Favorable Impact of an Infection Control Network on Nosocomial Infection Rates in Community Hospitals
Objective. To describe an infection control network (the Duke Infection Control Outreach Network [DICON]) and its impact on nosocomial infection rates in community hospitals.
Design. Prospective cohort study of rates of nosocomial infections and exposures of employees to bloodborne pathogens in hospitals during the first 3 years of their affiliation with DICON. Attributable cost and mortality estimates were obtained from published studies.
Setting. Twelve community hospitals in North Carolina and Virginia.
Results. During the first 3 years of hospital affiliation with DICON, annual rates of nosocomial bloodstream infections at study hospitals decreased by 23% (
). Annual rates of nosocomial infection and colonization due to methicillin‐resistant Staphylococcus aureus decreased by 22% (
), and rates of ventilator‐associated pneumonia decreased by 40% (
). Rates of exposure of employees to bloodborne pathogens decreased by 18% (
).
Conclusions. The establishment of an infection control network within a group of community hospitals was associated with substantial decreases in nosocomial infection rates. Standard surveillance methods, frequent data analysis and feedback, and interventions based on guidelines and protocols from the Centers for Disease Control and Prevention were the principal strategies used to achieve these reductions. In addition to lessening the adverse clinical outcomes due to nosocomial infections, these reductions substantially decreased the economic burden of infection: the decline in nosocomial bloodstream infections and ventilator‐associated pneumonia alone yielded potential savings of $578,307 to $2,195,954 per year at the study hospitals.
Received December 30, 2003; accepted May 24, 2004; electronically published February 28, 2006.
Most studies of the incidence and economic impact of nosocomial infections in the United States have been done in university hospitals and tertiary care centers.1‐3 Data from such studies cannot be directly applied to community hospitals. In community hospitals, challenges pertaining to nosocomial infection abound: infectious diseases specialists are often lacking, and staffing and support dedicated to infection control programs are often inadequate.4‐7 Infection control practitioners (ICPs) are often called upon to act in several roles and to perform multiple tasks in addition to infection control activities. In addition, many community hospitals do not have access to physicians trained in epidemiology, data analysis, and statistical methods.
The Duke Infection Control Outreach Network (DICON) was established in 1997 and now includes 22 community hospitals in the southeastern United States. The 3 physician epidemiologists and 3 full‐time ICPs associated with DICON collaborate with and help supervise a total of 31 ICPs working in affiliated community hospitals. DICON’s goals are to collect data from community hospitals and to use these data to educate key personnel in member hospitals and motivate them to understand local problems and patterns related to nosocomial infection. By fostering such understanding, DICON strives to improve the performance and quality of care delivered to patients at member hospitals. The objectives of this report were to provide a functional overview and description of DICON and to describe the impact of DICON on nosocomial infection rates at participating community hospitals.
Methods
Structure of DICON
The 3 physician epidemiologists and 3 network ICPs frequently provide consultation and make regular on‐site visits to hospitals affiliated with DICON. The liaison ICPs visit each hospital and work with local ICPs once or twice each month. In total, DICON’s liaison ICPs annually make approximately 115 visits to network hospitals and travel approximately 15,000 miles. Additional contacts occur between DICON personnel and local community infection control personnel via e‐mail and telephone conferences. During the course of each year, physician epidemiologists give periodic summaries of collected data and provide educational programs to medical staff members at affiliated hospitals.
For each hospital, data pertaining to hospital‐acquired primary bloodstream infection (BSI), colonization and infection due to methicillin‐resistant Staphylococcus aureus (MRSA), ventilator‐associated pneumonia (VAP), catheter‐associated urinary tract infection, and exposure of employees to bloodborne pathogens are collected according to standard definitions.8 In addition, procedure‐specific surgical site infection data are collected. Nosocomial infection rates are analyzed routinely according to the definitions of the Centers for Disease Control and Prevention (CDC) and are benchmarked against national rates provided by the National Nosocomial Infection Surveillance System (NNIS).8,9 In addition, infection rates are benchmarked both against data from other DICON institutions and longitudinally within each institution. Quarterly, semiannual, or annual reports summarizing surveillance activities, data analysis, and recommendations are provided to member hospitals. At an annual conference each winter, data trends from the preceding year are summarized and new goals and objectives for the coming year are discussed. Strategies for reducing nosocomial infection rates are based primarily on interpretation and analysis of surveillance data and CDC guidelines.
Study Design, Data Collection, and Study Sites
This was a prospective cohort study. All nosocomial infections were identified through standard surveillance methods by ICPs using NNIS definitions.8 Exposures of employees to bloodborne pathogens were self‐reported by affected personnel; additional detailed data were obtained from these persons by interviews with ICPs using standardized data forms.
The study included all hospitals that had joined DICON between 1998 and 1999 and for which DICON had collected 3 or more years of consecutive data. In this report, year 1, year 2, and year 3 were defined as the first, second, and third years, respectively, after a hospital joined DICON. Thus, the data used in this study were collected from January 1998 through December 2001 but included only 3 years of data from each hospital.
Statistics
All analyses were performed with Epi Info 6 (CDC). Fisher exact test and the 2‐tailed χ2 test were used to compare proportions, and the χ2 test for trend was used to detect trends in multiple strata. Infection rates were compared by using the incidence density ratio of year 1 and subsequent years; these rates were analyzed by using the z test statistic.10
Attributable Cost and Mortality Data
Estimates for savings attributable to reductions in the rates of nosocomial BSIs and VAP were obtained from recently published studies.11‐17 In one study, the direct attributable cost of nosocomial BSIs was reported in euros. We converted these costs to US dollars by using the lowest exchange rate during the study period (1 
on October 24, 2000).18
Results
Characteristics of study hospitals are detailed in the Table. The study hospitals were community hospitals in North Carolina and Virginia that ranged in size from 27 to 537 patient beds (mean, 223; median, 198.5). Seventeen ICPs worked in the study hospitals and collaborated with 2 DICON‐liaison ICPs during the study period. During the study period, the total number of patient‐days at the 12 study hospitals was 1,456,084 (mean, 505,795 days per year; median, 485,903 days per year). The total number of patient‐days at each institution ranged from 21,119 to 319,576 over the study period. The mean number of patient‐days per hospital per year was 40,447 (median, 33,404; range, 7,040 to 106,225).
Twelve hospitals provided data about exposure of employees to bloodborne pathogens, the incidence of nosocomial BSI, and infection and colonization with MRSA; 10 hospitals provided data about the incidence of VAP during the study period. Data about catheter‐associated urinary tract infections were not routinely collected by all participants and thus were not included in this analysis.
There were 251 episodes of nosocomial BSI in the first year, 188 in the second year, and 166 in the third year (Figure 1). During the study period, BSI rates decreased in 6 of the hospitals, remained stable in 3 hospitals, and increased in 3 hospitals. Among DICON study hospitals as a whole, rates of nosocomial BSI decreased by 23% between years 1 and 3 (
; rates for each of the 3 years, 0.49 per 1,000 patient‐days in the first year, 0.40 per 1,000 patient‐days in the second year, and 0.37 per 1,000 patient‐days in the third year; 
for trend).
Figure 1. Annual incidence of nosocomial bloodstream infections (BSIs) in 12 hospitals during the initial 3 years of DICON affiliation. There was a statistically significant trend toward reduced rates of BSI during the study period (
). Comparison of data from year 1 with data from year 3 data showed a 23% reduction in the incidence of BSI (
).
There were 370 episodes of nosocomial infection or instances of nosocomial acquisition of colonization due to MRSA in the study hospitals in year 1 of the study, 328 episodes in year 2, and 278 episodes in year 3 (Figure 2). During the study period, rates of nosocomial infection and colonization due to MRSA decreased in 6 of the hospitals, remained stable in 3 hospitals, and increased in 3 hospitals. Among DICON study hospitals as a whole, rates of nosocomial infection and colonization due to MRSA decreased by 22% between years 1 and 3 (
; rates for each of the 3 years: 0.76 per 1,000 patient‐days in the first year, 0.65 per 1,000 patient‐days in the second year, and 0.59 per 1,000 patient‐days in the third year; 
for trend).
Figure 2. Annual incidence of nosocomial methicillin‐resistant Staphylococcus aureus (MRSA) infections in 12 hospitals during the initial 3 years of DICON affiliation. There was a statistically significant trend toward reduced rates of MRSA during the study period (
). Comparison of data from year 1 with data from year 3 showed a 22% reduction in the incidence of MRSA infection (
).
During the study period there was a total of 38,431 ventilator‐days (mean per hospital, 1,281 ventilator‐days per year; median, 894.5) and a range of 200 to 15,486 ventilator‐days (mean annual range per hospital, 67 to 5,162 ventilator‐days; median range, 54 to 4,965). Rates of VAP decreased in 6 hospitals, remained stable in 2 hospitals, and increased in 2 hospitals. Among DICON study hospitals as a whole, there were 115 episodes of VAP in the first year, 108 episodes in the second year, and 73 episodes in the third year. The incidence of VAP decreased by 40% between years 1 and 3 in study hospitals (
; rates for each of the 3 years, 10 per 1,000 ventilator‐days in the first year, 8 per 1,000 ventilator‐days in the second year, and 6 per 1,000 ventilator‐days in the third year; 
for trend) (Figure 3).
Figure 3. Annual incidence of ventilator‐associated pneumonia (VAP) in 10 hospitals during the initial 3 years of DICON affiliation. There was a statistically significant trend toward reduced VAP rates during the study period (
). Comparison of data from year 1 with data from year 3 showed a 40% reduction in the incidence of VAP (
).
There were 495 employee exposures to bloodborne pathogens in the first year, 379 in the second year, and 360 in the third year. During the study period, rates of employee exposures to bloodborne pathogens decreased in 6 hospitals, remained stable in 2 hospitals, and increased in 4 hospitals. The rate of employee exposures to bloodborne pathogens decreased by 18% between years 1 and 3 in DICON study hospitals (
; rates for each of the 3 years, 0.96 per 1,000 patient‐days [20.6 per 100 hospital beds] in the first year, 0.76 per 1,000 patient‐days [15.8 per 100 hospital beds] in the second year, and 0.77 per 1,000 patient‐days [15.0 per 100 hospital beds] in the third year; 
for trend) (Figure 4).
Figure 4. Annual incidence of exposure of employees to bloodborne pathogens (BBPEs) in 12 hospitals during the initial 3 years of DICON affiliation. There was a statistically significant trend toward reduced rates of exposure of employees to bloodborne pathogens during the study period (
). Comparison of data from year 1 with data from year 3 showed an 18% reduction in the exposure of employees to bloodborne pathogens (
).
In 3 hospitals, decreases were noted during the study period for all 3 nosocomial infections studied (and for exposures to bloodborne pathogens). No hospital had increases in the rates for all 3 nosocomial infections during the study period.
Discussion
To our knowledge, this is the first study to examine the impact of an infection control network on nosocomial infection rates in community hospitals. Using prospectively collected data, we demonstrated significant reductions in the rates of nosocomial BSI, VAP, and colonization and infection due to MRSA during the first 3 years after community hospitals joined DICON. In addition, rates of exposure of employees to bloodborne pathogens decreased significantly. These successes were attained through standard surveillance methods; frequent analysis, interpretation, and feedback regarding results; and interventions based largely on CDC guidelines and protocols.
A limited number of studies have examined the impact of infection control programs on community hospitals.6,19‐21 This report provides important data pertaining to the incidence of nosocomial infections in a network of community hospitals and the positive effect that an infection control program can have on patient care in these institutions. It is generally assumed that instituting an infection control program decreases nosocomial infection rates. However, this assumption is based on few data.22 This report provides additional data supporting the causal relationship between the implementation of an infection control program and decreases in nosocomial infection rates.
DICON provides numerous benefits to member hospitals. Experienced liaison ICPs make routine visits to local hospitals to oversee infection control activities and to assist with data collection, analysis, and interpretation. In addition to these routine visits, liaison ICPs are available to answer questions and to help manage emergencies and outbreaks. Physician epidemiologists make 2 to 12 regular routine visits to local hospitals each year. Physicians are also available to address urgent questions and concerns. In addition, physician epidemiologists help analyze and interpret surveillance data and oversee the design and implementation of infection control interventions. DICON also provides benchmarks for nosocomial infection rates in our affiliated community hospitals. These “benchmarked” data compare rates among community hospitals that are of similar size and have similar volumes of surgical procedures. Such data are generally of intense interest to local ICPs, medical and surgical staff, and administrators. Our experience has been that comparisons such as these often serve as a catalyst for better adherence to hand hygiene and barrier precaution guidelines and that they motivate local practitioners to pay attention to infection control standards and local policies. DICON also provides a forum at which local ICPs can discuss issues and interventions with one another. Some of these interactions have led to networkwide initiatives (eg, feedback about findings related to the appropriateness of preoperative antimicrobial prophylaxis for patients who experience postoperative surgical site infections).
There are several potential reasons for the decreases in infection rates and in exposure of employees to bloodborne pathogens after hospitals joined DICON. First, standardized data‐collection methods were instituted. Hospital‐wide surveillance was limited to life‐threatening infection (ie, BSI) and antimicrobial‐resistant organisms (eg, MRSA), whereas surveillance for device‐related infections was restricted to intensive care units. These focused surveillance activities helped to streamline the effort and limit the workload of local ICPs. None of the hospitals changed their surveillance policies during the study period. One hospital performed targeted “active” surveillance throughout the study period by using nares cultures to screen patients newly admitted to the intensive care unit for MRSA carriage. Second, surveillance data were routinely analyzed, and summaries were given to providers. Surveillance data were used to design interventions and focus them on high‐risk units.
Third, on the basis of the recommendation of epidemiologists working with DICON, the use of antimicrobial‐ or antibiotic‐coated vascular catheters was instituted at 3 hospitals during the study period. Hospital ICPs were intensively educated about optimal sterile technique for vascular catheter insertion, and several of them implemented individualized educational programs within their hospitals. Both of these interventions, along with regular feedback about central venous catheter–associated BSI rates, may have played a role in the remarkable reduction in nosocomial BSI rates that was observed during the study period.
Fourth, waterless, alcohol‐based hand hygiene products were introduced in 7 study hospitals during the study period. Fifth, patients with a history of MRSA upon admission to the hospital were promptly placed in contact isolation. During the study period, 2 hospitals began electronically flagging the charts of patients with a history of MRSA. Sixth, the time between the initial positive culture for MRSA and the placement of the patient in contact isolation was monitored for every patient in all study hospitals; these data were analyzed and compared with those from other DICON hospitals, and the results were provided to local ICPs. These time‐trended and comparative data often stimulated personnel from participating hospitals to improve their local performance. Finally, during the study period, new medical devices with safety features aimed at preventing needle sticks were implemented in all hospitals, and regular reports about the rate of exposure of employees to bloodborne pathogens were provided to each hospital.
The primary limitation of this observational study is that there was no control arm: all hospitals were affiliated with DICON, and it was impossible to study “control” hospitals outside the DICON network. In addition, several infection control–related events and interventions occurred in different hospitals during the study period. Thus, it is difficult to independently attribute changes in infection rates to affiliation with DICON. DICON’s physicians and liaison ICPs were, however, involved in the oversight of all infection control–related interventions during the study period. An additional limitation is that employee exposures to bloodborne pathogens were self‐reported and, thus, may have been under‐reported throughout the study period.
Nosocomial infections have a substantial adverse impact on patient outcomes and costs in community hospitals. Between years 1 and 3 of DICON affiliation, 95 fewer episodes of nosocomial BSI occurred, and 103 cases of hospital‐acquired colonization and infection due to MRSA and 83 episodes of VAP may have been prevented. Using recently published estimates of the cost and mortality attributable to nosocomial infections, we suggest that the reductions in nosocomial BSIs and VAP saved 1011,14 to 3313,14 lives and reduced hospital costs by $1,734,92012,15 to $6,587,86213,16 during the study period ($578,307 to $2,195,954 per year). These findings suggest that the implementation of an infection control network was effective in improving patient care and decreasing hospital costs. Moreover, the total cost savings greatly exceeded the total cost of participation in DICON (data not shown).
Local participation in DICON includes benefits that cannot be measured in terms of economics. We believe that the attitudes of local ICPs and of professional and administrative staff were positively affected by their participation in DICON. Time‐trended and comparative surveillance data allowed these professionals to realize the magnitude of their local problems with nosocomial infections and exposures to bloodborne pathogens. Feedback of findings obtained by DICON’s analyses of surveillance data also probably helped personnel in participating hospitals to realize that their infection control problems were similar to those of other community hospitals.
Finally, we believe that the realization by community‐based ICPs and physicians that reductions in nosocomial infection rates were actually occurring in DICON‐affiliated hospitals during the study period was a powerful stimulus to the morale and resolve of these local hospital personnel. Indeed, it is possible that the most important motivator for continued and improved adherence to infection control policies and procedures was the realization that rates of nosocomial infection were falling. Unfortunately, we do not have measurements of compliance with existing infection control measures in participating hospitals during the study period, but further studies examining the impact of surveillance data on the behavior of medical personnel could test this hypothesis.
References
- 1. Gaynes R, Richards C, Edwards J, et al. Feeding back surveillance data to prevent hospital‐acquired infections. Emerg Infect Dis 2001; 7:295‐298.
- 2. Richards C, Emori TG, Edwards J, Fridkin S, Tolson J, Gaynes R. Characteristics of hospitals and infection control professionals participating in the National Nosocomial Infections Surveillance System 1999. Am J Infect Control 2001; 29:400‐403.
- 3. Centers for Disease Control and Prevention (CDC). Monitoring hospital‐acquired infections to promote patient safety—United States, 1990‐1999. MMWR Morb Mortal Wkly Rep 2000; 49:149‐153.
- 4. Engemann JJ, Fulmer E, Clark CC, Sexton DJ, Kaye KS. The impact of resistance on infection control units in community hospitals in the Duke Infection Control Outreach Network (DICON). Paper presented at 12th Annual Scientific Meeting of the Society for Healthcare Epidemiology of America; April 6‐9, 2002; Salt Lake City, Utah; abstract 9.
- 5. Britt MR, Burke JP, Nordquist AG, Wilfert JN, Smith CB. Infection control in small hospitals: prevalence surveys in 18 institutions. JAMA 1976; 236:1700‐1703.
- 6. Boyce JM. Hospital epidemiology in smaller hospitals. Infect Control Hosp Epidemiol 1995; 16:600‐606.
- 7. de Oliveira TC, Branchini ML. Infection control in a Brazilian regional multihospital system. Am J Infect Control 1999; 27:262‐269.
- 8. Horan TC, Emori TG. Definitions of key terms used in the NNIS System. Am J Infect Control 1997; 25:112‐116.
- 9. National Nosocomial Infections Surveillance System. National Nosocomial Infections Surveillance (NNIS) System Report, data summary from January 1992 to June 2002, issued August 2002. Am J Infect Control 2002; 30:458‐475.
- 10. Kleinbaum DG, Kupper, L.L., Morgenstern, H. Epidemiologic Research. Principles and Quantitative Methods. London: Lifetime Learning Publications; 1982.
- 11. Maki DG, Stolz SM, Wheeler S, Mermel LA. Prevention of central venous catheter‐related bloodstream infection by use of an antiseptic‐impregnated catheter: a randomized, controlled trial. Ann Intern Med 1997; 127:257‐266.
- 12. Orsi GB, Di Stefano L, Noah N. Hospital‐acquired, laboratory‐confirmed bloodstream infection: increased hospital stay and direct costs. Infect Control Hosp Epidemiol 2002; 23:190‐197.
- 13. Pittet D, Tarara D, Wenzel RP. Nosocomial bloodstream infection in critically ill patients: excess length of stay, extra costs, and attributable mortality. JAMA 1994; 271:1598‐1601.
- 14. Pittet D, Harbarth S. What techniques for diagnosis of ventilator‐associated pneumonia? Lancet 1998; 352:83‐84.
- 15. Shorr AF, O’Malley PG. Continuous subglottic suctioning for the prevention of ventilator‐associated pneumonia: potential economic implications. Chest 2001; 119:228‐235.
- 16. Zack JE, Garrison T, Trovillion E, et al. Effect of an education program aimed at reducing the occurrence of ventilator‐associated pneumonia. Crit Care Med 2002; 30:2407‐2412.
- 17. Warren DK, Shukla SJ, Olsen MA, et al. Outcome and attributable cost of ventilator‐associated pneumonia among intensive care unit patients in a suburban medical center. Crit Care Med 2003; 31:1312‐1317.
- 18. OANDA. Foreign exchange services and trading. Available at: http://www.oanda.com/. Accessed February 21, 2006.
- 19. Kirkland KB, Briggs JP, Trivette SL, Wilkinson WE, Sexton DJ. The impact of surgical‐site infections in the 1990s: attributable mortality, excess length of hospitalization, and extra costs. Infect Control Hosp Epidemiol 1999; 20:725‐730.
- 20. Whitehouse JD, Friedman ND, Kirkland KB, Richardson WJ, Sexton DJ. The impact of surgical‐site infections following orthopedic surgery at a community hospital and a university hospital: adverse quality of life, excess length of stay, and extra cost. Infect Control Hosp Epidemiol 2002; 23:183‐189.
- 21. Haley RW, Tenney JH, Lindsey JO 2nd, Garner JS, Bennett JV. How frequent are outbreaks of nosocomial infection in community hospitals? Infect Control 1985; 6:233‐236.
- 22. Haley RW, Culver DH, White JW, et al. The efficacy of infection surveillance and control programs in preventing nosocomial infections in US hospitals. Am J Epidemiol 1985; 121:182‐205.