Tackling the spread of C Difficile in a priority for the NHS. A team explain how they adapted and tested the Waterlow score to assess patients at risk of infection
Judith Tanner, PhD, MPhil, BN, is professor of clinical nursing research, De Montfort University and University Hospitals of Leicester Trust; Debra Khan, MSc, RGN, is senior research fellow, De Montfort University; Stephen Timmons, PhD, MA, BA, is associate professor, University of Nottingham.
Tanner J et al (2010) Using the Waterlow risk assessment tool to predict C difficile infection risk in hospital settings. Nursing Times; 106: 36, early online publication.
Clostridium difficile (C difficile) is the most important cause of infectious healthcare acquired diarrhoea and risk of infection is a public concern. Many acute and primary care trusts are working to further reduce rates of C difficile and targets have been set. Global reduction strategies include prudent antibiotic prescribing and intensive cleaning programmes.
While risk assessment tools for C difficile exist, they are not easy to use in hospital. This article introduces an inexpensive and easy to use assessment tool and outlines how nurses can apply a reduction strategy using it to identify patients who are at risk on admission and then target interventions to reduce this risk.
Keywords:Clostridium difficile, Risk assessment tool, Waterlow score, Infection control
- This article has been double-blind peer reviewed
Recommendations for practice
- Review the use of the Waterlow risk assessment tool and consider whether it could be used for Clostridium difficile risk assessment.
- Audit Waterlow scoring for completeness and accuracy.
- Discuss the implementation of the interventions aimed at high risk patients with the multidisciplinary team.
What is Clostridium difficile?
Around 2% of healthy adults are colonised with Clostridium difficile (C difficile), though rates are higher among nursing home residents and hospitalised patients (Simor et al, 2002).
C difficile is a bacterium found in soil and the human gastrointestinal tract. In the latter it mainly causes asymptomatic colonisation as the bacteria are kept under control by normal healthy intestinal flora.
When these flora are altered, the C difficile bacteria flourish and produce two types of toxins, called A and B. Toxin A causes mucosal damage and intestinal inflammation. Toxin B is 1,000 times more potent than toxin A and causes significant diarrhoea by stimulating gastrointestinal contractions.
Ultimately, the toxins A and B can cause major tissue damage with colon ulceration leading to colitis. Additional symptoms include dehydration, hypotension and bacteraemia (Kuijper et al, 2006).
C difficile associated disease (CDAD) refers to the spectrum of diseases caused by a C difficile infection, ranging from mild diarrhoea to a severe life threatening condition. In 2008, C difficile was mentioned on nearly 6,000 death certificates, and was directly responsible for 49% of these deaths (Office for National Statistics, 2008).
Risk factors for developing C difficile include clinical interventions which alter normal gastrointestinal flora, exposure to toxigenic strains of C difficile and other factors such as increasing age, reduced immunity and co-morbidities which predispose patients to infection. Clinical interventions that alter normal intestinal flora include repeated enemas, prolonged nasogastric feeding, gastrointestinal tract surgery and overuse of antibiotics, especially amoxicillin, clindamycin and cephalosporins.
C difficile bacteria are spread easily as diarrhoea from an infected patient can release spores into the environment which can survive for 70 days (O’Neill et al, 1993). The spores can be transported through contact with contaminated surfaces such as floors, bedpans, toilets, and also the hands of healthcare staff and patients.
C difficile infections in the UK
The C difficile organism was identified in the 1930s and by the 1980s was a widely recognised cause of diarrhoea in hospitals. Infection rates increased in the early 1990s and outbreaks became more common. An outbreak in Manchester in 1991 affected 175 patients and was directly responsible for 17 deaths (Cartmill et al, 1994). Increased incidence occurred again in 1996 and rose steadily from 2001, peaking in 2006 with 55,635 cases (Health Protection Agency, 2008).
In the UK in the early 2000s a new strain of C difficile, which had been associated with outbreaks in Canada, North America and Europe, began to emerge (Healthcare Commission, 2006). This new 027 strain produces toxins at a faster rate, which results in a more severe disease with increasing numbers of complications and deaths. The 027 strain is also resistant to fluoroquinolone antibiotics (Drudy et al, 2007).
In 2004, surveillance of C difficile in England became mandatory for all patients over the age of 65, with hospitals having to report cases to the Health Protection Agency (HPA, 2008). Surveillance was extended in 2007 to include all cases of C difficile in patients over two years (HPA, 2008).
Tackling the spread of C difficile infections became a priority for the NHS, as set out in the 2008/09 NHS Operating Framework (Department of Health, 2007a) and the Public Service Agreement Ensure Better Care for All(Cabinet Office, 2007).A national target was set to reduce cases by 30% in the financial year beginning 2010 (DH, 2007a), and from 2007, primary care trusts were required to agree local targets with NHS hospital trusts (DH, 2007b).
Interventions to prevent C difficile infection
Five main interventions have been identified to reduce or prevent the spread of C difficile infection (Healthcare Commission, 2005) and included in HPA (2009a) guidance, listed in Box 1.
Box 1. Interventions to prevent C difficile infections
- Intensive cleaning programmes
- Prudent antibiotic prescribing
- Increased hand hygiene
- Patient isolation
- Personal protective equipment
These interventions are either implemented across an entire trust for all patients (intensive cleaning programmes, prudent antibiotic prescribing and increased hand hygiene) or for specific patients who already show symptoms of C difficile infection (patient isolation, barrier nursing and personal protective equipment). A combination of these interventions appears to have been successful in reducing C difficile as rates have dropped. The latest report from 2008/2009 shows a 35% reduction in cases compared with the previous year (HPA, 2009b).
While recommended interventions have achieved some success, they are not ideal. For example, trust wide interventions require significant resources to implement and identifying at risk patients would be more cost effective. Similarly, actions aimed at preventing the spread from infected patients are crucial but the emphasis should be on preventing the infection at the outset.
The most effective strategy is to prevent new cases by targeting at risk patients. This approach of identifying those at risk and intervening has been used successfully in healthcare, for example using assessment tools to identify those at risk of falling (Oliver et al, 2004).
The DH (2008) published an assessment tool which aims to identify patients at risk of acquiring C difficile, as part of the Clean, Safe Care campaign (www.clean-safe-care.nhs.uk). Patients are assessed using two scoring systems. The first score is based on three critical risk factors:
- Is the patient over 65?
- Do they have a severe underlying disease?
- Are they taking antibiotics or have they taken them in the last eight weeks?
The second scoring system asks more specific questions such as recent contact with healthcare settings and use of proton pump inhibitors.
There are two main problems with this system. The first is the sensitivity of this assessment. Patients who answer yes to each of the three questions in the first assessment will be labelled as at high risk of developing a C. difficle infection. However, the proportion of hospital patients who are over 65, have an underlying disease and have taken antibiotics recently is considerable. This means that large numbers will be identified as at risk and it is not possible to implement resource intensive interventions for large numbers.
The second concern is not specific to this assessment tool, but new assessment tools in general. Numerous studies have shown that nursing staff are poor at assessing and documenting risk assessment scores (O’Conner, 2007). Implementing another new risk assessment tool will merely add to the burden of paperwork on admission, resulting in inaccurate or incomplete assessments, rendering this tool worthless.
An alternative risk assessment
Recent HPA guidance on C difficile calls for “prognostic risk scores [which] will require development and validation” (HPA, 2009a).
The Waterlow risk assessment tool has been found to be easy to use, inexpensive and imposes no extra burdens on nurses. It appears to be the only tool that can be used specifically for identifying patients at risk of C difficile on admission (Tanner et al, 2009).
The Waterlow is an assessment tool used to identify patients at risk of developing pressure ulcers. It was designed in 1984 by Judy Waterlow for use by student nurses (Waterlow, 1985) and revised version, available on judy-waterlow.co.uk, includes new categories for nutrition, organ failure and time spent in surgery. The Waterlow tool is widely used by nurses in many countries around the world and risk factors for a high score are shown in Box 2.
Box 2. Factors included in the Waterlow riak assessment tool
- Body mass index
- Skin type
- Sex and age
- Malnutrition screening tool (MST)
- Special risks:
major surgery or trauma
A team at De Montfort University, Leicester, successfully adapted the Waterlow tool for use as a risk assessment for C difficile infection. To show that the Waterlow is a valid and reliable tool for assessing C difficile infection, as well as pressure ulcers, the team carried out a study (reported in more detail in Tanner et al, 2009). In order to be truly rigorous, this study tested the Waterlow tool using a three stage process. The first stage identified risk factors, the second tested the tool’s validity in one hospital and the third assessed its validity in a different hospital.
For the first stage, 25 patients with C difficile infection were matched (in pairs) with another 25 who had not developed the infection. They were matched by age, gender, admission date, presenting diagnosis and specialty. The notes from the 25 matched pairs of patients were examined for known risk factors for C difficile as well as the Waterlow score and the a frailty score (Woods et al, 2005). Statistical analysis of this data showed that a high Waterlow score (20 or above) was significantly associated with patients having a C difficile infection.
The second stage tested the validity of the Waterlow score in predicting which patients would be likely to develop C difficile infection. All those admitted to an adult medical assessment unit over three months were followed up to determine which ones went on to develop C difficile infection. Waterlow scores for all patients were documented on admission. Some 1,468 patients were included in this stage of the study and 83 had a high score (20 or above). Twenty patients from the overall total developed C difficile infection, and 14 of these patients were in the small group of 83 patients with the high Waterlow scores. By identifying 70% of the C difficile cases, this shows that the Waterlow score is effective in predicting patients at risk of developing C difficile infection.
To further validate the reliability of the Waterlow score as a risk assessment tool for C difficile infection, the tool was tested in a second hospital using a larger sample. The electronic records of 29,211 patients were analysed to identify Waterlow scores on admission and C difficile status. Some 1,347 patients had a high Waterlow score of 20 or above and 98 (from the overall total) developed C difficile infection. Thirty-two infected patients were in the small group with high Waterlow scores of 20 or above. The Waterlow tool identified 33% of those who went on to develop C difficile infections. While not appearing to be as supportive of this use of the Waterlow tool, this still confirms that it is effective in predicting patients at risk of developing C difficile infection.
Advantages of using Waterlow
The major benefits of using the Waterlow score to predict those at risk of developing C difficile infection include:
- It is already widely used;
- Nurses are familiar with it;
- They do not have to carry out extra assessments.
It is therefore inexpensive and easy to implement, alhough it needs to be used accurately.
Another advantage of this tool is that it identifies a small group of at risk patients, allowing resource intensive interventions to be targeted. These are outlined in Box 3.
Box 3. Interventions to reduce the risk of C difficile
- Fast track discharge schemes are effective in reducing C difficile infection as they keep patients in hospital for the minimum time necessary.
- Dedicated bays in medical assessment units can be established for those with risk factors so staff are alerted to their needs. They also facilitate dedicated ward rounds by gerontologists.
- The majority of medical patients who develop C difficile infection are over 85. Rapid review on admission by gerontologists will speed up the assessment, treatment and discharge of at risk patients.
- Medication reviews ensure prescription of appropriate antibiotics (Ansari et al, 2003).
- Specific care pathways for patients at risk of developing C difficile infection help to minimise time spent in hospital, maximise patients’ defences and reduce exposure to C difficile.
While some of these initiatives have been implemented, current low levels of C difficile infection in the hospital studied mean it is not possible to evaluate them rigorously.
This article has suggested an inexpensive, effective and easy to use assessment tool for identifying hospital inpatients at risk of developing C difficile infection. It makes use of an existing assessment already used with most inpatients in the UK.
A rigorous study has shown that the Waterlow is appropriate for identifying those most at risk of developing C difficile infection, and, as an existing assessment, it is easy to use and implement, thus reducing costs.
This tool will enable nurses to target resource intensive interventions at those patients most at risk, thus preventing a number of these potentially life threatening infections.
Ansari F et al (2003) Outcomes of an intervention to improve hospital antibiotic prescribing: interrupted time series with segmented regression analysis. Journal of Antimicrobial Chemotherapy; 52: 842−8.
Cabinet Office (2007) PSA Delivery Agreement 19: Ensure Better Care for All. Norwich: HMSO.
Cartmill TD et al (1994) Management and control of a large outbreak of diarrhoea due to Clostridium difficile. Journal of Hospital Infection; 27:1-5.
Department of Health (2008) Clean, Safe Care: Reducing Infections and Saving Lives. London: DH.
Department of Health (2007a) The NHS in England: The Operating Framework for 2008/9. London: DH.
Department of Health (2007b) Saving Lives: Reducing Infection, Delivering Clean and Safe Care. High ImpactIntervention No 7. London: DH.
Drudy D et al (2007) gyrA Mutations in fluoroquinolone-resistant Clostridium difficile PCR-027. EmergingInfectious Diseases; 13: 504-505.
Health Protection Agency (2009a) Clostridium Difficile Infection: How to Deal with the Problem. London: HPA.
Health Protection Agency (2009b) Healthcare-Associated Infections in England: 2008-2009 Report. London: HPA.
Health Protection Agency (2008) Surveillance of Healthcare Associated Infections Report: 2008. London: HPA.
Healthcare Commission (2006) Investigation into Outbreaks of Clostridium difficile at Stoke MandevilleHospital, Buckinghamshire Hospitals NHS Trust. London: HC.
Healthcare Commission (2005) Management, Prevention and Surveillance of Clostridium Difficile: InterimFindings From a National Survey of NHS Acute Trusts in England. London: HC.
Kuijper EJ et al (2006) Emergence of Clostridium difficile-associated disease in North America and Europe. Clinical Microbiology and Infection; 12: Suppl 6: 2-18.
O’Conner K (2007) Introducing improved nursing documentation across a trust. Nursing Times; 103: 6, 32-33. tinyurl.com/improved-document
Office for National Statistics (2008) Health Statistics Quarterly 39. London: ONS.
O’Neill G et al (1993) A molecular characterization of Clostridium difficile isolates from humans, animals and their environments. Epidemiology and Infection; 111: 257-64.
Oliver D et al (2004) Risk factors and risk assessment tools for falls in hospital in-patients: a systematic review. Age and Ageing; 33: 122-130.
Simor AE et al (2002) Clostridium difficile in long term care facilities for the elderly. Infection Control andHospital Epidemiology; 23: 696-703.
Tanner J et al (2009) Waterlow score to predict patients at risk of developing Clostridium difficile associated disease. Journal of Hospital Infection; 71: 239-244.
Waterlow JA (1985) A risk assessment card. Nursing Times; 81: 49-55.
WoodsNF et al (2005) Frailty: emergence and consequences: method. Journal of American Geriatric Society; 53: 1321-1330.