infection control — a problem for patient safety.pdf

n engl j med 348;7 www.nejm.org february 13, 2003

The new england journal of medicine

651

health policy report

Infection Control — A Problem for Patient SafetyJohn P. Burke, M.D.

Nosocomial, or hospital-acquired, infections (moreappropriately called health care–associated infec-tions) are today by far the most common complica-tions affecting hospitalized patients. Indeed, theHarvard Medical Practice Study II found that a sin-gle type of nosocomial infection — surgical-woundinfection — constituted the second-largest catego-ry of adverse events.1 Long considered the greatestrisk that the hospital environment poses to pa-tients,2 nosocomial infections abruptly became theprovince of public health officers at the time of anationwide epidemic of hospital-based staphylo-coccal infections, in 1957 and 1958.3 Since then, thestudy and control of nosocomial infections havebeen profoundly shaped by the discipline of pub-lic health, with its emphasis on surveillance andepidemiologic methods. These infections are notonly the most common types of adverse events inhealth care; they may also be the most studied.

Currently, between 5 and 10 percent of patientsadmitted to acute care hospitals acquire one or moreinfections, and the risks have steadily increased dur-ing recent decades (Table 1).4,5 These adverse eventsaffect approximately 2 million patients each year inthe United States, result in some 90,000 deaths, and

add an estimated $4.5 to $5.7 billion per year to thecosts of patient care.6,7 Infection control is thereforea critical component of patient safety. In this articleI describe the common ground shared by these twodisciplines. I also discuss the major problems in in-fection control, approaches to their solutions, therole of the National Nosocomial Infections Surveil-lance (NNIS) System of the Centers for Disease Con-trol and Prevention (CDC) as a model, and the needfor renewed commitment to and innovations in in-fection control to help ensure patient safety.

Four types of infection account for more than 80percent of all nosocomial infections: urinary tractinfection (usually catheter-associated), surgical-siteinfection, bloodstream infection (usually associat-ed with the use of an intravascular device), and pneu-monia (usually ventilator-associated) (Fig. 1).8,9

One fourth of nosocomial infections involve pa-tients in intensive care units, and nearly 70 percentare due to microorganisms that are resistant to oneor more antibiotics — an emerging public healthcrisis that is due in large part to indiscriminate useof antibiotics.10

Nosocomial infections can also be ranked accord-ing to their frequencies, associated mortality rates,costs, and relative changes in frequency in recentyears.4,7 Catheter-associated urinary tract infectionsare the most frequent (accounting for about 35 per-cent of nosocomial infections) but carry the lowestmortality and lowest cost. Surgical-site infectionsare second in frequency (about 20 percent) and thirdin cost. Bloodstream infections and pneumonia areless common (about 15 percent each) but are asso-ciated with much higher mortality and costs. Blood-stream infections and methicillin-resistant Staphy-lococcus aureus infections share notoriety for beingboth the highest-cost infections and the most rap-idly increasing in frequency; the current incidenceof bloodstream infections is nearly three times theincidence in 1975.4,11 The rates of both urinary tract

the nature of nosocomialinfections

*Data are from Weinstein4 and Jarvis.5

Table 1. Nosocomial Infections in the United States.*

Variable Year

1975 1995

No. of admissions (¬10¡6) 37.7 35.9

No. of patient-days (¬10¡6) 299.0 190.0

Average length of stay (days) 7.9 5.3

No. of inpatient surgical proce-dures (¬10¡6)

18.3 13.3

No. of nosocomial infections (¬10¡6)

2.1 1.9

Incidence of nosocomial infections (no. per 1000 patient-days)

7.2 9.8

The New England Journal of Medicine Downloaded from nejm.org on June 22, 2012. For personal use only. No other uses without permission.

Copyright © 2003 Massachusetts Medical Society. All rights reserved.


652 n engl j med 348;7 www.nejm.org february 13, 2003

and surgical-site infections have declined slightly,perhaps because of surveillance artifacts caused bydecreases in the length of hospital stays and increas-ing numbers of infections that develop after dis-charge from the hospital.

Each of the main types of infection comprisesmore than one syndrome and has multiple patho-genetic pathways. For example, ventilator-associ-ated pneumonia, a cause of one fourth of the deathsattributed to nosocomial infections, commonly oc-curs as a result of infection with one or more bacte-rial species, but it may also occur with less commonpathogens, such as legionella, respiratory viruses,or Aspergillus fumigatus.12 For each of the device-asso-ciated infections, multiple risk factors are related tothe patient, the personnel caring for the patient, theprocedures they use, and the device itself.

Identification of risk factors permits elucidationof those that are alterable from those that are notand facilitates the development of targeted inter-

ventions to reduce the risk of infection. For exam-ple, avoiding the use of invasive devices altogetherby means of alternative strategies (for example, per-forming urinary drainage by condom catheter) andshortening the duration of use of the device (for ex-ample, reducing the number of days of mechanicalventilation) have been proposed in many guidelines.Strategies to prevent infections have been subdi-vided into several groups (education-based, proc-ess-based, and systems-based),13 but many of thesuggested interventions — such as “use antibioticswisely” or “educate and train staff”12 — have beenvague and difficult to implement.

Behavioral change remains a formidable ob-stacle. For example, cross-infection of patients byhealth care workers with contaminated hands is amajor source of infections. Despite educational ef-forts, health care workers, including physicians,continue to fail to adhere to standards for hand hy-giene, which is universally considered the singlemost important method for infection control. Theaverage level of compliance has varied among hos-pitals from 16 percent to 81 percent.14 Barriers tocompliance include understaffing and poor designof facilities, confusing and impractical guidelinesand policies, failure to apply behavioral-change the-ory fully, and insufficient commitment and enforce-ment by infection-control personnel.14,15 Remark-ably, the use of waterless antiseptic hand rubs, whenpart of a multifaceted campaign that encouragesappropriate hand washing, has been shown to bemore practical than standard hand washing aloneand has been shown to improve the adherence ofhealth care workers to hand-hygiene guidelines andto prevent the transmission of methicillin-resistantS. aureus to patients.16

The new Guideline for Hand Hygiene in Health-Care Settings, developed by a multidisciplinarytask force,17 may facilitate system improvementsby resolving many inconsistencies among previousguidelines from the CDC and other groups. It alsoincludes a requirement for monitoring adherenceto the guideline, along with suggested methods fordoing so. The guideline bans the use of artificialnails when providing patient care, defines the dif-ferent indications for hand washing as opposed todecontamination, and calls for the use of alcohol-based, waterless antiseptics for decontaminatingthe hands before and after any direct contact with apatient’s intact skin.

The history of infection control is littered withcommercial products and devices to prevent infec-

prevention of nosocomial infections

Figure 1. Number of Nosocomial Pathogens, According to Infection Site, Identified in the Hospital-Wide Component of the National Nosocomial Infec-tions Surveillance System from January 1990 to March 1996.

The hospital-wide component of the National Nosocomial Infections Surveil-lance System consists of a subgroup of hospitals reporting data on nosoco-mial infections from all patients. In January 1999, this component was eliminated from the system.

0

5,000

10,000

15,000

20,000

25,000

30,000

35,000

40,000

Nos

ocom

ial P

atho

gens

(no.

of i

sola

tes)

UrinaryTract

Infection

Surgical-Site

Infection

BloodstreamInfection

Pneumonia OtherSites





653

tion that were widely promoted after limited test-ing and have since been discredited.18 The devel-opment of safer devices (for example, needles withsafety features and antimicrobial-coated catheters)has produced incremental gains in infection con-trol, but devices constructed of biomaterials thatfully prevent infections remain a tantalizing pros-pect. Conversely, the actual infection-control bene-fits of many technological improvements that werenot designed primarily to prevent infections, suchas improvements in anesthesia equipment and prac-tice,19 are inestimable but probably great.

Because of the limitations of infection-controlmethods, the fundamentals of prevention have nec-essarily been grounded in epidemiology through thedevelopment of standard definitions and classifi-cations; surveillance and early reporting of infec-tions, with feedback to “those who need to know”(i.e., responsible authorities); evaluation of risk-based interventions; and production of evidence-based guidelines.20 This process has been guided bythe CDC, with the help of the American HospitalAssociation and the regulatory efforts of the JointCommission on Accreditation of Healthcare Or-ganizations.

Epidemiologic analysis, often by means of case–control studies, is a powerful tool for identifyingthe cause or source of nosocomial infections. Oneexample among hundreds is the recognition of ahospital outbreak in which 11 cases of neonatalsepsis over a period of four years were traced to asingle human carrier.21 Root-cause analysis of in-dividual cases would have been incapable of identi-fying the source of these or most other hospital in-fections. Another example is the recognition oferroneous handling of closed urinary-drainage sys-tems as a cause of catheter-associated urinary tractinfections.22 In this analysis, too, there was epide-miologic evidence of the importance of errors, eventhough most of the errors were not followed by in-fection. Moreover, voluntary reporting of frequentlyoccurring infections has been found to underesti-mate greatly the true rate of avoidable infections, be-cause most infections are considered unfortunate,inevitable consequences of medical procedures.

Active surveillance is necessary to identify alter-able risk factors (sometimes called process indica-tors). Various indicators for infection control havebeen extensively evaluated, for example, in the devel-opment of a collaborative project to monitor healthcare processes and outcomes.23 The growing im-portance of monitoring process indicators is a be-

lated but major trend in infection control, but onethat does not diminish the need for surveillance ofoutcomes. Without surveillance, we will not knowthe effect of our efforts to prevent infection. Twoexamples illustrate the value and limitations of proc-ess indicators and the need for continued surveil-lance: surgical-site infections and outbreaks inhospitals.

surgical-site infectionsMany quality-improvement projects have identi-fied errors in the administration of antibiotic pro-phylaxis before surgery as an independent risk fac-tor for some postoperative infections. Incorrecttiming of surgical prophylaxis is associated withincreases by a factor of two to six in the rates ofsurgical-site infection for operative procedures inwhich prophylaxis is generally recommended.24

Failure to administer the first dose of antibioticwithin the two-hour window before incision (toachieve adequate blood levels of the antibiotic dur-ing surgery) remains a common error, occurring,for example, in 27 to 54 percent of all selected op-erations in a 1996 New York State study.25 Effec-tive programs have recognized and addressed theroot causes of errors that result from faulty systemsof care.26 In most patients who receive inappropri-ate prophylaxis, however, infections do not devel-op, and therefore relatively stable (and seeminglylow) rates of surgical-site infections in an individualhospital can mask the problem and create compla-cency. Therefore, some limited monitoring of proc-ess indicators, such as timely prophylaxis, is nec-essary to detect system problems.

Improving the timing of antibiotic prophylaxisdoes not supersede other elements of infection con-trol. In several early studies, surveillance of surgi-cal-site infections with confidential feedback of therelevant data to surgeons was found to reduce therisk of infection.27 Regardless of the reasons forthese results, the reasons for surveillance are no lesspressing today and have additional justifications,with the use of ever more complex surgical proce-dures and with the development of most postop-erative infections after discharge from the hospi-tal. Voluntary reporting of wound infections bysurgeons has not worked, and effective surveillancerequires active identification of cases by trained per-sonnel and consideration of the use of automateddetection systems.28 The downsizing of many in-fection-control programs due to hospitals’ financialconstraints29 has further increased the need for





new types of surveillance and process indicators toidentify surgical-site infections.

outbreaks in hospitalsAt least 5 to 10 percent of infections occur in clus-ters, or outbreaks, that can be detected from care-ful review of surveillance information.30 Many out-breaks are recognized only by astute clinicians orlaboratory workers. Most, if not all, infections inoutbreaks can be construed as accidental injuries.Therefore, the detection, investigation, and controlof outbreaks are a critical issue in patient safety andrequire vigilance.

Though occasionally dramatic, outbreaks maybe insidious and may be protracted causes of sub-stantial morbidity and mortality.21,31 They occur inall health care settings and with all classes of infec-tious agents, especially antibiotic-resistant bacteria,and because of their sometimes widespread nature,often have a considerable effect on the public. Of114 health care–associated outbreaks investigatedby CDC personnel over a 10-year period, 6 were na-tional in scope and were traced to contaminatedproducts or devices.32 Contamination of commer-cially distributed products may be detected only byspontaneous reporting from infection-control unitsin hospitals.

Recently, data-mining tools have been appliedto detect previously unrecognized outbreaks.33

Molecular techniques have been used to show thatseemingly unrelated infections have been causedby interspecies transfer of genes encoding antibiot-ic resistance, suggesting that the true rate of cross-infection in hospital settings remains greatly un-derestimated.34 These data indicate that the role oflaboratory-based surveillance in public health islikely to increase.

The importance of the patient-safety movement inenergizing infection control is already manifest.Many infection-control units have broadened theiractivities in monitoring the use of antibiotics andin preventing adverse drug events due to antibiot-ics. (Antibiotic resistance may even be considereda special type of adverse drug event, one with soci-etal consequences.)

More than 25 years ago, the Department of Clin-ical Epidemiology and Infectious Diseases of theLDS Hospital, in Salt Lake City, devised “clinicaltriggers” to facilitate the detection and surveillance

of infections and to improve the use of isolation andbarrier precautions for infection control.35 Alsocalled “signals” or “alerts,” clinical triggers are ele-ments drawn from patients’ electronic medical rec-ords by means of programmed logic or algorithmsthat suggest ongoing or potential adverse events,including infections. Continuous, real-time scan-ning of laboratory and pharmacy records, for ex-ample, facilitates cost-effective surveillance and ac-tive interventions to prevent or ameliorate adverseevents. The LDS Hospital team monitored drugdoses, renal function, the prescription of commonantidotes, and other triggers to track and preventadverse drug events.36 Interventions by a clinicalpharmacist reduced the use and misuse of antibiot-ics and showed that the potential to stabilize anti-biotic resistance existed.37 Voluntary reporting ofmedication errors had little overlap with adversedrug events detected by this method. These conceptsare now being widely adopted by hospitals acrossthe country through collaborative efforts coordi-nated by the Institute for Healthcare Improvement.

Recently, the Agency for Healthcare Researchand Quality released a controversial report that re-viewed the evidence in favor of 79 patient-safetypractices, of which 22 (28 percent) involved infec-tion control.38 Further illustrating the commonground shared by these two disciplines, 5 of the 11practices that were judged worthy of widespreadimplementation involved infection control. Two ofthese five practices — the appropriate use of antibi-otic prophylaxis in surgical patients and the use ofmaximal sterile barriers during the placement ofcentral venous catheters — were readily accepted.Curiously, the Agency for Healthcare Research andQuality reported that there was weaker evidencesupporting methods to improve adherence to handhygiene and limitations in antibiotic use — practic-es that some infection-control experts believe offerthe greatest potential benefit. These and other in-fection-control practices were listed as prioritiesfor further research.

The NNIS System of the CDC is a voluntary,confidential, hospital-based reporting system thathas been influential in guiding infection-controlefforts in hospitals across the United States andaround the world; it is the only national source ofsystematically gathered data on hospital infections.

the patient-safety movement

is the nnis system a modelfor infection-control programs?





655

Monthly reports of nosocomial infections frommore than 300 hospitals (a nonrandom sample ofU.S. hospitals, all with at least 100 beds and nearly60 percent academic medical centers) have allowedbenchmarks for infection rates to be establishedthrough the use of standardized case definitionsand data-collection methods and computerizeddata entry and analysis.30 Analysis of NNIS Systemdata helps reveal changes in patterns of incidence,distribution, antibiotic resistance, sites of infection,outcomes, and risk factors for infection. In March2000, the NNIS System reported that during the1990s, the rates of infection for respiratory tract,urinary tract, and bloodstream sites, after adjust-ment for the duration of the use of invasive devices,had decreased in intensive care units in selected hos-pitals.39 The multiple reasons for these reductions,however, cannot be attributed to any specific inter-ventions, nor does the report mean that all hospi-tals providing data to the NNIS System obtainedthese salutary results, since only a subgroup of hos-pitals participated.

The NNIS System is viewed as a benchmark onthe basis of the reasonable expectation that the par-ticipating infection-control programs possess thecomponents for effectiveness identified by the CDCin previous studies: intense surveillance, intensecontrol measures, and an adequate number of in-fection-control professionals. Though not fully ad-justed for patient risk factors, the rates of endemicinfections in participating hospitals have been usedto help drive improvement efforts. A few successstories have been reported from selected hospitalsin which problems (such as excessive use of certaininvasive devices and deviations from national guide-lines) were identified and addressed, but evaluationis still incomplete.

Each hospital participating in the NNIS Sys-tem provides data on only one or two high-risk com-ponents of surveillance, such as intensive care orselected surgical procedures. In addition, case ascer-tainment is time-consuming and costly for hospi-tals, and the definitions for infections are complexand difficult to apply. Therefore, the NNIS System isa model for focused surveillance but not for overallinfection control. This system has not yet addressedmany important safety issues, such as clinical errorsof omission leading to failures to diagnose infec-tion or delays in the diagnosis of infection. Further-more, the definitions used during surveillance (forexample, the definitions for ventilator-associated

pneumonia and for infections developing after hos-pital discharge) are a work in progress.

Perhaps the most important outcome of theNNIS System is the infrastructure of trained infec-tion-control professionals that it has nurtured andthe cadre of CDC-trained infectious-disease physi-cians who have migrated to university and commu-nity hospitals during the past 30 years. These hu-man resources are now endangered because of theeconomic forces shaping health care and the down-sizing of many, if not most, infection-control unitsin hospitals. The voluntary nature of NNIS may bean important factor in its success, but participationalso helps hospitals meet regulatory requirements.In addition, the support of CDC epidemiologists isa vital asset. More than a decade ago, the Instituteof Medicine called for further development of theNNIS System and its expansion to include more U.S.hospitals40; indeed, the system has grown rapidly,from 120 hospitals in 1991 to more than 300 in2001. The call for broader participation among allU.S. hospitals is even more urgent today.

From the Department of Clinical Epidemiology and InfectiousDiseases, LDS Hospital; and the Department of Internal Medicine,University of Utah School of Medicine — both in Salt Lake City.

1. Leape LL, Brennan TA, Laird N, et al. The nature of adverseevents in hospitalized patients: results of the Harvard Medical Prac-tice Study II. N Engl J Med 1991;324:377-84.2. Rothman KJ. Sleuthing in hospitals. N Engl J Med 1985;313:258-60.3. Langmuir AD. The Epidemic Intelligence Service of the Centerfor Disease Control. Public Health Rep 1980;95:470-7.4. Weinstein RA. Nosocomial infection update. Emerg Infect Dis1998;4:416-20.5. Jarvis WR. Infection control and changing health-care deliverysystems. Emerg Infect Dis 2001;7:170-3.6. Public health focus: surveillance, prevention, and control ofnosocomial infections. MMWR Morb Mortal Wkly Rep 1992;41:783-7.7. Stone PW, Larson E, Kawar LN. A systematic audit of economicevidence linking nosocomial infections and infection control inter-ventions: 1990-2000. Am J Infect Control 2002;30:145-52.8. Haley RW, Culver DH, White JW, Morgan WM, Emori TG. Thenationwide nosocomial infection rate: a new need for vital statistics.Am J Epidemiol 1985;121:159-67.9. National Nosocomial Infections Surveillance (NNIS) report,data summary from October 1986–April 1996, issued May 1996: areport from the National Nosocomial Infections Surveillance (NNIS)system. Am J Infect Control 1996;24:380-8.10. Eggimann P, Pittet D. Infection control in the ICU. Chest 2001;120:2059-93.11. Wenzel RP, Edmond MB. The impact of hospital-acquiredbloodstream infections. Emerg Infect Dis 2001;7:174-7.12. Fleming CA, Steger KA, Craven DE. Host- and device-associatedrisk factors for nosocomial pneumonia: cost-effective strategies forprevention. In: Jarvis WR, ed. Nosocomial pneumonia. New York:Marcel Dekker, 2000:53-92.13. Olsen MA, Fraser VJ. Proving your value in healthcare epidemi-ology and infection control. Semin Infect Control 2002;2:26-50.





14. Pittet D. Improving adherence to hand hygiene practice: a mul-tidisciplinary approach. Emerg Infect Dis 2001;7:234-40.15. Farr BM. Reasons for noncompliance with infection controlguidelines. Infect Control Hosp Epidemiol 2000;21:411-6.16. Pittet D, Hugonnet S, Harbarth S, et al. Effectiveness of a hospi-tal-wide programme to improve compliance with hand hygiene.Lancet 2000;356:1307-12. [Erratum, Lancet 2000;356:2196.]17. Boyce JM, Pittet D. Guideline for hand hygiene in health-caresettings: recommendations of the Healthcare Infection Control Prac-tices Advisory Commmittee and the HICPAC/SHEA/APIC/IDSAHand Hygiene Task Force. MMWR Morb Mortal Wkly Rep 2002;51(RR-16):1-45. (Also available at http://www.cdc.gov/ncidod/hip/default.htm.)18. Burke JP. Randomized controlled trials in hospital epidemiol-ogy: sixth annual National Foundation for Infectious Diseases lec-ture. Am J Infect Control 1983;11:165-73.19. Hunt TK. Surgical wound infections: an overview. Am J Med1981;70:712-8.20. Scheckler WE, Brimhall D, Buck AS, et al. Requirements forinfrastructure and essential activities of infection control and epide-miology in hospitals: a consensus panel report. Infect Control HospEpidemiol 1998;19:114-24.21. Burke JP, Ingall D, Klein JO, Gezon HM, Finland M. Proteus mira-bilis infections in a hospital nursery traced to a human carrier.N Engl J Med 1971;284:115-21.22. Garibaldi RA, Burke JP, Dickman ML, Smith CB. Factors predis-posing to bacteriuria during indwelling urethral catheterization.N Engl J Med 1974;291:215-9.23. Kritchevsky SB, Simmons BP, Braun BI. The Project to MonitorIndicators: a collaborative effort between the Joint Commission onAccreditation of Healthcare Organizations and the Society ofHealthcare Epidemiology of America. Infect Control Hosp Epide-miol 1995;16:33-5.24. Classen DC, Evans RS, Pestotnik SL, Horn SD, Menlove RL,Burke JP. The timing of prophylactic administration of antibioticsand the risk of surgical-wound infection. N Engl J Med 1992;326:281-6.25. Silver A, Eichorn A, Kral J, et al. Timeliness and use of antibioticprophylaxis in selected inpatient surgical procedures. Am J Surg1996;171:548-52.26. Burke JP. Maximizing appropriate antibiotic prophylaxis forsurgical patients: an update from LDS Hospital, Salt Lake City. ClinInfect Dis 2001;33:Suppl 2:S78-S83.

27. Scheckler WE. Feedback of surgical-site infection rates to sur-geons: recommendations, the data, and the current reality. SeminInfect Control 2002;2:81-5.28. Platt R. Progress in surgical-site infection surveillance. InfectControl Hosp Epidemiol 2002;23:361-3.29. Calfee DP, Farr BM. Infection control and cost control in the eraof managed care. Infect Control Hosp Epidemiol 2002;23:407-10.30. Gaynes RP, Richards C, Edwards J, et al. Feeding back surveil-lance data to prevent hospital-acquired infections. Emerg Infect Dis2001;7:295-8.31. Weinstein RA. Lessons from an epidemic, again. N Engl J Med2001;344:1544-5.32. Jarvis WR. Hospital Infections Program, Centers for DiseaseControl and Prevention on-site outbreak investigations, 1990 to1999. Semin Infect Control 2001;1:74-84.33. Peterson LR, Brossette SE. Hunting health care-associatedinfections from the clinical microbiology laboratory: passive, active,and virtual surveillance. J Clin Microbiol 2002;40:1-4.34. Leverstein-van Hall MA, Box ATA, Blok HEM, Paauw A, FluitAC, Verhoef J. Evidence of extensive interspecies transfer of inte-gron-mediated antimicrobial resistance genes among multi-drugresistant Enterobacteriaceae in a clinical setting. J Infect Dis 2002;186:49-56.35. Burke JP, Classen DC, Pestotnik SL, Evans RS, Stevens LE. TheHELP system and its application to infection control. J Hosp Infect1991;18:Suppl A:424-31.36. Classen DC, Pestotnik SL, Evans RS, Burke JP. Computerizedsurveillance of adverse drug events in hospital patients. JAMA 1991;266:2847-51. [Erratum, JAMA 1992;267:1922.]37. Evans RS, Pestotnik SL, Classen DC, et al. A computer-assistedmanagement program for antibiotics and other antiinfective agents.N Engl J Med 1998;338:232-8.38. Making health care safer: a critical analysis of patient safetypractices. Evid Rep Technol Assess (Summ) 2001;43:1-668.39. Monitoring hospital-acquired infections to promote patientsafety — United States, 1990–1999. MMWR Morb Mortal Wkly Rep2000;49:149-53. [Erratum, MMWR Morb Mortal Wkly Rep 2000;49:189-90.]40. Lederberg J, Shope RE, Oaks SC Jr, eds. Emerging infections:microbial threats to health in the United States. Washington, D.C.:National Academy Press, 1992:121-2.Copyright © 2003 Massachusetts Medical Society.

clinical problem-solving series

The Journal welcomes submissions of manuscripts for the Clinical Problem-Solving series. This regular feature considers the step-by-step process of clinical decision making. For more information, please see http://www.nejm.org/hfa/articles.asp.



infection control — a problem for patient safety.pdf

Documents