Prevalence surveys in different countries have shown that about 1 in 10 hospitalized patients at any one time have acquired an infection and the incidence of infection is usually about 3% to 5%. Can this infection rate be significantly reduced and is there an irreducible minimum below which further reduction cannot be obtained? Cross-infection could probably be almost eliminated by the use of plastic isolators or life islands, and endogenous infections reduced by the wider and more rational prophylactic use of topical and systemic antimicrobial agents, and possibly a range of vaccines. However, the use of these techniques on all patients would clearly be impractical and too expensive for use in the foreseeable future. Excessive isolation techniques would have a detrimental psychological effect on many patients. In addition, failure to eradicate the organisms of the normal flora might be associated with more infections caused by antibiotic-resistant bacteria, fungi and viruses which would be more difficult to treat. Even if organisms could be eradicated, morbidity and mortality might not be correspondingly reduced. Priorities must therefore be set on the most effective use of staff and finances. These not only involve a consideration of infection, but all areas of clinical practice, eg, costs of cardiac transplants and provision of geriatric facilities. It therefore seems likely that little additional finance will be available for control of infection, other than possibly for transmissible infections with high mortality such as AIDS.

Urinary tract infections associated with the temporary use of indwelling bladder catheters remain a vexing problem in acute care hospitals. The appreciable morbidity and mortality from these infections are still under active investigation. Because of the large numbers of patients at risk, increased attention is also being given to the costs for prevention, detection, treatment, and follow-up of catheter-associated infections.

Before the advent of commercially available systems for maintaining closed sterile urinary drainage, infection was believed to be inevitable. In the past 25 years, the successful prevention of infection in 70% to 80% of catheterized patients by aseptic methods has undoubtedly prevented thousands of deaths from gram-negative bacteremia and stimulated efforts to further reduce the infection rate.

Infection of the urinary tract is acknowledged to be the most common hospital infection, associated closely with the presence of an indwelling catheter. Patients are catheterized for a variety of reasons and for different periods of time, ranging from postoperative catheterization of a few days following urological surgery to the long-term catheterization over many months or years of patients who are unfit for operation, with spinal injuries or neuropathic bladders. The extent of the problem includes patients returning home infected, requiring catheterization before readmission to the hospital or needing nursing at home with a long-term catheter. The risks of infection and its complications as well as methods of control may differ between each group.

It is of fundamental importance to distinguish between patients who came to surgery with an existing infection and those with sterile preoperative urine. Other factors include recent previous catheterization and, importantly, the length of time the catheter is inserted. For the patient with sterile preoperative urine, postoperative bacteriuria seldom causes severe symptoms and can be treated with antibiotics or left to clear spontaneously after removal of the catheter. A minority of patients suffer consequences of their infection especially when bacteriuria starts before the catheter is removed postoperatively. Catheter removal often causes transient bacteremia, also induced by instrumentation or operation on infected urine, which may lead to serious complications, particularly of septicemia. In our experience in Bristol, about 1 in 4 patients admitted for urological surgery already has infected urine as defined by the presence of ≥105 bacteria/ml midstream urine or ≥=104 bacteria/ml catheter urine. Of those admitted with infected urine, 3 in 4 have a catheter already inserted compared with only 1 catheterized patient in 4 admitted to operation with sterile urine.

Much attention has been given to ventilators, respiratory tubing and humidifiers as sources of bacteria causing pneumonias in intensive therapy unit (ITU) patients. However, efficient decontamination and safe storage of ventilators and humidifiers between patients, with frequent changing of respiratory tubing and the use of Chlorhexidine in humidifier reservoirs, are now undertaken in most ITUs. In addition, chest aspiration, which is an important route for bacterial entry to the chest, is performed by trained ITU staff, who are well aware of this danger and wear gloves and use meticulous techniques. As a result, it can be argued that infection of the lungs in ITU patients from these sources and routes has reached an “irreducible minimum.”

Unfortunately, many patients in ITUs continue to develop pneumonias. In my own experience patients yielded Pseudomonas in tracheal aspirates as frequently after the introduction of these techniques, together with the insertion of efficient bacterial filters, between the patient and the ventilator in both the inspiratory and expiratory lines, as before their adoption. What other sources and routes of infection of the lungs should be considered?

For clean elective surgery, the goal of nosocomial infection control is nothing less than an infection rate of zero. Although infection rates of less than 1% are being posted regularly in selected surgical operations today, infection rates of 2% to 6% persist in many procedures such as coronary artery bypass surgery and vascular surgery. Predictable achievement of a zero infection rate in all of clean surgery will require new directions in both basic and clinical research.

Research into the prevention of infection in clean surgery has, to date, focused upon 1) identifying and controlling the routes of wound contamination, 2) decontamination of a contaminated wound through the prophylactic use of antimicrobials, and 3) modifying the host immune system. The latter is a fascinating area of new research which will not be discussed in detail here.

Routes of contamination are often categorized as endogenous or exogenous. “Endogenous” generally refers to bacterial seeding of the wound from the flora of the patient's own skin, nose, perineum, and GI tract. “Exogenous” contamination has come to mean bacteria originating from the operating room environment or operating room personnel which reach the wound by direct inoculation (eg, improperly sterilized instruments, hands of the surgeon via torn gloves) or by the airborne route. These categories may not be broad enough to satisfy the complexities of the hospital environment, however. The addition of the category “acquired endogenous” may be of help in describing contamination which occurs when patients become colonized with resistant hospital flora and subsequently carry this newly acquired flora into the operating theater.

Any consideration of infection following clean surgery, particularly cardiothoracic must include both exogenous and endogenous sources. The MRC Study on Hip Surgery presents a particular challenge. Although uncontrolled antibiotic prophylaxis reduced the infection rate almost as well as a laminar flow operating theater, further analysis of the data correlated the incidence of infection with the theater air counts and more significantly considered the sources of Staphylococcus aureus infections. Data on swabs from patient and theater staff at the time of surgery were available for 14 of the 28 patients who subsequently developed deep infections with S. aureus. Overall 25% of patients and 33% of staff were carriers. Bacteriophage typing revealed that 2 patients, 2 surgeons and 7 theater staff carried the infecting type at the time of the surgery. In two instances there was no correlation and in the last a member of the surgical staff on leave at the time carried the relevant strain. This at least provides a clue that exogenous infection can be important, and justifies attempts to clean the environment as well as the patient.

The skin is itself the site of infection by many different bacteria. Dermatologists may see primary skin infection due to Staphylococcus aureus, beta hemolytic streptococci and a host of less common pathogens including gonococci, mycobacteria and even enterobacteriaceae, or secondary infection which is mostly staphylococcal in eczema or weighted toward Proteus and Pseudomonas in venous stasis ulcers. Geriatricians see a mixed microbial flora in decubitus ulcers while those who insert plastic prostheses, intravenous or abdominal catheters can expect infection with coagulase-negative or positive cocci.

The skin also forms a natural reservoir of organisms available to infect a variety of patients and it is with the skin as a reservoir that this article is concerned. The normal flora of the skin in healthy subjects is composed chiefly of coagulase-negative cocci and coryneforms with gram-negative bacilli chiefly, but not entirely, represented by Acinetobacter. Patients admitted to the hospital who are severely ill may become more heavily colonized by gram-negative bacilli while frank skin lesions constitute a more obvious source of potential pathogens.

Whereas infections of the skin per se comprise only a fraction of all institutionally-acquired infections, the skin has become one of the most important reservoirs of nosocomial pathogens in the hospital. Professor Noble has provided a scholarly review of the increasing importance of the major constituents of the cutaneous microflora as nosocomial pathogens and what we know of their epidemiology. Unfortunately, the empiricism and limited scientific data which underlie essential infection control measures in this area, particularly in regard to cutaneous antisepsis and handwashing, is almost incongruous in an era in which controlled clinical trials have dominated most other areas of medicine. The numerous outbreaks traced to contaminated antiseptics and disinfectants over the past two decades, stand as mute testimony to the inadequate investigative attention this area has received.

The international recognition of the emergence of methicillin-resistant Staphylococcus aureus as major nosocomial pathogens dates to the late 1970s. The weight of evidence from many institutions suggests that transmission involves spread from patient to patient of the same or very similar strain and that effective control, if not elimination, involves isolation of infected and colonized patients.

More recently, a number of reports have suggested the emergence of methicillin-resistant S. epidermidis and other coagulase-negative staphylococci as significant nosocomial pathogens. Although the organisms undoubtedly share many similarities with methicillin-resistant strains of S. aureus, it is certain that biologic differences exist which are important for infection control.

The “second” and “third” generation cephalosporins offer striking antimicrobial activity against a wide spectrum of Enterobacteriaceae. Nevertheless, mutants resistant to these drugs have emerged in both laboratory and clinical settings. For example, before the commercial availability of the third-generation agents, we treated three cardiac surgery patients for Enterobacter mediastinitis with aminoglycosides and high doses of cefamandole. In two, initial treatment failed due to emergence of strains that were not only resistant to cefamandole, but also to then experimental third-generation drugs. Despite such reports and in vitro studies of the mechanisms of resistance, the frequency with which broad-spectrum cephalosporin resistance develops in clinical practice is not clear. To help delineate this problem, we have reviewed our hospital's experience with Enterobacter strains resistant to newer cephalosporins (using cefamandole and cefotaxime as prototypes) and the relation of resistant strains to cephalosporin use, with special attention to our cardiac surgery patients.

Reducing the concentration of filamentous fungal spores in the air inhaled by immunosuppressed patients is an important goal. Although it is not clear that air is the exclusive vector of nosocomial filamentous fungal disease, it is quite likely that it is the most important vector. Uncertainties about this assertion are compounded by the great variety of fungal species involved. Aspergillus fumigatus, A. flavus, A. terreus, A. niger, Petriellidium boydii, Fusarium sp., Mucoraceae, Phoma sp., Alternaria, Penicillium, and others have caused invasive disease in immunosuppressed patients. But detailed studies of environmental correlates of nosocomial disease are largely restricted to A. flavus and A. fumigatus. Spores of all of these species may be recovered from unfiltered air. It is probably reasonable to extrapolate from Aspergillus to all of the species involved.

Disseminated aspergillosis in the compromised host is an important and intractable problem in Great Britain, having a general incidence, when post-mortem data are included, approaching that quoted for the United States. It emerges as a most important cause of death in units where prolonged granulocytopenia accompanies treatment.

Rhame has presented a persuasive account of the environmental origin of aspergillus infection and the possibility of its control by environmental measures. This is an important issue for those involved in the planning of new units, for the use of protective isolation is under criticism as failing to provide significant benefit, and even general provision of filtered air to the unit may be beyond the budget allowed. We have to consider the minimum necessary measures to reduce the risk. Providing HEPA-filtered air to the unit may not be sufficient in itself; British heart transplant patients have died of aspergillosis (A. fumigatus), in some cases initiated soon after surgery, despite nursing in rooms with filtered ventilation (Newsom SWB, personal communication); and respiratory ventilators have been suspected as one source within such a unit. We also have to consider other approaches to the problem. Is there an endogenous element in the production of aspergillosis? Is there any prospect of successful chemoprophylaxis?

In the hospital, both HTLV-III and hepatitis B virus are most likely to spread via infected blood or contaminated blood products. A summary of the main factors affecting spread of these viruses and some observations made at the author's hospital in central London, will be discussed here.

AIDS was first recognized in Britain in 1981 and up to August 1985 there were 206 confirmed cases of AIDS with 114 deaths. The majority of patients with AIDS have been male homosexuals and less than 15% of the cases occurred outside of London. By August 1985 over 10,000 cases of AIDS had been reported in the US and it is clear by comparison that the current epidemic in Britain is in its early stages. Only a minority of London homosexuals have been infected. It is possible that an aggressive public health campaign could slow the course of the epidemic. Measures needed would include an increase in education, publicity, facilities for HTLV-III antibody tests, trained social workers and counseling on a massive scale. In the meantime the number of hospitalized patients with known HTLV-III infections has increased considerably during the last 2 years in certain hospitals in London (Figure 1). Approximately 33% of homosexuals attending sexually transmitted diseases (STD) clinics in London have serological evidence of HTLV-III infection. A few hospitals in central London have seen most of the patients with known HTLV-III or hepatitis B infections mainly because of the concentrations of homosexuals, drug addicts or other “risk groups” in the local catchment areas of these hospitals. A small number of hospitals in London and elsewhere in Britain have special hemophiliac centers and currently these centers have generally noted that more than 50% of their hemophiliac patients already have serological evidence of HTLV-III infection. Recent reports have documented the rate of increase of HTLV-III infections in British hemophiliac patients. During the last few years in London less than 10% of male homosexuals known to have HTLV-III infection and less than 1% of infected hemophiliac patients have developed AIDS.

Hepatitis B virus and HTLV-III are quite dissimilar in virology, organ tropism and clinical consequences. However, they are remarkably similar in epidemiology, modes of transmission, affected risk groups, and potentials for nosocomial infection. Presently, the two diseases pose much different problems for hospital epidemiologists. Over the past 10 years, the risks of nosocomial hepatitis B have dropped off sharply. In fact, hepatitis B is now theoretically a totally preventable disease. On the other hand, the epidemic of HTLV-III infection in high-risk groups continues unabated. The fear of acquiring infection has precipitated panic reactions among some groups of health care workers and the lay public.

Over the past 25 years group B streptococci have become established as one of the main bacterial pathogens of the neonate in Western Europe and the United States. The attack rate of 0.25/1,000 live births found by Mayon White in Great Britain1 appears typical of many European countries. However, in some centers in the United States attack rates can be over 10 times higher.

Two types of neonatal group B streptococcus (GBS) diseases exist, “early” and “late” onset. Early onset disease usually presents within the first few days of life. Often the most serious infections are present at birth or seen within a few hours. Early onset disease presents with pneumonia, respiratory distress and shock. Bacteremia is normally present and meningitis may occur. Mortality is high (50% to 75%). The portal of entry is probably the respiratory tract. Infants normally acquire the infecting organism from their mothers. Heavy maternal and infant colonization, prolonged rupture of membranes, prematurity, and obstetric complications are all risk factors.

Delayed onset disease, as its name suggests, presents after the first week of life, primarily with bacteremia and meningitis. Mortality is much lower than for the early onset form, but still appreciable for a bacterial infection (14% to 18%). Its epidemiology is uncertain.

Despite increased awareness and the availability of potent chemotherapeutic agents, the Group B streptococcus remains the leading cause of life-threatening bacterial infections in the neonatal period. The majority of infections occur in the first few days of life. These so-called “early-onset” infections are often fulminant and are associated with a very high case fatality rate, ranging from more than 20% in recent reports to as high as 80% in older series. “Late-onset” disease occurs, by definition, after the first week of life. It usually presents with meningitis, although a wide spectrum of infections has been reported, including cellulitis, adenitis, septic arthritis, and osteomyelitis. Late-onset disease is usually less serious than early-onset disease, although deaths do occur and major sequelae are not rare. Accordingly, attempts to prevent group B streptococcal infections have concentrated on the more common and frequently lethal early-onset disease. Strategies that have been considered to date include antibiotic- and immuno-prophylaxis. In the preceding article, Easmon advocates an additional approach; topical use of the antiseptic Chlorhexidine gluconate in the vagina during labor.

Two decades have elapsed since the first detailed studies were undertaken to determine which factors were associated with an increased incidence of life-threatening sepsis during the treatment of leukemia and lymphoma. The state of granulocytopenia is now acknowledged to be the major factor predisposing to infection and accordingly, attention has focused on the merits of preventive measures applied during this risk period. A tremendous effort has been invested in the evaluation of different protocols and there is little doubt that major advances have been made even though a simultaneous improvement in survival rates as a result of more intensive chemotherapy has not always been in evidence. While it is appropriate to initiate prophylaxis before the treatment of most adult acute leukemias and for marrow transplantation, the reduced duration and severity of granulocytopenia during chemotherapy for both childhood acute leukemia and lymphoma generally obviate the need for this approach.

Progress in the development of potent broad-spectrum antibiotics coupled with better clinical awareness of the value of early treatment, as well as changing practices in supportive care have each contributed to the recent pattern of infection that has been observed. In addition, the impact of new technologies, as illustrated by recent innovative applications of bone marrow transplantation, further underlines the need to reappraise current infection control procedures.

Between 1972 and 1973, Britt and colleagues conducted 1-day infection prevalence surveys in 18 hospitals each with fewer than 75 beds and all located in the rural intermountain west. The lessons learned from that study were incorporated into a book chapter entitled “Infection Control in Small Hospitals” by Dr. Britt. Since that study no substantial nor systematic review of nosocomial infections in small community or rural hospitals has been reported. A recent editorial in Infection Control posed four questions that still needed to be answered for the smaller (less than 100-bed) hospital:

1. Given the statistical realities of small hospitals, what types of surveillance methods—periodic prevalence surveys, general surveillance, focused surveillance, etc.—are the most reliable? Are any of them necessary?

2. Which of the many infection control practices recommended by the CDC, the Joint Commission on Accreditation of Hospitals, and others are applicable to small hospitals?

3. What resource sharing of existing expertise, from larger hospitals, health departments, groups such as the Association of Practitioners of Infection Control and the Society of Hospital Epidemiologists of America, and others can be developed for smaller hospitals in cost effective and realistic ways?

4. Should smaller hospitals be required to have the same type of multidisciplinary infection control committees required of larger hospitals or can the responsibilities of the committee be delegated to a smaller group such as one nurse and one staff physician?

The purpose of this report is to provide detailed infection control information obtained from 15 hospitals located in rural areas of Southwestern Wisconsin. This article will focus on data derived from an initial comprehensive prevalence survey that replicated the Britt study in the Wisconsin sites in 1983. Additional data were derived from ongoing bi-monthly prevalence studies done over 6 consecutive months in each of the 15 rural hospitals. Data from 6 months of ongoing comprehensive surveillance from each of the 15 hospitals and final conclusions from the project will be the subject of a subsequent paper.