been validated in this way outside the US or outside of the ICU setting,

and so no recommendation regarding this approach can be made at this time.

In summary, the results of the NOW study suggest that the current mandatory MRSA screening

policy is followed in less than two-thirds of admissions, but that even if compliance was 100%, it

would still not be cost-effective in any trust type or scenario. While no active screening strategy

was optimal at current MRSA levels, the most cost effective policy is one based on screening

admissions to high risk specialties; cost-effectiveness increases as MRSA prevalence rises.

Trusts would need however, to take measures to ensure high compliance with this strategy.

Objective

To focus and maximise the clinical impact for patients (adults and children) who are most likely

to benefit (i.e. those patients for whom MRSA colonisation carries the greatest risk of

infection/poor outcome), it is recommended that the current practice of mandatory MRSA

screening of acute and elective admissions to NHS hospitals in England is streamlined to the

following:

• All patients admitted to high risk units (defined below).

• All patients previously identified as colonised with or infected by MRSA.

High risk specialties/units

High risk specialties are defined as vascular, renal/dialysis, neurosurgery, cardiothoracic

surgery, haematology/oncology/bone marrow transplant, orthopaedics/trauma, and all

intensive care units (adult/paediatric ICUs, Neonatal Intensive Care Units, High dependency

units, Coronary Care Units).

• In addition, local risk assessment should be used to define other potential high MRSA

risk units/specialties; for example, according to provision of specialised services (e.g.

transplant, neonatal), and units with a history of high endemicity of MRSA. Local risk

assessment may increase the proportion of detected MRSA colonised individuals,

notably those patients at risk of poor outcome from MRSA.

Implementation of modified admission MRSA screening guidance for NHS (2014)

Interventions

1. Clear guidance on the local policy for MRSA screening should be made available to all

staff; this should be used as a standard against which audit is carried out.

2. Trusts should identify and screen patients in high MRSA risk specialties (as above).

3. Trusts should identify and re-screen any patient previously known to be MRSA positive

and isolate these pre-emptively, pending the results of laboratory tests.

4. NICE accredited Standards for Microbiology Investigations (SMIs) are available for

MRSA screening.

The most commonly practised methods of MRSA screening that

were recorded in the NOW study were based on culture on chromogenic agar (5% of

trusts routinely used PCR for emergency screens; 15% used PCR for some emergency

screening). Average MRSA screen turnaround times were 2.87 days for MRSA+ve and

1.75 days for MRSA-ve samples.

5. The most frequently sampled body sites included in current MRSA screening, as

identified in the NOW Study (i.e. as practised by at least 75% of trusts), are the nose,

groin/perineum and ‘other’ sites where appropriate (e.g. wounds, indwelling devices,

throat, etc).

6. The frequency of (repeat) screening should be determined locally and made explicit in

local guidance.

7. If MRSA transmission is detected/suspected in hitherto ‘low risk’ specialties then local

risk assessment should be used to determine the need for and extent of MRSA

screening, i.e. as befits normal infection prevention & control practice when investigating

clustering of cases or transmission.

8. All patients identified as MRSA positive must receive decolonisation/suppression

therapy; MRSA positive patients should normally be isolated until such time that MRSA

colonisation has been shown to be no longer present or local risk assessment

determines that isolation can be safely discontinued.

9. Local risk assessment in terms of isolation and decolonisation/suppression therapy will

be required for some persistently positive MRSA patients.

10. Trusts should continue to monitor MRSA rates.

Compliance and audit

1. Trusts should make every effort to ensure very high levels of screening in the patient

groups identified above.

2. Regular measurement/audit should be carried out to demonstrate compliance with local

MRSA screening guidance.

3. Audit should specifically include measurement of the appropriateness of decolonisation

(correct agents/dosages for correct time) and follow up.

4. Trusts should make performance/compliance data on MRSA screening and

decolonisation widely available within their organisations.

5. The Director of Infection Prevention and Control should provide assurance to the Trust

Board on the level of compliance with the local policy on MRSA screening

/decolonisation.

6. A qualitative study on the patient experience of MRSA screening and the impact of a

positive result found that the following are essential to securing and sustaining patients’

satisfaction and confidence in the care they receive in relation to preventing MRSA

infection:

1,2

i. patients need to be informed of the result of their screen, even if negative;

ii. information needs to be provided in an individualised way (both written and

verbal);

iii. staff need to be sufficiently knowledgeable and confident to invite patients’ and

carers’ questions and communicate information in a sensitive way;

iv. specific and comprehensive guidelines for home-based decolonisation are

required;

v. patients expect to see that standards of cleanliness and infection prevention, such

as hand hygiene, are practised; and

vi. measuring and acting on feedback from the patient experience is necessary.

Implementation of modified admission MRSA screening guidance for NHS (2014)

References

1. Available at: http://www.hpa.org.uk/

2. Health Protection Agency. Surveillance of surgical site infections in NHS hospitals in

England, 2011/2012. London: Health Protection Agency, December 2012. Available

from: www.hpa.org.uk

3. Health Protection Agency. Sixth report of the mandatory surveillance of surgical site

infection in orthopaedic surgery, April 2004 to March 2010. London: Health Protection

Agency, December 2010.

4. ONS. Deaths Involving MRSA: England and Wales, 2008 to 2012.

5. Coia JE, Duckworth GJ, Edwards DI, Farrington M, Fry C, Humphreys H, Mallaghan C,

Tucker DR; Joint Working Party of the British Society of Antimicrobial Chemotherapy;

Hospital Infection Society; Infection Control Nurses Association. Guidelines for the

control and prevention of meticillin-resistant Staphylococcus aureus (MRSA) in

healthcare facilities. J Hosp Infect 2006; 63 Suppl 1: S1–44.

6. Public Health Wales. Clostridium difficile and Staphylococcus aureus Bacteraemia

Monthly Update. Available at:

http://www.wales.nhs.uk/sites3/page.cfm?orgid=379&pid=67899 Last accessed 30 June

2014.

7. Fuller C, Robotham J, Savage J, Deeny S, Hopkins S, Cookson B, Stone S. Final report

of national one week prevalence audit of MRSA screening. Available at:

www.idrn.org/audit Last accessed 30 June 2014.

8. Fuller C, Robotham J, Savage J, Hopkins S, Deeny SR, et al. (2013) The National One

Week Prevalence Audit of Universal Meticillin-Resistant Staphylococcus aureus (MRSA)

Admission Screening 2012. PLoS ONE 8(9): e74219. doi:10.1371/journal.pone.0074219

9. Advisory Committee on Antimicrobial Resistance and Healthcare Associated Infection

(ARHAI). National one week (NOW) prevalence audit of MRSA screening report (ARHAI

22-13(02)). Available at: https://www.gov.uk/government/policy-advisory-

groups/advisory-committee-on-antimicrobial-resistance-and-healthcare-associated-

infection#minutes Last accessed 22 November 2013.

10. Huang SS, Septimus E, Kleinman K, et al. Targeted versus universal decolonization to

prevent ICU infection. N Engl J Med 2013;368:2255-65.

11. Public Health England. B 29 - Investigation of Specimens for Screening for MRSA.

Available at:

http://www.hpa.org.uk/ProductsServices/MicrobiologyPathology/UKStandardsForMicrobi

ologyInvestigations/TermsOfUseForSMIs/AccessToUKSMIs/SMIBacteriology/smiB29Inv

estigationofSpecimensforScreeningforMRSA/ Last accessed 22 November 2013.

12. Loveday H et al. The patient experience of the MRSA screening process and the impact

of a MRSA-positive result: a qualitative study. Antimicrobial Resistance and Infection

Control 2013;2 (Suppl 1):P56 http://www.aricjournal.com/content/2/S1/P56

MRSA Screening Implementation Group

April 2014

Updated post-consultation, June 2014

Appendix

Table 1. NOW Study: Proportion of admissions screened for MRSA

Admission Category

All

Acute

Specialist

Teaching

Emergency

52788/87165

(61%)

Median 67.3%

IQR (47.5-85.8%)

n=132 trusts

38127/63577

(60%)

Median 67.1%

IQR (47.4-85.8%)

n=91

657/1166

(56%)

Median 85.9%

IQR (68.3-100%)

n=16

13736/21988

(62.5%)

Median 59.4%

IQR (48.9-89.2%)

n=22

Elective (not including

day-cases)

22773/27838

(81.8%)

Median 90%

IQR (58-118%)

n=115 trusts

14477/16497

(87.7%)

Median 92%

IQR (59-136%)

n=77

1652/2191

(75.4%)

Median 86%

IQR (62-100%)

n=16

6569/9044

(72.6%)

Median 73%

IQR (30-102%)

n=20

Day-cases (not

including dermatology,

endoscopy, ophthalmic

and paediatrics)

22416/46777

(47.9%)

Median 90%

IQR (23.2-78.9%)

n=110 trusts

14255/32927

(43.3%)

Median 36.5%

IQR (17.4-73.9%)

n=77

1153/1568

(73.5%)

Median 67.3%

IQR (42.6-100%)

n=13

6894/11927

(57.8%)

Median 48.3%

IQR (36.1-77.7%)

n=19

Table 2. NOW Study: Proportion of Admission Screens MRSA positive

Admission Category

All

Acute

Specialist

Teaching

Emergency

1075/52064

(2.1%)

Median 1.6%

IQR (1.1-2.7%)

n=129 trusts

836/37408

(2.2%)

Median 2%

IQR (1.2-2.7%)

n=90

5/652

(1%)

Median 0%

IQR (0-0.2%)

n=16

230/13736

(1.7%)

Median 1.7%

IQR (1.1-2.4%)

n=22

Elective (not including

day-cases)

188/20798

(0.9%)

Median 0.7%

IQR (0-1.9%)

n=101 trusts

110/13532

(0.8%)

Median 0.7%

IQR (0-1.8%)

n=68

25/1488

(1.7%)

Median 0.7%

IQR (0-2.5%)

n=15

53/5703

(0.9%)

Median 0.5%

IQR (0.3-1.5%)

n=14

Day-cases (not

including dermatology,

endoscopy, ophthalmic

and paediatrics)

150/21501

(0.7%)

Median 0%

IQR (0-1%)

n=112 trusts

58/13509

(0.4%)

Median 0%

IQR (0-0.6%)

n=76

6/1062

(0.6%)

Median 0%

IQR (0-1.1%)

n=16

85/6816

(1.2%)

Median 0.7%

IQR (0.3-1.2%)

n=19

Implementation of modified admission MRSA screening guidance for NHS (2014)

Table 3. NOW Study: Proportion of admission screens newly positive for MRSA

Admission Category

All

Acute

Specialist

Teaching

Emergency

498/50739

(1.0%)

Median 0.9%

IQR (0.4-1.3%)

n=127 trusts

374/36083

(1.0%)

Median 1.0%

IQR (0.5-1.5%)

n=88

4/652

(0.6%)

Median 0%

IQR (0-0%)

n=16

119/13736

(0.9%)

Median 0.8%

IQR (0.5-1.3%)

n=22

Elective (not including

day-cases)

107/19283

(0.6%)

Median 0.4%

IQR (0-1.2%)

n=98 trusts

68/12953

(0.5%)

Median 0.4%

IQR (0-1.7%)

n=68

16/1346

(1.2%)

Median 0.4%

IQR (0-1.5%)

n=14

23/4909

(0.5%)

Median 0.5%

IQR (0-1.4%)

n=15

Day-cases (not

including dermatology,

endoscopy, ophthalmic

and paediatrics)

79/20461

(0.4%)

Median 0%

IQR (0-0.1%)

n=110 trusts

27/12469

(0.2%)

Median 0%

IQR (0-0.2%)

n=74

5/1062

(0.5%)

Median 0%

IQR (0-0%)

n=16

47/6816

(0.7%)

Median 0.1%

IQR (0-0.7%)

n=19

Table 4. NOW Study: Cost-effectiveness of screening strategies by trust type and MRSA

admission prevalence. Incremental cost per QALY gained is shown for strategies considered

both cost-effective

†

and non-cost-effective. Any remaining strategies for each prevalence

scenario were dominated

††

Trust

type

MRSA prevalence

on admission

Cost-effective strategies : Mean

(range

†††

Non cost-effective strategies : Mean

(range

†††

ACUTE

Baseline -

Strategy 3: £45,198/QALY

(£35,314-£61,390)

Strategy 5: £216,449/QALY

(£112,948-£1,770,724)

Moderate (x2) -

Strategy 3: £48,655/QALY

(£38,417-£64,899)

Strategy 6: £70,930/QALY

(£43,581-£161,673)

Strategy 5: £72,545/QALY

(£42,552-£206,075)

High (x3)

Strategy 3 :£29,565/QALY

(£25,660-£34,389)

Strategy 5: £159,566/QALY

(£95,476-£412,113)

Very high (x4)

Strategy 3: £28,708/QALY

(£25,479-£32,609)

Appendix

Trust

type

MRSA prevalence

on admission

Cost-effective strategies : Mean

(range

†††

Non cost-effective strategies : Mean

(range

†††

TEACHING

Baseline -

Strategy 3: £47,936/QALY

(£34,585-£74,757)

Moderate (x2) -

Strategy 3: £33,751/QALY

(£26,828-£43,702)

Strategy 2: 1,425,323/QALY

(£213,225-cost/health-loss)

High (x3) -

Strategy 3: £43,686/QALY

(£33,921-£59,612)

Strategy 5: £175,973/QALY

(£78,509-cost/health-loss)

Very high (x4) -

Strategy 3: £37,369/QALY

(£31,218-£37,369)

Strategy 6: £239,808/QALY

(£129,780-£11,238,684)

SPECIALIST

Baseline -

Strategy 5: £62,566/QALY

(£47,979-£89,425)

Moderate (x2) -

Strategy 5: £31,248/QALY

(£27,276-£36,499)

High (x3)

Strategy 3: £24,009/QALY

(£20,764 - £28,362)

Strategy 5: £26,411/QALY

(£17,071-£54,549)

Very high (x4)

Strategy 3: £19,331/QALY

(£17,295-£21,860)

Strategy 5: £24,503/QALY

(£16,421 - £45,503)

† An ICER of less than £30,000 per QALY is considered cost-effective. An ICER of more than £30,000 is not

considered cost-effective. £30,000 is the upper limit of the usual NHS willingness to pay range.

†† Dominated strategies are those that are more costly and provide less benefit than one other strategy or a

combination of two other strategies. Since it can never be cost-effective to pay more for less benefit, ICERs were

not calculated for these strategies.

††† ICER ranges were calculated using mean costs and QALYs ±1standard error, where the minimum = smallest

possible difference in cost / greatest possible difference in health benefits, and maximum = greatest possible

difference in cost/ smallest possible difference in health benefits.

*Further bootstrapped confidence intervals will be added.

Implementation of modified admission MRSA screening guidance for NHS (2014)

Table 5. Cost-effectiveness acceptability frontiers. Lines depict the optimal strategies (i.e. those

with the highest expected net monetary benefit) dependent on the willingness to pay for health

benefits, while dotted vertical lines show the willingness to pay values at which the decision

changes.

Acute

Teaching

Specialist

Baseline

prevalence

Acute: 1.4%

Teaching : 1.3%

Specialist: 1.04%

2x prevalence

Acute: 2.8%

Teaching: 2.6%

Specialist: 2.08%

3x prevalence

Acute: 4.2%

Teaching: 3.9%

Specialist: 3.12%

4x prevalence

Acute: 5.6%

Teaching: 5.2%

Specialist: 4.13%

Twice

transmission

Annex: Consultation

Consultation on ‘Implementation of modified admission MRSA screening guidance for NHS (2014)’

Comments

(the names of individuals providing comments are withheld here)

Response

A sensible and well written document. However it is rather "England -centric"

for the BIA. There is no mention of the Welsh experience where universal

screening was always eschewed and where rates of MRSA have also fallen.

Thank you.

Advice for Wales is not our remit. However, the observation

about declining MRSA rates in Wales in the absence of a

mandatory screening programme is pertinent – we have

added a note to our discussion.

Thank you for these draft guidelines, I do support the strategies proposed.

My only comment is that I would like to see similar guidelines provided in due

course for community hospital and mental health settings as specific

guidance has been previously issued, for example:

http://webarchive.nationalarchives.gov.uk/20101125133833/http://www.clean-

safe-care.nhs.uk/Documents/MRSA_Emergency_Screening_-_FAQs_-

_Apr_2010.pdf

Our own experience in mental health is that if all patients listed within the

possible risk groups are screened, very few are detected that were not known

MRSA positives. As such we have stepped down to screening only known

previous MRSA positives on admission to our mental health wards. It would

be useful to see a more cost effective strategy also advocated for non acute

trusts.

Thank you. This is however beyond our remit.

Implementation of modified admission MRSA screening guidance for NHS (2014)

We wish to respond as follows to 'Implementation of modified admission

MRSA screening guidance for NHS (2014)', the consultation document

prepared by the Department of Health’s 'MRSA Screening Implementation

Group', whose membership we note is not provided.

A) One of the arguments cited in favour of targeted rather than universal

screening is that compliance with universal screening is poor. However this is

not in itself a reason to prefer a different strategy, unless there are reasons to

believe that compliance with a different strategy would change in such a way

that outcomes would improve. We can't think of any plausible reason why this

should be so.

B) A second argument cited is the low prevalence of MRSA. Unlike the poor

compliance argument, this is relevant to choice of screening strategy, but

only insofar as it affects the cost-benefit analysis - it is not an argument in its

own right. It could also be turned on its head: it is possible that the current

low prevalence is a consequence of universal screening and that, without

universal screening, MRSA’s ‘R-nought’ value would exceed one and the

prevalence would begin to rise again. In other words we are at risk of

abandoning a strategy simply because it has been successful. Under the

‘targeted screening’ proposal, is there a prevalence at which we would go

back to universal screening?

C) The low prevalence of MRSA is also pertinent to the comparative

performance of different screening methodologies (because

prevalence affects the predictive value of the test) so it is a shame that the

proposal says nothing about this. This is a missed opportunity to define best

practice on questions that would benefit from a standardised NHS approach,

such as pooled processing vs separate processing, nose only vs nose and

throat vs nose/throat/perineum, selective media vs selective differential

media, solid media vs broth enrichment, place of PCR etc. What were the

The omission of names was an oversight that has now been

corrected.

A) Universal screening has not been optimally implemented.

Poor implementation is however not the most pressing reason

to consider alternatives. We have altered the wording in the

guidance document to make this point clearer and reflect the

finding in the report that, even if compliance were 100%,

routine screening would still be the least cost-effective policy

in all trust types and scenarios. The NOW study concluded

that a move to targeted screening of high risk plus checklist

activated screening of low risk, which in certain circumstances

in specialty hospitals is marginally more cost effective than

high risk screening, would be even harder to implement.

B) The guidance clearly states the need to review screening

policy based on prospective surveillance. A threshold has not

been set. Even if MRSA prevalence quadruples universal

screening remains cost ineffective.

C) This is outside of the remit of the NOW study; its aim was

to evaluate universal vs targeted approaches. i.e. who should

be screened rather than how. However, the effect of

prevalence on test predictive value was accounted for in the

assumptions in the NOW study.

The NOW study modelled what actually happens in the NHS,

which is that PCR based screening is rarely used routinely.

The use of standard methods is cited in the implementation

Annex: Consultation

terms of reference of the MRSA Screening Implementation Group if not to

debate such questions?

D) The cost effectiveness modelling on which the group’s proposals are

based is very difficult to assess. The members have drawn heavily from the

NOW study, but the NOW study has not been published in a peer-reviewed

journal and seems to exist only as a 320-page document on the UCL and

IDRN websites. It describes limited aspects of its modelling but directs the

reader for further information to a reference (Robotham JV, Graves N,

Barnett AG, et al (2011). Model-based evaluation and cost-effectiveness

analysis of MRSA intervention policies) that again is not peer-reviewed and in

fact doesn’t seem to be available any longer. This is a serious criticism of a

study that seeks to inform public policy, and of the current consultation

process.

E) It is important that the membership of the MRSA Screening

Implementation Group is disclosed. Have they been asked to appraise

critically the NOW study or are they the same people who wrote it?

F) Acknowledging that we are not given full details about the NOW study’s

modelling, the information that is provided leads us to regard the cost-

effectiveness calculations with scepticism. In particular it is proposed that,

compared to universal screening, high-risk area screening would save the

NHS as a whole £250m per year and the average NHS trust £1.6m per year,

at the cost of a small number of cases of infection. These figures seem

designed to catch the chief executive's eye (or to justify withdrawal of funding

from NHS trusts) but are implausible when set against the model's

parameters of swabbing costs of £3.20, laboratory costs of £4.24 (negative

test) or £7.24 (positive test), and an estimated universal screening load for

the average 553-bed NHS trust of 790 screens per week. The answer is

almost certainly in figure A3 (see below), which confirms that it is the costs of

isolation, not the costs of screening, that make by far the largest impact on

guidance.

D) The reference that is mentioned is a DH report (from the 3

year DoH funded modelling study MECAMIP), which was peer

reviewed; although it is no longer available online, it is

available on request. This resulted in a BMJ publication

(Robotham JV, et al. BMJ 2011;343:d5694. doi:

10.1136/bmj.d5694) and an accompanying editorial. There

have been three associated publications further detailing the

methodologies (Worby CJ, et al. Am J Epidemiol

2013;177:1306-13. Barnett AG et al. Am J Epidemiol

2009;170:1186-94. Deeny SR, et al. J Hosp Infect

2013;85:33-44). The NOW study report itself was peer

reviewed by four referees including a health economist. The

audit data from the NOW study have been published (Fuller

C, et al. PLoS One 2013;8:e74219). The main study

publication is in final stages or preparation before submission.

The current version of the NOW study report is available at

www.idrn.org/audit. This reference has been updated in the

implementation guidance.

We believe that it would not be ideal to wait for publication

and then consider how to implement the findings of the study,

not least considering the costing conclusions.

E) The names of the Implementation Group have been added

to the report, along with its remit.

F/G) We agree that much of the anticipated savings come

from reductions in isolation, and accept that this is a

Implementation of modified admission MRSA screening guidance for NHS (2014)

the modelled costs of each strategy. These isolation costs themselves are

plausible (£333 per day for a standard bed and an additional £88.43 per day

for isolation nursing) but it is wholly disingenuous to use these to inform

decisions about screening. The argument becomes essentially that we

should screen fewer patients in order to isolate fewer patients. Setting aside

the point acknowledged by the NOW study itself that many carriers identified

are not isolated anyway (so these savings are not there to be had), as an

attempt to inform national screening policy this is either cynical or muddle-

headed.

G) Using the modelling parameters in the NOW study, for the average acute

trust we calculate that the saving in screening costs of high-risk area

screening compared to universal screening would be about £260k pa. If we

accept that only 61% of universal screens are actually taken then the saving

would be at most £158k (if only 61% of targeted screens were taken) and

possibly as little as £63k (if all targeted screens were taken). These

calculations assume that the costs of laboratory processing are independent

of screening volume – in reality the fixed nature of laboratory costs will mean

that the savings are less – but as a ball-park figure we might say that

targeted rather than universal screening has the potential to save the

average trust about £100k per annum. This, not £1.6 million, and is the figure

to weigh carefully against the costs and risks.

contentious issue. However, there is insufficient capacity to

isolate patients for infection reasons. MDR GNB threat needs

to be managed. The current recommendations help rather

than hinder the MDR GNB challenge.

The costs calculated from the model are not simply cost

savings resulting from screens ‘prevented’. They encapsulate

all bed day costs, infection-related costs, and intervention

related costs associated with onward transmission within the

whole hospital population (and therefore differences in lengths

of patient stay, movement around the hospital and risk of

death).

It is correct that the differences in cost of isolation outweigh

differences in cost of screens. However, given that the

effectiveness of any screening strategy is entirely dependent

on the control measures that accompany the screening,

screening and control were necessarily modelled, and

evaluated, together. Screening in itself does not create any

effect ( i.e. health benefits), and so to evaluate the cost-

effectiveness of screening in and of itself makes little sense in

this case. The only difference in effect brought about by

different screening options will be the difference in level of

control imposed. It is therefore the effect that these differing

levels of control result in, that are estimated. Both costs and

effects therefore need to be evaluated in terms of both

screening and accompanying control. This is also the

practice used elsewhere - see for example the Scottish

Pathfinder MRSA evaluation of MRSA screening.

The more important point though, is that it is the difference in

bed day costs, far exceed any intervention related costs (e.g.

either screening or isolation costs), and it is differences in

mortality and length of stay between infected and uninfected

patients, that dominate the cost results.

Annex: Consultation

H) As well as over-estimating the benefits of the proposed change, the

proposal underestimates the costs and risks. In particular, knowledge that a

patient's recent MRSA screen was negative is useful information when

choosing empiric antimicrobials: absence of a negative screening result might

lower the threshold for expensive (and potentially toxic) treatments such as

linezolid and daptomycin in patients with suspected staphylococcal infections,

while absence of a positive screening result might delay the initiation of

appropriate anti-MRSA treatment and cause treatment failure. Another

advantage of universal screening is that it facilitates detection of in-hospital

MRSA acquisitions, events that can be a bell-wether for lapses in infection

control practice generally. As far as we can tell, none of these was included

in the modelling.

I) Finally, the proposed small increase in transmission events is impossible to

affirm without seeing the calculation, but even if we take the figures at face

value these will translate into real infections and real deaths in NHS patients.

If that harm is really felt to be outweighed by the benefits then the Dept of

Health must disavow its previous zero tolerance of MRSA infections.

H) The perspective of the study is that of a regional or

national level healthcare decision maker, i.e. with the whole

health economy to consider. We agree that there may be

other benefits of universal screening, not included in the

model; for example, antibiotic choice. However, without data

on these additional benefits, for example the impact of

‘knowledge of MRSA status’ on antibiotic decision making,

and the subsequent effects that that decision making has

(primarily on patient’s length of stay and mortality) – any effect

of screening on antibiotic choice could only be included as a

scenario analysis. Given the given the current low prevalence

of MRSA, the number of scenarios in which infections that

would otherwise have been ‘detected’ (by ‘universal’

screening) is likely to be very small. The model does

acknowledge that a very small number of MRSA infections will

be missed.

With reference to the reviewer’s point regarding ‘missing’ any

in hospital transmission – this is true. We acknowledge that

the cost/QALY approach used here will not identify the ‘best’

policy if the desired outcome is monitoring (and subsequent

reduction) of colonisation within the hospital. The evaluation

was conducting to identify the ‘best’ policy option in terms of

cost per quality adjusted life year gained (which, in this case,

largely related to those policies best able to reduce infections

as opposed to colonisations).

I) There is no contradiction between the Government’s stated

policy of zero tolerance and an effective, targeted, evidence-

based screening policy for MRSA bloodstream infections.

The transmission dynamic modelling based approach was

used specifically to capture the infections and deaths

Implementation of modified admission MRSA screening guidance for NHS (2014)

In summary, we accept the suggestion that compared to universal screening,

some kind of targeted screening strategy might detect 80% of colonised

patients while incurring only half as many screens. However we do not

accept that a robust cost-benefit assessment has yet been made –

specifically we do not accept that the costs of isolation nursing are germane

to a decision on screening strategy. Removing those costs, we calculate that

this change would save the average NHS trust something like £100,000 per

annum at the expense of increases in MRSA infections, empiric antibiotic

costs, the consequences of delayed treatment of MRSA infections and other

non-trivial adverse outcomes: this is the cost-effectiveness calculation that

needs to be undertaken before an evidence-based policy change can be

made.

Finally, we wish to express surprise at the way in which the results of an

unpublished and methodologically opaque study are being misrepresented to

influence this important debate.

Fig A3 (NOW study p72) showing that the cost-effectiveness modelling is

overwhelmingly about the costs of isolation, not the costs of screening.

associated with each of the strategy options (and especially to

the knock on infections and deaths brought about due to

transmission). Total quality adjusted life years gained under

each strategy are compared. The cost/QALY approach is

therefore used in order for these infections and deaths to be

included within the decision making process. The purpose of

presenting health benefits as QALYs is to allow decisions

across the health sector to be compared (using the same

units) and so to enable rational decision making.

The differences in the numbers of infections and colonisations

are the output of the transmission dynamic model, simulating

transmission throughout the hospital for each scenario over a

5 year period. i.e. 13,000 lines of computer code – run 1000’s

of times. It is therefore difficult to provide a ‘calculation’. This

is the case with any such model-based cost-effectiveness

evaluation. It is the model structure and the inputs that are

driving the simulation.

The computer code itself is available on request. However,

reassurance should come from the fact that the code has

been developed by a team of mathematical modellers, subject

to a multitude of checks, and resulted in numerous peer

reviewed publications.

Methodology (A)

We note that the recommendations are based on the findings of the NOW

study. It would appear that the NOW report has not been published in any

peer-reviewed format, although it is available on certain websites. Page

numbers below refer to the copy of the publication on the idrn.org web site.

Furthermore, the consultation paper refers to a “report on file” in the DH

(reference 8).

A) Please see above answers (4) regarding NOW study

availability, publication and peer review.

Annex: Consultation

Assumptions

(B) The rationale for some of the assumptions are puzzling. For example,

“Direct infection from a susceptible state cannot occur in low risk specialty

settings and patients must first become colonised” (pg 193)

We have reviewed the assumed parameters (pg 210). Unfortunately, due to

the time scales available, detailed review of the sources of the information,

and the extent that the data could be extrapolated from one setting to another

was not possible.

to be the bed day costs for the admission of individual patients, set at an

average of £333 (pgs 210, 234) and incremental costs of placing a patient in

an isolation room, set at £88.43/day. We note that 100% bed occupancy is

assumed in the model (pg 193). Therefore, there would not be any

incremental costs directly resulting from patient admission or placement in a

single room. Indeed, if the report’s conclusion that there will be an increase,

however “minimal” (pg. 12), in infections consequent to the proposed

changes in screening practice is accepted, a increase in total bed days would

be expected, bed capacity permitting. It is therefore unclear why bed days

costs would be lower in strategy 1 (pg 234).

(D) We also note that contact precaution costs per se were set at £19.53/day.

In our experience, wards do not normally increase their staffing complement

in reaction to small changes in the number of patients being nursed in single

rooms. Furthermore, we presume that this cost cannot relate exclusively to

consumables; if this includes costs such as enhanced cleaning, in a context

where highest possible standards of cleaning are expected on an ongoing

basis (and indeed demanded by regulatory authorities), it is difficult to see

how such additional costs would be incurred. Indeed, if the consequence of a

reduction in screening were to be considered to its natural conclusion, this

B) The value of this parameter for a high risk setting was

extremely low, and therefore it can reasonably be assumed to

be lower for low risk specialities. As this is such a rare event,

it was computationally sensible to assume this to be zero.

This assumption enabled transmission estimates for this

setting to be estimated using hospital data. It is worth noting

that patients could acquire colonisation on day x, and

subsequently become infected on day x+1, and the probability

of each of these transitions was estimated using individual

level hospital data. Furthermore, given that susceptible and

colonised patients were assumed equally infectious, there

would be no difference in their transmission potential.

C) Increases in numbers of infections resulted in (on average)

longer patient stays. While we absolutely agree with the

reviewer that under the assumption of 100% bed occupancy,

this does not, in itself, result in increased bed day costs.

However, it is the opportunity cost of bed days lost to treat

other patients that we describe. Results presented are

divided by the number of admissions. Therefore, longer stays

will prevent admissions, decreasing this denominator, and

increase bed day costs per admission.

D) Contact precaution costs were taken from the Scottish

MRSA Screening pathfinder study, in which they updated the

estimates from the previous HTA Report (Ritchie K, et al.

Consultation report on health technology: Clinical and cost

effectiveness of screening for MRSA. NHS Quality

Improvement Scotland, 2006. Available at:

http://www.nhshealthquality.org/nhsqis/3780.html

In brief, this included: 12 patient contacts per day by

Implementation of modified admission MRSA screening guidance for NHS (2014)

will lead to a modest increase in infections, and logically a greater inevitable

increase in unrecognised colonised individuals - cleaning standards would

therefore need to be increased across the board, in order to compensate

from the inevitably increased risk of transmission of MRSA.

(E) The report states that the average charge made by laboratories for MRSA

screening was £5.68, (sd = £4.44), pg. 19. In the model, different costs were

used, £7.24 for a positive, £4.24 for a negative test (pg.210). It is widely

known that the cost of reagents and associated laboratory staff costs, ranging

from specimen reception, through processing and result issue would scarcely

lead to cash releasing savings of more than around £1 per screen set, as all

fixed costs (ranging from estate and major equipment to managerial and

medical costs) would still have to be met, but in the NHS economic model,

these fixed costs would be included in quoted costs (only 2% of hospitals use

private providers, where a reduction in total specimen volume might lead to a

cash-releasing savings in excess of reagent and direct staff costs).

Incidentally, the author recently attended a UEMS Microbiology section

meeting where it was reported that private laboratories in Germany charge

between Euro 1.50 to 3.50 per patient screen (these quoted costs would

include a contribution to fixed costs), thereby providing some validation of the

suggested true marginal costs above, assuming that UK laboratories are

organised in an optimal efficient manner.

Similarly, swabbing cost of £4.20 was used (pg. 210). Other than the true

marginal costs of the swabs themselves, a reduction in nursing time spent

swabbing a patient would not lead to any reduction in the staffing

complement on the ward and would therefore not release any savings.

(F) Modelling

No sensitivity analysis to changes in the cost parameters was presented. It

would be instructive to see the effect of re-running the model with marginal

costs to take into account reasonably expected cash releasing.

healthcare staff (3 minutes per contact needed to ensure

compliance). Plus consumables for each contact of one pair

of gloves and one plastic apron. Providing a total daily cost

including overheads. (see pg 64 of HTA report for more

information).

E) Again, screen costs were taken form the best available

evidence, which was again from the Scottish Pathfinder study.

It is worth reemphasising that all of these costs will be

dwarfed by differences in cost brought about by differences in

length of stay and mortality (due to differences in numbers of

infections between strategies). Note the differences in scale

between each of the cost components on any of the cost

plots.

Having said this, large reductions in testing could release

estate/space/staffing, or make these available for other

purposes e.g. CPE screening, nurse directed care, etc.

F) Indeed, it would be ideal to include uncertainty in these

cost estimates, in the full probabilistic sensitivity analysis.

However, it is unlikely that the decision between strategies

would change under different cost assumptions (other than

the cost of bed days which dominate the evaluation).

Annex: Consultation

The model also appears to be static in time and does not take into account

the reasonable expectation of an inevitable increase in colonisation rates that

would occur over time in those strategies that entail lower screening and

therefore detection rates. Over a period of several years, it can be expected

intuitively that as colonisation rates increase, so will transmission, with an

inevitable acceleration in infections.

(G) We could find no evidence that the model took into account changes in

compliance with screening in the different models. One of the implicit

criticisms of universal screening in the consultation paper is the observed

compliance rate of 61%. Intuitively, compliance with various forms of check-

list activated compliance is unlikely to exceed this figure. Experience of

practitioners practising in the field of IPC would argue that compliance is very

much aided when embedded in routine practice. Therefore a change away

from universal screening can be reasonably expected to lead to a greater

proportion of unisolated MRSA patients, with a consequently greater risk of

transmission to others.

(H) Screening practice

The consultation paper states that screening compliance at the time of the

NOW study data collection period was 61%.

However, the data collection period was in May 2011 (pg. 8), only a few

months after the introduction of universal mandatory emergency admission

screening. It is reasonable to assume that in the early months, Trusts were

still developing their systems and that compliance has improved considerably

since. It would therefore be inappropriate to use data that applied in 2011 to

inform current practice, particularly as there is considerable anecdotal

evidence from IPC practitioners in London who are members of this Forum,

that compliance with established screening policies is very high in most

organisations.

Admission colonization rates fell in the years before

implementation of the current screening guidance.

Also, guidance states that local decisions about who/where to

screen may be required according to local epidemiology.

G) This is true. However, we took the decision early on in the

NOW study to compare each of the strategies against one

another ‘on a level playing field’ and to see how each fared

under the assumption of 100% compliance.

Actually, recently published evidence from Scotland

(http://www.bmj.com/content/348/bmj.g1697/rapid-responses

)

provides further evidence that risk-based approaches to

MRSA screening can be successful in clinical practice and

maximise cost-effectiveness.

H) The modelling assumed 100% compliance with each

policy.

Implementation of modified admission MRSA screening guidance for NHS (2014)

(I) The report states that at the time of the survey, the turnaround time for a

positive result was nearly 3 days, and nearly 2 days for a negative result (pg.

19). In our opinion, there is considerable scope for improvement, which may

well have been achieved, at least in some laboratories, particularly those that

have embraced continuous quality improvement initiatives. For example, in

the author’s laboratory (until 2013) negative results and presumptive positive

results were routinely issued within 24 hours, 7 days per week.

For patients whose results become available after discharge, in our

experience, systems have been put into place in many hospitals to notify

primary care of the result, enabling GPs to prescribe suppressive /

decolonisation therapy should this be considered beneficial. In any case

given the high frequency of re-admission, this provides additional useful

information to inform patient placement immediately on presentation should

re-admission be required.

(J) Patient placement and decolonisation / suppression management

The report states that the one-day audit demonstrated that only 55% of new

MRSA patients were isolated and that decolonsation / suppression therapy

had been started in only 73% (pg. 19). The audit did not appear to collect

data on time from MRSA detection to implementation of isolation and

commencement of decolonisation / suppression treatment. It is therefore

unclear whether these deficiencies were merely temporal and the proportion

of unisolated / untreated patients at, say 6 and 24 hours does not appear to

be know in this study cohort.

Sensitivity analysis of costs at various colonisation rates

The authors of the NOW report have run the model at various prevalence of

colonisation. There are limited data on prevalence of MRSA among

emergency admissions. However, a “very high” rate of 5% may well be less

than the rate that may be encountered should national guidelines

I) This evaluation was an extension to previous work, and at

the time of being conducted, the turnaround times

represented synthesised estimates from a review of the

available evidence in the literature; these accurately reflected

the turnaround times for MRSA positive patients reported in

the NOW audit. We acknowledge that these may indeed be

reducing all the time. This is an inherent problem in the

evaluation of screening, with ever faster (and cheaper) tests

being developed.

(J) National audit data from individual patients were used

here. The study recorded if all positive patients that week, or

randomly chosen negative patients, were isolated before or

after screening result became available. Isolation usually

occurred as soon as a positive result was known, but many

patients were home before results were available.

The model reviewed multiple scenarios, including varying

MRSA rates. Guidance clearly cites the need for prospective

surveillance and review of policy.

Note that previous modelling work (DH MECAMIP study)

found that while screening and control is likely to be cost

effective at a prevalence of 10% (including universal

screening) in ICU settings, this was not the case in general

medical wards. The NOW study included data on all

emergency admissions in England during one week. We are

Annex: Consultation

recommend selective screening once again. For example, G Rao et al (JHI.

2007; 66: 15) had observed MRSA colonisation rates of 8.6% among

admissions (6.7% of individual patients). The author had also observed

colonisation rates of 6-8% among emergency admissions in 2008 prior to

introduction of universal MRSA screening at his hospital (unpublished data).

(K) Additional, practical, considerations

Most experienced practitioners do not consider that it is practical or feasible

to identify high risk patients in an A&E environment. There is published

evidence to support this (for example G Rao et al. JHI. 2007; 66: 15).

(L) At high levels of bed occupancy, patients are frequently transferred

between wards, including those that are classified as high and low risk.

Allowance for this effect does not appear to have been made in the model.

Practical considerations around this would be considerable.

Experienced practitioners recognise that IPC operates in a human

environment. Consequently, application of precautions intended to minimize

transmission (e.g. effective hand hygiene) are more likely to be applied

consistently when risks are known, such as in known MRSA colonised

patients, who are being nursed with contact precautions.

Knowledge of an individual patient’s MRSA status is valuable in empirical

antimicrobial choices, thereby potentially leading to increases in antimicrobial

costs if this information is not available. This element was not included in the

costing model.

unaware of any more extensive data, including in the Scottish

pathfinder study.

(K) This is not true. Patients at high risk of infections are/can

reasonably be screened in A&E e.g. influenza, viral

gastroenteritis, SARS, etc. Infection control teams should

turn their attention from screening everyone to ensuring that

high risk patients are screened. For a fraction of the costs of

routine screening, one could invest in audit and feedback

systems to ensure that compliance with high risk screening is

high.

(L) High risk and low risk ward transfers were estimated using

individual-level hospital data, and therefore aimed to reflect

real hospital movements. These transfers were included in

the model.

All of which comes at a cost. National audit data were used in

order to make the modelling representative of clinical practice.

This is not in accordance with the application and practice of

standard precautions.

Please see above response regarding empirical prescribing.

The aim was to estimate the optimal policy; ‘optimal’ was

defined in terms of cost/QALY. We agree that if the definition

of optimal differed, e.g. instead the policy that best enabled

monitoring (or reduction) of colonisations, then the decision

would have differed.

Implementation of modified admission MRSA screening guidance for NHS (2014)

(M) Centres that implemented universal MRSA screening at a time when

‘avoidable’ cases of MRSA bacteraemia had already been all but eliminated

observed a rapid and marked reduction in true hospital acquired infections or

colonisations. For example G Rao et al. JHI. 2007; 66: 15, Sarma et al. ARIC

2013; 2: 2, A. Mifsud unpublished observations.

(N) Once systems have been established for screening, these can be

relatively easily adapted to take into account changing needs, such as

screening for CROs and other emerging infections. Dismantling of universal

programmes will make subsequent re-instatement difficult.

(O) It has been suggested that universal use of antisepsis may reduce the

impact of reduced screening. However, the recent recognition of MRSA

strains carrying the antiseptic resistance genes, such that they are clinically

resistant to chlorhexidine, should strike a note of caution against its

unfettered use (Edgeworth JAC. 2011: 66 s2: ii41)

(P) DIPC Forum recommendation

We have concerns around the assumptions used in the economic model,

such that we suspect that the financial case may not be as stark as

presented. We are certain that genuine cash releasing savings may well ot

ensue, and indeed, if the history of management of MRSA in the UK in the

1980s and 1990s is repeated, a large increase in infections and therefore

costs can be anticipated with a high degree of probability.

M) The Rao et al study was observational over a one year

period and noted ‘The study was not designed to establish

whether this reduction was causally associated with the

screening programme.’ In the Sarma et al single centre

observational study, multiple interventions occurred

(‘Following the introduction of Root Cause Analysis in May

2006 a number of interventions were made in quick

succession as part of the MRSA improvement programme’).

Whilst there was a significant association with the introduction

of screening, it is not possible to conclude that other

measures drove or part drove the observed decline in MRSA

infection.

N) We are not advocating dismantling systems for screening.

Targeted MRSA screening reasonably can release resources

to implement other screening e.g. for CROs.

O) We have not advocated ‘universal use of antisepsis.’

P) There is no contradiction between the Government’s stated

policy of zero tolerance and an effective, targeted, evidence-

based screening policy for MRSA bloodstream infections.

Furthermore, nothing in this revised guidance negates the

need for effective surveillance, screening and reporting

generally. In fact, the guidance makes it clear that ‘Trusts will

need to regularly review (and where necessary improve) their

compliance with national screening guidance for each

Annex: Consultation

In principle we are not supportive of mandatory measures. However, in this

particular case, we are convinced that withdrawal of universal MRSA

screening will inevitably lead to an increase in MRSA colonisation rates

which, over time, will lead to an increase in colonisation rates in the

community and will impact on the spread of MRSA in other hospitals in the

vicinity. Relaxed control of MRSA in one hospital can be expected to give rise

to a geometric build up of cases within the community and will impact on

cases in adjacent hospitals.

Furthermore, the removal of mandated MRSA screening is incongruous in

the context of the DH’s stated objective of zero tolerance of avoidable

infections. The current performance management regime around MRSA

bacteraemia management includes stringent criticism and sanctions. A

system that allows sub-ideal performance in some respects but not in others

appears perverse.

We would suggest that efforts are focussed on fine-tuning the current MRSA

management pathways, for example by reviewing screening in situations

where transmission is unlikely, e.g. in most day case surgeries,

improvements in compliance with universal emergency admission screening

(if poor performance still occurs), improving laboratory testing and reporting

processes, such that cash releasing savings could be accrued by withdrawal

of molecular testing , and improving the primary / secondary care interface to

ensure good flow of information and action across the sectors.

specialty, recent MRSA infection data, patient demographics

and types of services provided within individual organisations.’

The guidance also clearly states need for continued

local/national surveillance. If MRSA rates increase then can

revisit case for screening. The guidance also states that local

risk assessment can be used to make decisions about

who/where to screen.

Cost effectiveness is not perverse. It is one of the

requirements of good management of public resources. The

key reason to conduct evaluations such as the NOW study is

to provide a rational basis for decision making and allocation

of scarce resources. We do not have any other evidence

from RCTs etc for whole hospitals based on representative

national data. The model and report are from the perspective

of the regional health policy maker who is considering the

picture for the health economy.

The suggested alternatives remain conjecture.

Introduction section

* Not sure if this part will be included in final guidance??

Introduction section, 3rd paragraph

* Re: The NOW study showing that 'compliance with current mandatory

Yes it will but modified as above.

Implementation of modified admission MRSA screening guidance for NHS (2014)

screening policy was poor (e.g. only 61% were screened;...'

o More informative to give the breakdown by admission type as given by

NOW study: the NOW study showed this related to 61% of patients in the

emergency admissions category and 41% in the eligible day case admissions

category but it was highest in the elective admissions category (81% of

patients).

* The next point: ..'about half of new positives were isolated when their

result became known'

o Better to make clear that overall, 55% of new positives were isolated

once results were confirmed

* The next point: 'and about a quarter did not receive decolonisation

therapy..'

o First this information is not included in the Appendix of the consultation

o Second, the result quoted is confusing. The NOW report showed that

34% of Trusts did not use pre-emptive suppression/decolonisation on any

admitted patient (see Table 9 in NOW report). Or was consultation referring

to 27% of patients that did not receive decolonisation once their MRSA result

was known.

Introduction section, 5th paragraph

The reader can refer to the NOW study for more detail if

required - this is an overview.

OK.

Do not understand the point.

The latter.

Annex: Consultation

* Re: 'Approximately 60% of MRSA colonised individuals will be

detected by screening those known to be previously MRSA positive and high

risk specialty admissions'.

o I did not see this result anywhere in the NOW report. This statement

implies that three pieces of information needed, a denominator for all risk

factors, the numerator for prior MRSA and numerator for high risk specialty.

In fact if 60% refers to these two groups they must be mutually exclusive and

another numerator should be included which comprises patients with both

these risk factors present (even if 0). Is his based on Table 29a? Whatever

the 60% refers to, this needs to be clarified.

Summary

* Re: 'Importantly, focussed screening should be adopted in line with

local risk assessments..'

o If focussed screening to be main national policy then better to say:

'Importantly, focussed screening should be adopted, where necessary, as

indicated by local risk assessments..'

* Re: 'Trusts will need to consider current compliance with mandatory

screening guidance for each specialty...'

o This sentence refers to the 'current' situation regarding compliance to

existing guideline. Once the existing guideline (screening all admissions)

becomes obsolete everything else will by association be irrelevant. Better to

say: 'Trusts need to regularly review and where necessary, improve their

compliance with national screening guidance for each specialty...'

This is calculated from data in the report. Simplistically, 50%

are previous MRSA +ves plus high risk screening identifies a

further 10%.

Suggested rewording is not what we mean.

OK.

Implementation of modified admission MRSA screening guidance for NHS (2014)

After Objective, add a short Background section - this will include definition of

MRSA/ habitat/transmission dynamics/impact on length of stay,

morbidity/mortality

The screening groups are confusingly written. Due to layout there is a

possibility of misinterpretation. Solution advised as follows.

After Background section, important to add a section called: Screening

categories.

* For clarity re-organise information under Screening categories:

o Keep: 'Elective or acute admissions to high risk specialties (defined

below)'.

o Remove the sub-bullet point: 'all patients admitted to critical care units' -

these are already included in the high risk specialties list.

o Remove the sub-bullet point: 'all patients previously identified as

colonised with or infected by MRSA' - these would be included in the local

risk assessment component. (The NOW study in fact identified high risk

specialty screening (reverting to previous strategy) as the most cost-effective

strategy. The NOW report did not specifically have findings for 'prior MRSA'

alone (in any case it was embedded in the strategy that included all

admissions i.e. strategy #6).

o Add: 'Additional specified patient groups identified through local risk

No, too much detail. This information is widely accessible.

Formatting will be checked in published guidance.

We are not just charged with implementing the NOW findings

but are to use our judgment in using that evidence to revise

guidance.

Annex: Consultation

assessment'

So the 'Screening categories' plus sub bullets would finally look like as

follows:

Screening categories

* Elective or acute admissions to high risk specialties (defined below):

o Sub-bullet point: High risk specialties are defined as vascular,

renal/dialysis, neurosurgery, cardiothoracic surgery,.........Coronary Care

Units

* Additional specified patient groups identified through local risk

assessment

o Sub-bullet point: Local risk assessment should be used to define other

potential high MRSA risk units/specialties; for example, according to

provision of localised specialised services.....endemicity of MRSA.

o Add after this: 'Include all patients previously known to be MRSA

positive'.

o The continue with: 'Local risk assessment may increase the proportion of

detected colonised individuals, notably.....poor outcome from MRSA. (bold as

Implementation of modified admission MRSA screening guidance for NHS (2014)

found in consultation)

Interventions, No.8, add word 'effective'

* All patients identified as MRSA positive must receive effective

decolonisation/suppression therapy

Compliance and audit

* Suggested edits as follows:

o Remove i. and keep the statement: 'Trusts should make every

effort....groups identified above' as opening descriptive statement.

o The next point ii should now be i

• vi (which is now v) should begin with: Patient feedback of results is

important. A qualitative study on the patient experience.......experience

is necessary'

•

Unnecessary – we would not advocate ineffective therapy.

Thank you for the suggestions, but we have elected to keep

the original version as others have not commented on a need

for change here.

(A) Concerns that the modelling used in the NOW study is flawed

because of the underlying assumptions.

Other modelling studies have suggested that a combined approach of

isolation and screening confers efficacy (e.g. Bootsma‐M et al. 2006 PNAS;

103: 5620–5625). Coia

‐

J (BMJ 2014;348:g1697) wrote: “A prospective case

‐

control study of more than 12000 patients showed that screening strategies

(A) Indeed, the transmission dynamic model was informed

assuming that isolating patients is an effective control

measure. Precisely, the model input was a 64% (S.D. 14%)

reduction in transmissibility due to isolation. This is reflected

in the model results – where reductions in transmissions,

infections and deaths are seen for all strategies involving

screening and isolation. Furthermore, the strategies were

Annex: Consultation

of sufficient sensitivity require screening of 65% of admissions (Harbarth S at

al. Evaluating the probability of previously unknown carriage of MRSA at

hospital admission. Am J Med 2006;119:275.e15‐23)

(B) Importance of Medical Devices in patients

Members experience suggests that serious MRSA infections and

bacteraemias have been seen in subgroups of patients with long term

medical devices in situ especially those devices that patients take with them

into community settings e.g. nephrostomies, central intravascular devices,

long term urinary catheters.

This is seen especially in patients with underlying malignancy. This needs to

be factored in when writing screening guidance for practical use, maybe at

the specific trust level.

In hospitals with >95% adult bed occupancy rate, many patients are

transferred several times during an admission, including transfers between

“low risk” wards and “high risk” wards. Thus, even if identifying populations at

high risk for MRSA carriage may be feasible, universal screening may be

easier to implement, as risk categorisation may have to change throughout

admissions due to the mix of patients in different wards and clinical areas.

(D) Realisation of projected cost savings

Cost savings make some assumptions that may not be practically realisable.

In addition there are additional costs to not having MRSA status from recent

screening. Savings are based on not screening (cost of implementing

screening) and not therefore isolating patients. However many patients are

not isolated in single rooms in low risk areas. Knowledge of MRSA status is

compared in terms of these effects combined with costs.

(B) There is no clinical and cost-effectiveness evidence to

support the screening of patients with medical devices and

this was not considered by the NOW Study. However, the

guidance allows local risk assessment to define other high

MRSA risk groups, which may include these groups of

patients.

be straightforward, appropriate risk assessment should be

able to categorise the majority of patient groups correctly.

Issues with implementation are not a justifiable argument to

sustain universal screening in the absence of cost-

effectiveness data.

(D) Please see answers above.

Briefly, savings are not due to not isolating patients, but

instead due (primarily) to reductions in infections (and

therefore length of stay). The cost figures demonstrate that

any differences in isolation costs are dwarfed by the

differences in bed day costs (brought about by reduction in

length of stay – in turn brought about by reduction in

infections).

The potential costs of not knowing MRSA status are difficult to

quantify, but are not likely to be significant in low risk patient

groups. Issues with implementation are not a justifiable

argument to sustain universal screening in the absence of

cost-effectiveness data.

Implementation of modified admission MRSA screening guidance for NHS (2014)

believed to prompt enhanced precautions in staff – encourages compliance

i.e. modifying behaviour to reduce transmission risk. There is a cost in

implementing selective screening in the training, auditing, time necessary to

achieve a more complex system. It is likely to be hard to successfully

implement and maintain compliance. Knowledge of MRSA screening results

gives confidence in not using agents such as linezolid or daptomycin in the

initial treatment of serious infections with a likely staphylococcal cause.

(E) Long term implications

Risk‐based screening for MRSA carriage is likely to miss too many carriers of

MRSA to achieve meaningful control of MRSA infections including prevention

of MRSA bloodstream infections in the long term (5‐10 years).

(F) Stratified Implementing of revised screening policies

NHS hospitals that suffer from >95% adult bed occupancy rates and continue

to have seen hospital‐acquired MRSA bloodstream infection in the last 24

months, should continue with universal admission screening. NHS hospitals

that have not seen any hospital‐acquired MRSA bloodstream infection in the

last 24 months could trial alternative practical strategies of screening for

MRSA that are less costly.

(E) There is no evidence to support that appropriate risk-

based screening will fail to achieve meaningful control of

MRSA. The guidance does advocate prospective surveillance

to determine if revised risk assessment and policy change is

required.

(F) There is no evidence to support this statement. The

guidance allows local risk assessment to identify appropriate

high risk groups of patients who should be screened The

guidance does advocate prospective surveillance to

determine if revised risk assessment and policy change is

required.

Compiled by Prof Mark Wilcox on behalf of MRSA Screening Implementation Group, June 2014