Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Arch Pediatr Adolesc Med. Author manuscript; available in PMC 2007 November 18.
Published in final edited form as:
PMCID: PMC2080680

Nursing resources

a major determinant of nosocomial infection?


Purpose of review

There is growing concern that changes in nurse workforce and hospital-restructuring interventions negatively impact on patient outcomes. This review focuses on the association between understaffing and health-care-associated infections.

Recent findings

There is a large number of studies showing that overcrowding, understaffing or a misbalance between workload and resources are important determinants of nosocomial infections and cross-transmission of microorganisms. Importantly, not only the number of staff but also the level of their training affects outcomes. The nurse workforce is ageing, mainly due to fewer individuals’ engaging in a nursing career. This phenomenon, combined with cost-driven downsizing, contributes to a nursing shortage, and this tendency is not expected to revert unless important system changes are implemented. The causal pathway between understaffing and infection is complex, and factors might include lack of time to comply with infection control recommendations, job dissatisfaction, job-related burnout, absenteeism and a high staff turnover.


The evidence that cost-driven downsizing and changes in staffing patterns causes harm to patients cannot be ignored, and should not be considered as an inevitable outcome. More research is needed to better define the optimal patient-to-nurse ratio in various hospital settings and to estimate the economical impact of the nursing shortage. All quality-improvement interventions should carefully take into account systems and processes to be successful, as the issue of staffing is essentially a structural problem.

Keywords: cross-transmission, nosocomial infection, nurse, understaffing, workload


Despite advances in patient care, nosocomial infections continue to be a serious threat, causing adverse outcomes and increased costs [1-7]. Several individual risk factors, such as exposure to invasive devices, have been identified and successfully targeted by preventive strategies [8,9], but relatively little is known about the effect of organizational and institutional factors on the risk of nosocomial infection. Since the early 1990s and the rise of managed care in the USA [10], health policy makers and administrators have tried harder than ever to rationalize the health-delivery system to decrease health-care costs, while at the same time trying to retain high-quality services. In addition, the main workforce of the health system, nurses, has shrunk over the past decade. There is growing concern that these two phenomena in combination will seriously jeopardize the quality of care and patient safety.

Numerous studies have been performed to investigate the association between staffing issues or increased workload and health-care-associated adverse events, such as mortality [11•,12-14], patient length of stay, costs [14], postoperative complications [14-17], medication errors [18], decubitus ulcers [18,19], delayed weaning from mechanical ventilation [20] or needlestick injuries [21]. In the present review, we will focus on nosocomial infection and cross-transmission of microorganisms, with special attention to absolute nurse shortage, nurse skill mix and the threshold above which infection risk increases.

Staffing and nosocomial infections

Several multicentre studies using large databases [18,19,22-24,25•,26] showed a consistent effect of the number and composition of the workforce on the incidence of nosocomial infections.

The study by Needleman et al. [24] used administrative data from 799 US hospitals in 11 states to investigate the association between several patient outcomes and nurse staffing. Outcomes sensitive to nurse staffing among medical patients were urinary tract infection, pneumonia, length of hospital stay and upper gastrointestinal bleeding. The authors built several models with various assumptions to estimate the impact of modification in nurse staffing on the rate of specific patient outcomes. Urinary-tract infection was estimated to be reduced by 9% when the nurse staffing was shifted towards higher numbers of registered nurses. Similarly, a higher number of registered nurses was expected to reduce the rate of pneumonia by 6%. These findings are supported by the study by Cho et al. [26], which showed that an increase of 1 h worked by registered nurses per patient day or a 10% increase in the proportion of registered nurses were associated with, respectively, 8.9 and 9.5% decreases in the chance of pneumonia. By knowing or estimating the attributable cost of an infection, this type of information gives some insight into economical issues related to staff shortage and interventions to reverse it.

In a prospective multicentre study involving over 4000 critically ill patients, Alonso-Echanove et al. [25•] documented the infection risk associated with nurse skill mix. Among patients not receiving total parenteral nutrition there was an increased risk (by over twofold) of catheter-associated bloodstream infection in patients cared for by ‘float’ nurses [nurses not assigned to the intensive care unit (ICU)] for more than 60% of the duration of the central vascular access. Of note, the authors did not find an association between the nurse-to-patient ratio and the risk of catheter-associated blood-stream infection. Compared to others, this study is remarkable by its prospective design and its quality in the risk adjustment. The importance of the qualification and level of training of nurses has been highlighted in other studies [11•,27,28]. By use of a case-control study performed in a surgical ICU in an endemic setting, Robert et al. [28] found that patients who experienced primary bloodstream infection were more likely to have been hospitalized at a time when the unit was staffed with fewer regular ICU nurses and more pool nurses (i.e. those not experienced in critical care).

In a large ecological study including 211 hospitals per year over 7 years, Unruh [19] related staffing patterns to six adverse events (atelectasis, decubitus ulcers, falls, pneumonia, posttreatment infections and urinary-tract infections). The number of licensed nurses was inversely correlated with the risk of urinary tract infection, and the proportion of licensed nurses with the risk of pneumonia. One important limitation of that study was the poor adjustment for individual risk factors; only a few hospital characteristics (such as ownership status, number of certified doctors and capacity utilization) were accounted for in the analysis.

Fridkin et al. [29] performed, as part of an outbreak investigation, a retrospective case-control and a cohort study to identify risk factors for central venous catheter-associated bloodstream infections in a surgical ICU. The investigators found that several surrogate markers of workload were unfavourable during the outbreak period as compared to the pre-outbreak period, and that high workload was an independent risk factor for infection. Interestingly, there was a temporal association between the number of infections and the patient-to-nurse ratio, with a threshold above which the number of infections was above the endemic level. Thus there may be a critical staffing threshold below which optimal patient care becomes difficult, causing suboptimal device management and increased infection rates.

Studying the association between staffing and nosocomial infection is not straightforward, as it requires a complex risk adjustment and standardized and valid case definitions. Moreover, the causal pathway between staff shortage and infection is not easy to grasp, due to the complexity of the infection process itself. Cross-transmission of microorganisms known to be transmitted mainly by health-care workers’ hands, such as methicillin-resistant Staphylococcus aureus (MRSA), is another issue: each transmission event is a failure of basic infection-control measures, such as hand hygiene, regardless of the patient case mix or severity of illness. Consequently, cross-transmission rate might be sensitive to compliance to infection control measures and to workload. Vicca [30] offered an excellent illustration of the contribution of understaffing to the spread of MRSA, as a surrogate marker for hand-hygiene compliance. In this study, the number of incident MRSA cases was closely associated in time with several markers of workload, and there was a weak but significant correlation between the MRSA attack rate and varying staffing levels. Recently, Andersen et al. [31] reported another outbreak of MRSA in a neonatal ICU, but provided only a descriptive epidemiological investigation. Before and during the outbreak there was high activity, overcrowding and a large proportion of relatively untrained nurses. The outbreak stopped after these and other infection control measures had been improved. A recently published study investigated the transmission of MRSA in a British ICU [32]. The first part of this well conducted investigation consisted of a risk factor analysis based on prospectively collected data. A relative staff deficit was an independent risk factor for belonging to a cluster of MRSA, whereas it was not for sporadic cases. The authors then fitted a stochastic model to their data to investigate the effectiveness of infection control measures, and found that improved compliance with hand hygiene would be effective to reduce the net reproductive number, i.e. the number of transmission events. Several other studies linking overcrowding, understaffing or nursing workload with cross-transmission of MRSA [33-35], extended-spectrum β-lactamase-producing Enterobacteriaceae [36,37], Klebsiella pneumoniae [38], Enterobacter cloacae [39] or gastrointestinal virus [40] have been published. Data from an outbreak investigation performed in our neonatal ICU showed the intermediate step between high workload and nosocomial infection: noncompliance with hand-hygiene recommendations [39]. We experienced an outbreak of E. cloacae involving eight patients over a 6-week period. During the outbreak period, the number of patients increased, the mean occupancy rate exceeded the standard by 50%, whereas staff on duty remained constant, and far below what was required. An observational survey of hand-hygiene practices performed in the middle of the outbreak showed that compliance with hand-hygiene recommendations was 38%. The outbreak was resolved after workload decreased and compliance with hand disinfection increased.

Nurse workforce and demographics

Many Western countries, if not all, are facing the problem of a nursing shortage, and this phenomenon is not expected to reverse unless deep structural changes are implemented. Reasons for this are numerous, correlated to each other, and include the following determinants. Ageing of the nurse workforce has been well documented and is mainly due to the decreasing number of individuals choosing a nursing career. It is expected that by 2020 the nursing workforce will be 20% below the projected requirement in the United States [41]. The situation is very similar in Canada, as the predicted shortage of nurses by 2011 will range between 59 000 and 113 000 [42], and in European countries, where there is presently a shortage of 30 000 health-care workers in the UK, 25 000 in France and 6000 in Switzerland. Added to that, most hospital systems have implemented restructuring interventions aiming to rationalize care delivery under the pressure of cost containment, resulting in changes in the nurse workforce and nursing skills [43]. One unsatisfactory option is to partially fill the gap with less-trained professionals - ‘float’ or ‘pool’ nurses - and the risks associated with that practice are presented above. We have reviewed the consequences of a nursing shortage on patients’ outcome; Aiken et al. [44] provide more insight into this complex issue by investigating the effect of several organizational items on nurses themselves. In a large international study [44], the percentage of nurses with a burnout score above the norm for medical personnel and the percentage of nurses dissatisfied with their job were extremely high - around 40%. Moreover, inadequate staffing and insufficient organizational support were both associated with job dissatisfaction and burnout. In another study, Aiken et al. [45] showed that each additional patient per nurse was associated with a 23% increase in burnout and 15% increase in job dissatisfaction among nurses and a 7% increase in patient mortality [45]. Job dissatisfaction, intention to leave the current position and turnover are related; job dissatisfaction and burnout might lead to absenteeism, and absenteeism negatively impacts on patient outcome [46]. And the loop is closed. To revert this will require addressing simultaneously: (1) the declining number of new nurses, (2) attracting nurses to stay in hospitals and (3) nurses leaving the workforce early [47]. It is beyond the scope of this review to discuss possible strategies, which are outlined elsewhere [41,47-49].

Limitations of the studies

Most of the studies, if not all, aiming to assess the association between understaffing and preventable adverse events in hospitals suffer from a variety of limitations, some of which may be unavoidable. Large multicentre studies use several administrative databases, often use multiple data sets that have to be standardized, are performed retrospectively and rely heavily on available data, such as discharge data. Consequently errors may have been introduced at each step of the data management, adverse events are very likely to have been underreported and adjustment for confounding factors is often minimal. Moreover, these studies share the problems of all studies with an ecological design, in the sense that the focus is to detect variations in outcome rates for hospitals rather than detecting individual cases of complications. Even if the outcome (nosocomial infection) is assessed on an individual basis, the risk factor under investigation (patient-to-nurse ratio) is by definition ecological in nature, as it applies to all patients at a point in time. Studies performed in single centres or units will be based on more homogeneous data but might be endangered by small sample sizes. The independency assumption is often violated when performing risk-factor analysis. Indeed, the occurrence of an infection in a patient is not necessarily independent from an infection in another patient; nor is a second infection in a patient from the first, nor the infection rate in a unit from that in another unit in the same hospital: this clustering should be accounted for in the analysis. Finally, many of these studies were performed in the context of outbreak investigations, which of note represent only 10% of all nosocomial infections. How these findings can be generalized to endemic infections remains unclear.

Is there enough evidence?

This question could be rephrased as ‘Why should we not be satisfied with all these data?’ Some might wait for the perfect study that will never come; they should rather accept the risk of imprecise risk assessment; this is life and epidemiology. Randomized controlled trials are not always the gold standard and should sometimes be replaced, for the good of the study participants, by common sense [50]. It might be helpful at this point to interpret these data with respect to Bradford-Hill criteria for causality: findings are consistent between studies, there are a large number of studies performed in various settings and using different study designs and outcomes, there is a temporal association between ‘exposure’ and ‘outcome’, the phenomenon is reversible and the association is plausible. The question of the causal pathway, although intuitively quite obvious, is not that straightforward, and we believe that the issue of staff shortage should be considered as a determinant of an adverse event, rather than strictly speaking a risk factor. As mentioned, one study documented a link between understaffing and cross-transmission: noncompliance with hand-hygiene recommendations [39]. This is the final step leading to cross-transmission, but it would be simplistic to view it in a mechanical and linear way: understaffing leads to increased workload, which leads to no time to perform adequate hand hygiene. There are surely more complex issues than just ‘time’, difficult to disentangle, such as job dissatisfaction, burnout, intention to leave and the institutional safety climate.

Future research and action

Numerous ways of assessing workload or misbalance between workload and resources have been used, all suffering from their ecological nature. More refined workload measurements are needed that would ultimately help to estimate what is adequate staffing according to various hospital settings. The estimation of threshold above which the infection risk increases would be of paramount importance for administrators and policy makers. Formal economical evaluation is needed, which will probably demonstrate that the benefit resulting from cutting down on nursing resources is by far outweighed by the cost of nosocomial infections attributed to staff shortages. Increasing recruitment of nursing students and keeping them in the job should be priorities. Any intervention bypassing principles of quality improvement and assessment as proposed by Donabedian [51] are doomed to fail, as the issue of staffing is deeply rooted in systems and processes. In this regard, more studies such as the one by Braun et al. [52•] are needed to better target modifiable determinants of nosocomial infections.


In summary, the association between understaffing, overcrowding and nosocomial infection is not a linear cause-effect relationship, but consists of the interaction of several factors exhibiting synergistic effects, and may be flawed by various methodological shortcomings, including publication bias. Nevertheless, whatever the limitations of previous studies, the evidence that cost-driven downsizing and changes in staffing patterns cause harm to patients cannot be ignored. Considering the tremendous number of deaths, infections or other adverse health-care-associated events due to under-staffing, this situation should not be considered as inevitable, and changes should be brought by acting on the system. Finally, it surely appears self-evident to state that we could easily do better with more, but we definitely can do better with what we have, the list of successful interventions to prevent health-care-associated infections being too long to mention.


We are indebted to Rosemary Sudan for providing editorial assistance. Stéphane Hugonnet was funded by a research grant from the Swiss National Science Foundation (grant number 32-68164.02).


intensive care unit
methicillin-resistant Staphylococcus aureus

References and recommended reading

Papers of particular interest, published within the annual period of review, have been highlighted as:

• of special interest

•• of outstanding interest

1. Haley RW. Measuring the costs of nosocomial infections: methods for estimating economic burden on the hospital. Am J Med. 1991;91:32S–38S. [PubMed]
2. Haley RW, Schaberg DR, Crossley KB, et al. Extra charges and prolongation of stay attributable to nosocomial infections: a prospective interhospital comparison. Am J Med. 1981;70:51–58. [PubMed]
3. 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. [PubMed]
4. Jarvis WR. Selected aspects of the socioeconomic impact of nosocomial infections: morbidity, mortality; cost, and prevention. Infect Control Hosp Epidemiol. 1996;17:552–557. [PubMed]
5. Fagon JY, Chastre J, Hance AJ, et al. Nosocomial pneumonia in ventilated patients: a cohort study evaluating attributable mortality and hospital stay. Am J Med. 1993;94:281–288. [PubMed]
6. Fagon JY, Novara A, Stephan F, et al. Mortality attributable to nosocomial infections in the ICU. Infect Control Hosp Epidemiol. 1994;15:428–434. [PubMed]
7. Sax H, Pittet D. Interhospital differences in nosocomial infection rates: importance of case-mix adjustment. Arch Intern Med. 2002;162:2437–2442. [PubMed]
8. Haley RW. The scientific basis for using surveillance and risk factor data to reduce nosocomial infection rates. J Hosp Infect. 1995;30(Suppl):3–14. [PubMed]
9. Eggimann P, Harbarth S, Constantin MN, et al. Impact of a prevention strategy targeted at vascular-access care on incidence of infections acquired in intensive care. Lancet. 2000;355:1864–1868. [PubMed]
10. Oberlander J. The US health care system: on a road to nowhere? CMAJ. 2002;167:163–168. [PMC free article] [PubMed]
11•. Aiken LH, Clarke SP, Cheung RB, et al. Educational levels of hospital nurses and surgical patient mortality. JAMA. 2003;290:1617–1623. [PMC free article] [PubMed]A well conducted study showing the association between staffing and mortality in surgical patients
12. Lapichino G, Gattinoni L, Radrizzani D, et al. Volume of activity and occupancy rate in intensive care units: association with mortality. Intensive Care Med. 2004;30:290–297. [PubMed]
13. Tarnow-Mordi WO, Hau C, Warden A, Shearer AJ. Hospital mortality in relation to staff workload: a 4-year study in an adult intensive-care unit. Lancet. 2000;356:185–189. [PubMed]
14. Dimick JB, Pronovost PJ, Heitmiller RF, Lipsett PA. Intensive care unit physician staffing is associated with decreased length of stay, hospital cost, and complications after esophageal resection. Crit Care Med. 2001;29:753–758. [PubMed]
15. Pronovost PJ, Jenckes MW, Dorman T, et al. Organizational characteristics of intensive care units related to outcomes of abdominal aortic surgery. JAMA. 1999;281:1310–1317. [PubMed]
16. Amaravadi RK, Dimick JB, Pronovost PJ, Lipsett PA. ICU nurse-to-patient ratio is associated with complications and resource use after esophagectomy. Intensive Care Med. 2000;26:1857–1862. [PubMed]
17. Dimick JB, Swoboda SM, Pronovost PJ, Lipsett PA. Effect of nurse-to-patient ratio in the intensive care unit on pulmonary complications and resource use after hepatectomy. Am J Crit Care. 2001;10:376–382. [PubMed]
18. Blegen MA, Goode CJ, Reed L. Nurse staffing and patient outcomes. Nurs Res. 1998;47:43–50. [PubMed]
19. Unruh L. Licensed nurse staffing and adverse events in hospitals. Med Care. 2003;41:142–152. [PubMed]
20. Thorens JB, Kaelin RM, Jolliet P, Chevrolet JC. Influence of the quality of nursing on the duration of weaning from mechanical ventilation in patients with chronic obstructive pulmonary disease. Crit Care Med. 1995;23:1807–1815. [PubMed]
21. Clarke SP, Rockett JL, Sloane DM, Aiken LH. Organizational climate, staffing, and safety equipment as predictors of needlestick injuries and near-misses in hospital nurses. Am J Infect Control. 2002;30:207–216. [PubMed]
22. Kovner C, Gergen PJ. Nurse staffing levels and adverse events following surgery in U.S. hospitals. Image J Nurs Scholarsh. 1998;30:315–321. [PubMed]
23. Lichtig LK, Knauf RA, Milholland DK. Some impacts of nursing on acute care hospital outcomes. J Nurs Adm. 1999;29:25–33. [PubMed]
24. Needleman J, Buerhaus P, Mattke S, et al. Nurse-staffing levels and the quality of care in hospitals. N Engl J Med. 2002;346:1715–1722. [PubMed]
25•. Alonso-Echanove J, Edwards JR, Richards MJ, et al. Effect of nurse staffing and antimicrobial-impregnated central venous catheters on the risk for bloodstream infections in intensive care units. Infect Control Hosp Epidemiol. 2003;24:916–925. [PubMed]Interesting prospective study placing emphasis on risk adjustment
26. Cho SH, Ketefian S, Barkauskas VH, Smith DG. The effects of nurse staffing on adverse events, morbidity, mortality, and medical costs. Nurs Res. 2003;52:71–79. [PubMed]
27. Yang KP. Relationships between nurse staffing and patient outcomes. J Nurs Res. 2003;11:149–158. [PubMed]
28. Robert J, Fridkin SK, Blumberg HM, et al. The influence of the composition of the nursing staff on primary bloodstream infection rates in a surgical intensive care unit. Infect Control Hosp Epidemiol. 2000;21:12–17. [PubMed]
29. Fridkin SK, Pear SM, Williamson T, et al. The role of understaffing in central venous catheter-associated bloodstream infections. Infect Control Hosp Epidemiol. 1996;17:150–158. [PubMed]
30. Vicca AF. Nursing staff workload as a determinant of methicillin-resistant Staphylococcus aureus spread in an adult intensive therapy unit. J Hosp Infect. 1999;43:109–113. [PubMed]
31. Andersen BM, Lindemann R, Bergh K, et al. Spread of methicillin-resistant Staphylococcus aureus in a neonatal intensive unit associated with understaffing, overcrowding and mixing of patients. J Hosp Infect. 2002;50:18–24. [PubMed]
32. Grundmann H, Hori S, Winter B, et al. Risk factors for the transmission of methicillin-resistant Staphylococcus aureus in an adult intensive care unit: fitting a model to the data. J Infect Dis. 2002;185:481–488. [PubMed]
33. Haley RW, Cushion NB, Tenover FC, et al. Eradication of endemic methicillin-resistant Staphylococcus aureus infections from a neonatal intensive care unit. J Infect Dis. 1995;171:614–624. [PubMed]
34. Haley RW, Bregman D. The role of understaffing and overcrowding in recurrent outbreaks of staphylococcal infection in a neonatal special-care unit. J Infect Dis. 1982;145:875–885. [PubMed]
35. Arnow PM, Allyn PA, Nichols EM, et al. Control of methicillin-resistant Staphylococcus aureus in a burn unit: role of nurse staffing. J Trauma. 1982;22:954–959. [PubMed]
36. Pessoa-Silva CL, Toscano CM, Moreira BM, et al. Infection due to extended-spectrum beta-lactamase-producing Salmonella enterica subsp. enterica serotype infantis in a neonatal unit. J Pediatr. 2002;141:381–387. [PubMed]
37. Soulier A, Barbut F, Ollivier JM, et al. Decreased transmission of Enterobacteriaceae with extended-spectrum beta-lactamases in an intensive care unit by nursing reorganization. J Hosp Infect. 1995;31:89–97. [PubMed]
38. Denny F, St John MA, Lewis DB, Daniel B. Nosocomial Klebsiella pneumoniae colonization in a neonatal special care unit. Ann Trop Paediatr. 1986;6:123–128. [PubMed]
39. Harbarth S, Sudre P, Dharan S, et al. Outbreak of Enterobacter cloacae related to understaffing, overcrowding and poor hygiene practices. Infect Control Hosp Epidemiol. 1999;20:598–603. [PubMed]
40. Stegenga J, Bell E, Matlow A. The role of nurse understaffing in nosocomial viral gastrointestinal infections on a general pediatrics ward. Infect Control Hosp Epidemiol. 2002;23:133–136. [PubMed]
41. Buerhaus PI, Staiger DO, Auerbach DI. Implications of an aging registered nurse workforce. JAMA. 2000;283:2948–2954. [PubMed]
42. Spurgeon D. Canada faces nurse shortage. BMJ. 2000;320:1030. [PMC free article] [PubMed]
43. Sochalski J, Aiken LH, Fagin CM. Hospital restructuring in the United States, Canada, and Western Europe: an outcomes research agenda. Med Care. 1997;35:OS13–OS25. [PubMed]
44. Aiken LH, Clarke SP, Sloane DM. Hospital staffing, organization, and quality of care: cross-national findings. Int J Qual Health Care. 2002;14:5–13. [PubMed]
45. Aiken LH, Clarke SP, Sloane DM, et al. Hospital nurse staffing and patient mortality, nurse burnout, and job dissatisfaction. JAMA. 2002;288:1987–1993. [PubMed]
46. Taunton RL, Kleinbeck SV, Stafford R, et al. Patient outcomes: are they linked to registered nurse absenteeism, separation, or work load? J Nurs Adm. 1994;24:48–55. [PubMed]
47. Berliner HS, Ginzberg E. Why this hospital nursing shortage is different. JAMA. 2002;288:2742–2744. [PubMed]
48. Jackson M, Chiarello LA, Gaynes RP, Gerberding JL. Nurse staffing and health care-associated infections: proceedings from a working group meeting. Am J Infect Control. 2002;30:199–206. [PubMed]
49. Needleman J, Buerhaus P. Nurse staffing and patient safety: current knowledge and implications for action. Int J Qual Health Care. 2003;15:275–277. [PubMed]
50. Smith GC, Pell JP. Parachute use to prevent death and major trauma related to gravitational challenge: systematic review of randomised controlled trials. BMJ. 2003;327:1459–1461. [PMC free article] [PubMed]
51. Donabedian A. Methods for deriving criteria for assessing the quality of medical care. Med Care Rev. 1980;37:653–698. [PubMed]
52•. Braun BI, Kritchevsky SB, Wong ES, et al. Preventing central venous catheter-associated primary bloodstream infections: characteristics of practices among hospitals participating in the Evaluation of Processes and Indicators in Infection Control (EPIC) study. Infect Control Hosp Epidemiol. 2003;24:926–935. [PubMed]Good example of a study highlighting modifiable system and process risk factors