The epidemiology of health-care-associated infections in home-care settings has not been defined, but infections certainly occur. Outbreaks have been documented in association with use of central venous catheters, parenteral nutrition, bathing practices, educational level of caregivers, and the introduction of new products, such as needleless devices for intravenous infusion (6–8).

Needleless devices are used for connecting and accessing intravenous infusion tubing, replacing traditional needles. These devices are used in both home and hospital settings and are perceived to be safe for patients and effective in reducing needlestick injuries.

From 1993 through 1995, the Hospital Infections Program, Centers for Disease Control and Prevention (CDC), investigated three outbreaks of bloodstream infections (BSI) in patients receiving infusion therapy in their homes. In all three outbreaks, needleless devices were associated with BSIs. The first outbreak occurred in Rhode Island in 1993-1994. The endcaps on these devices were changed every 7 days. BSIs were frequent when needleless devices were used to administer total parenteral nutrition (6). The second outbreak, in Oakland, California, during 1992-1994, occurred among pediatric hematology-oncology patients. The BSI rate was higher when needleless devices were used by Asian or Hispanic children but not by white or black children. The racial/ethnic differences were thought to stem from socioeconomic factors or possibly from language barriers that prevented full understanding of instructions on infection control (7). The third outbreak occurred in Houston, Texas, in 1994-1995. The BSI rate was higher when the needleless device endcaps were changed every 7 days and lower when they were changed every 2-3 days. Patients who showered may have had a higher BSI rate than those who took tub baths (8).

These outbreak investigations were, by necessity, retrospective, and some data were difficult to obtain. To better define the epidemiology of BSIs in the home-care setting, in 1995 the Hospital Infections Program conducted a prospective multicenter study of home infusion therapy patients. The objectives were to determine rates of BSI and to identify risk factors, especially the use of needleless devices. The study, which was conducted in Cleveland, Ohio, and Toronto, Canada, involved 827 patients (69,532 catheter-days) (9). The most common underlying diagnoses among this cohort were infections caused by organisms other than HIV (67%), malignancy (24%), nutritional and digestive disease (17%), heart disease (14%), organ transplantation (11%), and HIV infection (7%).

Overall, 7% of these patients had one or more BSIs during a median of 44 days of catheter use (range 1 to 395 catheter days). A multivariate analysis showed that independent risk factors included recent bone marrow transplant, receipt of total parenteral nutrition, receipt of infusion therapy outside the home (e.g., in a clinic or physician’s office), use of a multilumen catheter, and having had a previous BSI (9). Needleless devices were not associated with BSI.

Two prevalence surveys of infections among patients of Missouri home health agencies were conducted by CDC in collaboration with the Missouri Alliance for Home Care (MAHC) and the Missouri Department of Health, the first during summer (June 1-30, 1999) and the second during winter (February 15-March 15, 2000). Of 5,100 home-care patients enrolled in the summer survey, 16% (793) were reported to have infections; 8% (63) of these infections were reported as being acquired at home, 16% (127) as hospital acquired, 35% (278) as unknown source, and 41% (325) as community acquired. The infection sites reported were urinary tract (214 [27%]), respiratory tract (190 [24%]), skin or soft tissue (190 [24%]), surgical site (95 [12%]), or bloodstream (17 [2%]); 18% (143/793) of infections occurred at other body sites (e.g., gastrointestinal, bone) (10). Of 2,890 patients enrolled in the winter survey, 16% (466) had infections. The prevalence of respiratory tract infections was higher during the winter survey than during the summer survey. These results suggest that an estimated 1.2 million patients receiving home care in the United States have infections annually, supporting a need for surveillance of infections among home-care patients (11).

A nationwide hospital-acquired infection surveillance system and standardized infection definitions have been in existence since the 1970s (12–14). However, no national surveillance or standardized definitions exist for monitoring infections in the home-care setting. Recently, the Association for Professionals in Infection Control and Epidemiology, Inc. (APIC) published draft definitions for surveillance of infections in home-care patients (15). However, these definitions have not yet been validated.

National surveillance of health-care-associated infections in home care may potentially decrease infection rates, as has been documented in hospitals by the National Nosocomial Infections Surveillance system (NNIS). This voluntary, hospital-based reporting system was established to monitor hospital-acquired infections and to guide the prevention efforts of infection control practitioners. During 1990-1999, risk-adjusted infection rates in intensive-care units decreased by approximately 40% among hospitals participating in NNIS (16).

A national system for surveillance of health-care-associated infections in home care would not only provide useful data on incidence and types of infections but also simplify identification of risk factors for infection and development of national benchmarks for comparing infection rates. Risk-adjusted rates may assist individual home-care agencies to identify areas for performance and quality improvement and to evaluate the impact of prevention interventions on infection rates.

Home-care surveillance poses several unique challenges, including lack of nationally accepted standard definitions and surveillance methods, loss of patient follow-up, lack of trained infection control personnel in home-care settings, difficulty in capturing clinical and laboratory data, and difficulty in obtaining numerator and denominator data.

Several groups are collecting data on health-care-associated infections in home care and other outpatient areas. These data may prove useful in developing a national home health-care surveillance system.

MAHC is a nonprofit association that provides home care education, advocacy, and information for its 250-member agencies, most of which are located in Missouri. In the early 1990s, MAHC established an infection control committee composed of nurses who provided infection control activities for their agencies. In 1993, the committee implemented the MAHC Infection Surveillance Project (ISP) to monitor infections associated with central venous and urinary catheters. ISP is an active surveillance system that uses standardized criteria and definitions for tracking, aggregating, and reporting urinary infections and BSIs among home-care patients. Currently, 88 home-care agencies from 23 states participate in ISP. Although MAHC has contracted with the Hospital Industry Data Institute, Missouri Hospital Association, to organize and present the ISP data, the results have not yet been published. Although the ISP definitions have not been validated to determine sensitivity and specificity, the data allow participating agencies to compare their infection rates with those of other agencies.

On a broader scale, the Health Care Financing Administration (HCFA), in collaboration with the Center for Health Sciences and Policy Research, has developed the Outcome and Assessment Information Set (OASIS) to measure patient outcomes and improve quality in home care. HCFA requires all Medicare-certified home-care agencies to electronically submit data for their Medicare patients to a central OASIS database in Baltimore, Maryland. The outcomes monitored in OASIS are changes in patient health status, as indicated by need for emergency care or hospitalization, for example. Data collected include patient demographics and medical history, living arrangements, type of wound, urinary tract infection, respiratory devices, medications, emergency care received, transfer to an inpatient facility, and death. Most data items are obtained at start of care, every two calendar months, and at discharge. Since August 1999, eight million records have been entered into the OASIS database, and release of an initial report is anticipated in 2001 after the data are tested for accuracy (HCFA, pers. comm.)

Another national and international data source is the Outpatient Parenteral Antimicrobial Therapy (OPAT) registry, which aims to improve delivery of care and outcomes for outpatients receiving parenteral antimicrobial therapy. OPAT provides a broad database for assessing antimicrobial drug-prescribing practices and outcomes among patients with infections treated in outpatient settings. Data collected include patient demographics, diagnosis, pathogen, venous access device, infusion system, adverse events, clinical outcome, and patient satisfaction. Currently, 23 OPAT provider sites from 19 states are participating in the U.S. registry, and 26 provider sites from 6 countries are in the international registry. OPAT data have been presented at scientific conferences (23).

Surveillance methods that are commonly used in hospital programs may not be feasible for home care. Different strategies are needed to make surveillance in the home easier to implement, particularly if adequately trained staff and diagnostic services are limited. For example, the Dialysis Surveillance Network provides a novel way of tracking hospitalization, antimicrobial use, and selected infections in hemodialysis outpatients (24). Episodes of potential infection are identified by a clearly defined sequence of steps that involves completing an “incident form” for all patients admitted to a hospital or started on intravenous antimicrobial therapy. The presence (or absence) of symptoms indicating infection is recorded rather than the infections themselves, and a computer algorithm determines whether the infection case definitions are met; the data collector is not required to memorize case definitions. The lessons from this surveillance system, in addition to other traditional outpatient systems, may be useful in establishing national surveillance for home health-care-associated infections.

Nearly as many patients receive home care annually as hospital care. With the continued expansion of home health-care delivery and documented infection risk in this setting, a national system for surveillance of health-care-associated infections in the home-care setting is needed. Collaboration between home health-care agencies, state and federal health agencies, private industry, and national or managed-care organizations is essential to make this system feasible and functional. Development and implementation of such a system would foster better understanding of the epidemiology of health-care-associated infections in the home-care setting. Furthermore, this system would provide a means for monitoring the impact of interventions aimed at preventing the emergence of these infections in the home.