Hospital Acquired Infection:
A Framework for a National System of Surveillance for the NHS in Scotland

II HOSPITAL ACQUIRED INFECTION SURVEILLANCE

Definition of Hospital Acquired Infection

12. Expressed in the simplest of terms, a hospital acquired infection can be defined as one that was neither present, nor incubating, at the time of admission to hospital. In practice, there are several definitions whose wording differs in minor respects from the above. The most widely accepted definition is that a hospital acquired infection is one which manifests itself 72 hours or more after admission to hospital. This definition, however, may not suit every circumstance. The term .admission".., for example, is clearly not applicable in the context of A&E attendances or day-case surgery, and more detailed definitions exist in relation to specific types of infection.

13. For practical purposes, standard definitions should be determined by whichever body ultimately assumes responsibility for central data collection on HAI in Scotland. It will be essential to ensure that participants in the surveillance scheme work to precisely the same definitions to allow comparability of data.

Definition of Surveillance

14. Surveillance can be defined as the on-going systematic collection, analysis, evaluation and dissemination of data to effectively control HAI. Again. precise definitions for each context should be standardised so that trends and distributions can be compared over time and the success of preventive strategies properly assessed. The definitions must be simple to use and should be devised in consultation with all relevant clinical staff.

Aetiology of Hospital Acquired Infection

Pathogens versus Hosts

15. Infection occurs when a micro-organism invades a susceptible host and causes disease. Any micro-organism that is capable of producing disease is called a pathogen. The pathogen may be a bacterium, a virus, a protozoan, or a fungus. Most HAls result from bacterial infection. Not all pathogens produce disease with equal probability; much depends upon the pathogenicity of the invading micro-organism and the susceptibility of the host. Some micro-organisms regularly cause disease in immunocompetent individuals, whereas others may only cause disease if an individual's immune system is compromised or a natural barrier to infection has been breached. In-patients are more susceptible to infection than the general population because of their state of health and the extent of their exposure to risk of infection. The very young and the very old are at particular risk of acquiring infection in hospital. Other common risk factors include immunosuppression, vascular or urinary catheterisation, surgical drains and underlying diseases, such as diabetes mellitus and malignancy.

Routes of Transmission

16. Infective agents can be transmitted by exogenous or endogenous means. The vast majority, (95%), of surgical infections are endogenous in origin. The four main routes of exogenous transmission are contact, common vehicle, airborne and vector. Contact may be direct or indirect. In the case of direct contact, the source and the host come into physical contact which leads to the direct transfer of micro-organisms. Indirect contact refers to a more passive form of transfer, e.g. hand to mouth. In common vehicle transmission, the microorganism is transmitted via inanimate objects such as fomites, (towels, bedding., etc.), or food and water. Airborne transmission refers to the transmission of micro-organisms by air or dust. Vector transmission relates to transmission via living creatures, such as insects, lice and ticks, which move from one host to another.

Types of Surveillance

17. The main types of surveillance can be categorised as follows:

• Targeted surveillance
• Selective surveillance
• Continuous surveillance (e.g. of "'alert organisms" or ""alert conditions")
• Prevalence surveys.

18. Targeted surveillance makes use of data on HAI which occurs in a defined sub-group of patients, such as those in a particular ward or department, those undergoing a specific procedure or those acquiring a particular infection.

19. Selective surveillance makes use of data on HAI from a pre-determined "selection"" of patients with the aim of consistently identifying the majority of cases of HAI as a satisfactory alternative to the more time-consuming method of targeted surveillance.

20. Continuous surveillance, undertaken on an on-going basis, can be used in specific "high risk" areas, such as intensive care, neonatal intensive care., transplant, renal dialysis, burns and oncology units. It may also be used in the case of vulnerable patients, such as the immunocompromised, and the very young or the very elderly patient.

21. Surveillance of .'alert organisms" is widely employed throughout the UK to detect and prevent outbreaks of infection. An outbreak is suspected when two or more indistinguishable organisms are isolated from different patients in the same ward or unit. The success of this type of surveillance is dependent on ward staff taking appropriate samples for microbiology from all patients suspected of having the infection.

22. Surveillance of >"'alert conditions" is the process of reporting to the Infection Control Team clinical conditions, such as suspected infective diarrhoea and/or vomiting, food poisoning, measles, mumps, rubella, etc, which may give rise to hospital outbreaks.

23. Prevalence surveys are used to measure the proportion of patients infected during the time period of the survey. Repeated, well-designed surveys can provide useful data on infection trends and the efficacy of infection prevention and control measures. However, the results are usually of more limited value than those obtained from incidence studies. Prevalence studies are therefore probably best used as an adjunct to other surveillance methods.

Importance of Post-Discharge Surveillance

24. Research indicates that between 20% and 70% of surgical wound infections may present after discharge¹⁵". Little is known about other types of post-discharge infection. Many, however, have associated high economic costs that continue long after the original event.

Despite the cost and time required to do post-discharge surveillance, factors such as dramatic changes in patient management, (leading to shorter lengths of stay), and the advent of day case surgery, underscore the need for both post-discharge surveillance and out-patient surveillance to be undertaken as an integral part of the overall surveillance programme.

Prevalence, Incidence and Effect on Mortality

Prevalance

25. Prevalance may be expressed as the proportion of a population studied who have an HAI at a single point in time, usually referred to as a prevalence rate. Most of the information relating to HAI is derived from prevalence surveys. Two national prevalence surveys were undertaken in the United Kingdom in 1980 and 1994 respectively^16,17.The 1980 study surveyed 18,186 patients in 43 hospitals and identified a 9.2% prevalence rate of all inpatients with infections of the urinary tract, surgical wound and lower respiratory tract respectively being the most frequent. The 1994 study identified a similar rate. These figures are consistent with findings in other parts of Europe which show that HAI is generally found to have a prevalence rate of approximately 10%.

26. Although overall prevalence rates are similar, HAI prevalence will vary between different patient sub-groups because of their different risk factor profiles. It is interesting to note that in the two studies referred to above, the rate of surgical wound infection had fallen from 24% in 1980 to 16% in 1994 whereas the rate for infections of the lower respiratory tract had risen from 21 % in 1980 to 34% in 1994. However, these statistics should be interpreted with caution since neither study took into account the occurrence of HAI post-discharge which means that they may well have underestimated the full extent of the problem.

Incidence

27. Incidence may be defined as the number of new cases of HAI which occur within a particular time period in a specified population. Incidence rates are expressed either as the number of patients with an infection per thousand patients discharged within a specified period (which is then reduced to a percentage), or as the number of infections per thousand patients discharged (also reduced to a percentage) during this period. The latter method is often employed since it is not uncommon for a patient to develop more than one HAI. The majority of studies which attempt to estimate the incidence of HAI have focussed on particular patient groups or types of infection ^18,19. Few studies have tried to estimate the incidence of all types of HAI and those that have done so do not take into account the incidence of HAI which presents post-discharge ¹⁹. As with prevalence, incidence rates vary considerably according to the type of patient or specialty under review.

Effect on Mortality

28. Data from studies in the U S suggest that 10% of patients who acquire an HAI will subsequently die as a direct result of the infection ²⁰ . The studies also suggest that the infection will be a major contributory factor in the death of a further 30% of these patients. When the figures were extrapolated to the whole of the US population presenting with an HAI over the course of a single year, it was estimated that in 1982, some 20,000 deaths were directly attributable to HAI and that a further 60,000 were indirectly attributable to HAI. This means that, in the US, HAI is the eleventh leading cause of death. When deaths which are both directly and indirectly attributable to HAI are taken into consideration, HAI becomes the fourth leading cause of death.

29. In the UK, it has been estimated that 5,000 deaths per year are as a direct result of HAI and that HAI is a significant contributing factor in a further 15,000 deaths per year 3 . HAI might therefore be a more common primary cause of death than road traffic accidents or suicides.

Cost of Hospital Acquired Infection

Costs within the Hospital

30. Few studies have been published which examine the cost of HAI in UK hospitals but in 1989 Currie et a 1 ²¹ estimated that reductions in the incidence of HAI by 20%,35% and 50% across the UK could produce annual savings of £15.6m., £29.3m and £50m respectively to the NHS after off-setting the costs of infection control teams and their programmes.

31. Costs should take into account increased length of stay, extra investigations, drugs and dressings etc, and the opportunity costs, i.e. those spent on diagnosis and treatment of HAI which could have been spent on treating other patients. Increased length of hospital stay is likely to be the biggest cost factor of all, closely followed by prolonged antibiotic use and additional laboratory tests. Infections which result in outbreaks, [e.g. methicillin-resistant staphylococcus aureus, (MRSA)], can be very expensive and may result in ward closures and cancellations of operations and admissions. Staff may have to undergo tests and be off work.

32. In the case of an outbreak, additional costs will be incurred through the use of agency staff or overtime required to meet extra demands on staff; the use of antimicrobial agents; patient and staff screening; cleaning services; the provision of isolation wards; ward closures and reduced clinical activity, etc. One major MRSA outbreak in England ²² was estimated to cost in excess of £400,000 in direct costs alone, excluding those costs associated with increased length of stay, additional prescribing, and those arising from staff absence due to infection, or litigation.

33. Further "hidden" costs associated with an outbreak may be those of adverse publicity resulting in loss of confidence in the Trust (by the public and health care commissioners),³ and significant use of senior management and professional time

34. Results of cost-of-illness studies are reproduced in Table I.

TABLE I: RESULTS OF COST-OF-ILLNESS STUDIES OF HAI BY TYPE OF INFECTION, LENGTH OF STAY AND EXTRA COSTS OR CHARGES: 1978-1994.

Key: SWI = surgical wound infections; UTI = urinary tract infections; LRTI lower respiratory tract infections.
+ All costs have been converted into sterling using the OECD "'Health Data" database (1996).
They have been adjusted to 1995/96 prices by using a factor series that takes account of hospital input cost inflation in the UK (DoH 1997).
Patients in this study may have more than one infection.
Source: Plowman R M, Graves N, Roberts J A. Hospital Acquired Infection. Office of Health Economics, 1997.
Reprinted with kind permission from the Office of Health Economics.

Costs to the NHSiS

35. Cost of illness studies are frequently used to justify calls for resources. These studies often fail to take into account the ".avoidable" burden of disease. It is important to consider the avoidable burden of HAI because such an estimate would take into account both the effectiveness of any intervention and the standard of current practice. For example, not all HAI can be reduced by surveillance and some hospitals may already have good surveillance policies.

36. The total burden of HAI and the avoidable burden can be estimated in Scotland by involving a number of assumptions:

• The prevalence of HAI is 9% ¹⁷
• The effectiveness of surveillance at reducing HAI is 33% for an intensive programme and 18% for a moderately intensive programme ⁶
• Approximately one third of Trusts have an intensive surveillance programme, a third a moderately effective programme and a third a low intensity programme a
• On average, an infection increases length of stay by 2 days at a cost of £314 excluding fixed costs.'

a Estimate (although not dissimilar to data from the Surveillance Activities Questionnaire.

b Estimate. Two days. The average cost of a Scottish general medical bed is £157 per day + fixed costs at 23%, giving a total of £204 (Blue Book).

37. Table II illustrates that there are an estimated 69,000 infections in Scotland per annum, of which 12,500 are potentially avoidable.

TABLE II: PREVENTABLE HOSPITAL ACQUIRED INFECTION BY

SURVEILLANCE IN SCOTLAND

Notes: Continuous 1n-patient Stay, ISD record linkage, 1995. Linear trend to 1998
Expected prevalence (9.0)33(x+y+z), where y = x/0.82 and z = x/0.67.

38. Based on the above estimations and assumptions, the total cost of HAI to Scottish hospitals is approximately £21.6m per annum. The avoidable burden of HAI is substantially less at approximately £3.9m per annum. The avoidable burden could range from £0.54m to £7.88m depending upon the effectiveness of surveillance and the cost of an infection, (see Table II).

TABLE Ill: AVOIDABLE BURDEN OF HOSPITAL ACQUIRED INFECTION BY SURVEILLANCE

39. Table III illustrates the considerable uncertainty surrounding attempts to cost the avoidable burden of disease. It should be noted that the avoidable burden represents potential resource savings, not realisable financial savings. The costs have been calculated on the basis of the opportunity cost of resources used in treating those with HAI. In practice, the true cost is likely to be substantially higher since these figures do not include the cost of treating HAI manifesting in the community, costs to patients and their families or additional costs to society.

40. Economic evaluation of HAI is still in its infancy but the London School of Hygiene and Tropical Medicine in conjunction with the PHLS is in the process of developing a model to cost the socio-economic impact of HAI. The results of their deliberations should be published later this year.

Conclusion

41. For the individual patient, improved infection control measures, including HAI surveillance, will result in lower rates of infection, higher rates of survival, decreased morbidity, shorter duration of stay and a more rapid return to health. For the NHSiS, it will help to prevent the spread of infection., reduce diagnostic and treatment costs, reduce length of stay (which. in turn, will also help to reduce waiting lists), avoid litigation and negative publicity.