During the past 30 years, many important strides have been made in the prevention of nosocomial infections in the United States. Infection control programs have been established in hospitals throughout the country. Techniques for surveillance of nosocomial infections have been developed and utilized extensively. Results of the Study on the Efficacy of Nosocomial Infection Control (SENIC Project) and the experience with surveillance of surgical wound infections have documented the fact that surveillance is an integral component of an effective nosocomial infection control program. In recent years, a number of approaches to nosocomial infection surveillance have been proposed as alternatives to comprehensive or hospital-wide surveillance. In 1986, four surveillance components were introduced in the National Nosocomial Infections Surveillance (NNIS) system to provide participating institutions the option to tailor their surveillance program to their local needs and priorities while continuing to provide information to the national database on nosocomial infections. Infection control practitioners currently face a challenge to develop more meaningful nosocomial infection rates to permit identification of new infection control priorities for their institution and to assess progress toward specific prevention objectives.

This paper addresses the problems associated with defining and classifying events as nosocomial infections, discusses the methods by which rates of nosocomial infection are calculated and their rationales, and presents some specific rates useful in nosocomial epidemiology. Previously unpublished data demonstrate important differences between antibiotic susceptibility tallies produced by clinical laboratories and similar tallies derived from nosocomial infection surveillance data.

Conversion of real world events into categorical data presents formidable difficulties. Surveillance personnel must classify a given series of clinical events as 0,1, or more infections and make a determination as to whether each infection is nosocomial or community acquired. High-quality research studies to validate these efforts should compare the sensitivity and specificity of methods used to some “gold standard.” The gold standard is usually a review of medical records or patients by an infectious diseases physician. But even the standard is flawed. In clinical practice this flaw presents less of a problem because therapy for infectious diseases is generally quite safe and may be instituted when the probability of infection is 10%, 5%, or even lower. For surveillance purposes a higher standard is required, which is particularly important when surveillance information is used to provide feedback data to physicians who understandably bridle at overestimates of infection rates in their patients. The overestimation of infections based on weak evidence under-cuts feedback efforts.

The Centers for Disease Control's Study on the Efficacy of Nosocomial Infection Control (SENIC) showed that infection surveillance and control activities are associated with a decrease in nosocomial infection rates. Moreover, the intensity of activity correlated with the magnitude of the fall in infection rates. These results, plus the guidelines of regulatory agencies, mandate that infection control programs conduct surveillance activities. However, absolute standards for the content and nature of surveillance programs have not been established, and many descriptions of different types of surveillance programs are available. In this primer, we describe the considerations involved in development of a surveillance program with emphasis on issues concerning data collection.

Langmuir considers surveillance when applied to disease as meaning the collection of data, the analysis of those data, and the distribution of the resulting information to those needing to know. The definition implies that surveillance is observational and that surveillance activities should be clearly separated from other related activities such as control measures. The latter activities, including their initiation, approval, and funding, are administrative matters underpinned by a scientific base that are undertaken by the recipients of the surveillance data and their analyses. They should be clearly separated from surveillance activities per se. There is also the implication that action results from surveillance; surveillance without action should be abandoned.

Most information in hospital epidemiology comes from observational studies of hospitalized patients rather than planned experiments, and in such observational studies the characteristics of study patients may vary widely, even within a single hospital. Any comparison between hospital populations will usually contain additional, unintended contrasts among patients with varying degrees of health. Adult patients, for example, may have vastly different underlying diseases, and infants may be of substantially different birth weights. We used both underlying disease and birth weight as indices of the basic severity of illness in order to adjust for confounding by differences in underlying disease in reanalyses of several published studies. We give an example in which differing birth weights among groups of infants compared artifactually double the apparent effect of nosocomial infections as a cause of mortality, and another example in which differing degrees of severity of underlying illness artifactually halve the apparent effect of appropriate antibiotics in preventing death from bacteremia with gram-negative bacilli. We describe simple intuitive methods based on stratification, adapted from chronic disease epidemiology, to remove confounding effects during analyses.

Over 20 years ago, Philip Brachman advised us that “… the surveillance of all institutionally associated infections is important in order to minimize the risk of infection to all patients entering the institution and to members of the community.” Seven years later, in 1970, other staff members at the Centers for Disease Control (CDC) offered us more specific surveillance objectives:

A. To determine the frequency and kinds of endemic nosocomial infections, in order to identify deviations from the baseline so that infection control personnel can:

1. Determine where studies are needed.

2. Ascertain where control measures (long-term and emergency) need to be established and how effective new control measures are.

3. Establish policy.

B. To provide the patient and personnel (and in some instances the community) with all possible protection from infections of nosocomial origin.

C. To meet the requirements of the Joint Commission on Accreditation and the medical-legal guidelines of “accepted standards of patient care.”

D. To provide the medical and nursing staff with meaningful data on the level of nosocomial infection in their work areas.

If each single nosocomial infection represented sufficient deviation from the baseline occurrence, then analysis of surveillance data would be quite straightforward. However, that is not the case. A 1976 project report for the National Center for Health Statistics identified various “sentinel health events” whose occurrence should trigger “… scientific search for remediable underlying causes.”