Acknowledgments

Drafts of this document have been reviewed by leaders of numerous medical, scientific, public health, and labor organizations and others expert in tuberculosis, acquired immunodeficiency syndrome, infection control, hospital epidemiology, microbiology, ventilation, industrial hygiene, nursing, dental practice, or emergency medical services. We thank the many organizations and individuals for their thoughtful comments, suggestions, and assistance.

TB Infection-Control Guidelines Work Group

Carmine J. Bozzi Dale R. Burwen, M.D. Samuel W. Dooley, M.D. Patricia M. Simone, M.D. National Center for Prevention Services

Consuelo Beck-Sague, M.D. Elizabeth A. Bolyard, R.N., M.P.H.

William R. Jarvis, M.D. National Center for Infectious Diseases

Philip J. Bierbaum Christine A. Hudson, M.P.H.

Robert T. Hughes Linda S. Martin, Ph.D. Robert J. Mullan, M.D. National Institute for Occupational Safety and Health

Brian M. Willis, J.D., M.P.H.

Office of the Director

Executive Summary

This document updates and replaces all previously published guidelines for the prevention of Mycobacterium tuberculosis transmission in health-care facilities. The purpose of this revision is to emphasize the importance of a) the hierarchy of control measures, including administrative and engineering controls and personal respiratory protection; b) the use of risk assessments for developing a written tuberculosis (TB) control plan; c) early identifi- cation and management of persons who have TB; d) TB screening programs for health-care workers (HCWs); e) HCW training and education; and f) the evaluation of TB infection-control programs.

Transmission of M. tuberculosis is a recognized risk to patients and HCWs in health-care facilities. Transmission is most likely to occur from patients who have unrecognized pulmonary or laryngeal TB, are not on effective anti-TB therapy, and have not been placed in TB isolation. Several recent TB outbreaks in health-care facilities, including outbreaks of multidrug- resistant TB, have heightened concern about nosocomial transmission. Patients who have multidrug-resistant TB can remain infectious for prolonged periods, which increases the risk for nosocomial and/or occupational transmission of M. tuberculosis. Increases in the incidence of TB have been observed in some geographic areas; these increases are related partially to the high risk for TB among immunosuppressed persons, particularly those infected with human immunodeficiency virus (HIV). Transmission of M. tuberculosis to HIV-infected persons is of particular concern because these persons are at high risk for developing active TB if they become infected with the bacteria. Thus, health- care facilities should be particularly alert to the need for preventing transmission of M. tuberculosis in settings in which HIV-infected persons work or receive care.

Supervisory responsibility for the TB infection-control program should be assigned to a designated person or group of persons who should be given the authority to implement and enforce TB infection-control policies. An effective TB infection-control program requires early identification, isolation, and treatment of persons who have active TB. The primary emphasis of TB infection-control plans in health-care facilities should be achieving these three goals by the application of a hierarchy of control measures, including a) the use of administrative measures to reduce the risk for exposure to persons who have infectious TB, b) the use of engineering controls to prevent the spread and reduce the concentration of infectious droplet nuclei, and c) the use of personal respiratory protective equipment in areas where there is still a risk for exposure to M. tuberculosis (e.g., TB isolation rooms). Implementation of a TB infection-control program requires risk assessment and development of a TB infection-control plan; early identification, treatment, and isolation of infectious TB patients; effective engineering controls; an appropriate respiratory protection program; HCW TB training, education, counseling, and screening; and evaluation of the program's effectiveness.

Although completely eliminating the risk for transmission of M. tuberculosis in all health-care facilities may not be possible at the present time, adherence to these guidelines should reduce the risk to persons in these settings. Recently, nosocomial TB outbreaks have demonstrated the substantial morbidity and mortality among patients and HCWs that have been associated with incomplete implementation of CDC's Guidelines for Preventing the Transmission of Tuberculosis in Health-Care Facilities, with Special Focus on HIV-Related Issues published in 1990. * Follow-up investigations at some of these hospitals have documented that complete implementation of measures similar or identical to those in the 1990 TB Guidelines significantly reduced or eliminated nosocomial transmission of M. tuberculosis to patients and/or HCWs.

1. Introduction

1. Purpose of Document

   In April 1992, the National MDR-TB Task Force published the National Action Plan to Combat Multidrug-Resistant Tuberculosis (1). The publication was a response to reported nosocomial outbreaks of tuberculosis (TB), including outbreaks of multidrug-resistant TB (MDR-TB), and the increasing incidence of TB in some geographic areas. The plan called for the update and revision of the guidelines for preventing nosocomial transmission of Mycobacterium tuberculosis published December 7, 1990 (2).

   Public meetings were held in October 1992 and January 1993 to discuss revision of the 1990 TB Guidelines (2). CDC received considerable input on various aspects of infection control, including health-care worker (HCW) education; administrative controls (e.g., having protocols for the early identification and management of patients who have TB); the need for more specific recommendations regarding ventilation; and clarification on the use of respiratory protection in health-care settings. On the basis of these events and the input received, on October 12, 1993, CDC published in the Federal Register the Draft Guidelines For Preventing the Transmission of Tuberculosis in Health-Care Facilities, Second Edition (3). During and after the 90-day comment period following publication of this draft, CDC's TB Infection-Control Guidelines Work Group received and reviewed more than 2,500 comments.

   The purpose of this document is to make recommendations for reducing the risk for transmitting M. tuberculosis to HCWs, patients, volunteers, visitors, and other persons in these settings. The information also may serve as a useful resource for educating HCWs about TB.

   These recommendations update and replace all previously published CDC recommendations for TB infection control in health-care facilities (2,4). The recommendations in this document are applicable primarily to inpatient facilities in which health care is provided (e.g., hospitals, medical wards in correctional facilities, nursing homes, and hospices). Recommendations applicable to ambulatory-care facilities, emergency departments, home-health-care settings, emergency medical services, medical offices, dental settings, and other facilities or residential settings that provide medical care are provided in separate sections, with cross-references to other sections of the guidelines if appropriate.

   Designated personnel at health-care facilities should conduct a risk assessment for the entire facility and for each area ** and occupa- tional group, determine the risk for nosocomial or occupational transmission of M. tuberculosis, and implement an appropriate TB infection-control program. The extent of the TB infection-control program may range from a simple program emphasizing administrative controls in settings where there is minimal risk for exposure to M. tuberculosis, to a comprehensive program that includes administrative controls, engineering controls, and respiratory protection in settings where the risk for exposure is high. In all settings, administrative measures should be used to minimize the number of HCWs exposed to M. tuberculosis while still providing optimal care for TB patients. HCWs providing care to patients who have TB should be informed about the level of risk for transmission of M. tuberculosis and the appropriate control measures to minimize that risk.

   In this document, the term "HCWs" refers to all the paid and unpaid persons working in health-care settings who have the potential for exposure to M. tuberculosis. This may include, but is not limited to, physicians; nurses; aides; dental workers; technicians; workers in laboratories and morgues; emergency medical service (EMS) personnel; students; part-time personnel; temporary staff not employed by the health-care facility; and persons not involved directly in patient care but who are potentially at risk for occupational exposure to M. tuberculosis (e.g., volunteer workers and dietary, housekeeping, maintenance, clerical, and janitorial staff).

   Although the purpose of this document is to make recommendations for reducing the risk for transmission of M. tuberculosis in health-care facilities, the process of implementing these recommendations must safeguard, in accordance with applicable state and federal laws, the confidentiality and civil rights of persons who have TB.

2. Epidemiology, Transmission, and Pathogenesis of TB

   The prevalence of TB is not distributed evenly throughout all segments of the U.S. population. Some subgroups or persons have a higher risk for TB either because they are more likely than other persons in the general population to have been exposed to and infected with M. tuberculosis or because their infection is more likely to progress to active TB after they have been infected (5). In some cases, both of these factors may be present. Groups of persons known to have a higher prevalence of TB infection include contacts of persons who have active TB, foreign-born persons from areas of the world with a high prevalence of TB (e.g., Asia, Africa, the Caribbean, and Latin America), medically underserved populations (e.g., some African-Americans, Hispanics, Asians and Pacific Islanders, American Indians, and Alaskan Natives), homeless persons, current or former correctional-facility inmates, alcoholics, injecting-drug users, and the elderly. Groups with a higher risk for progression from latent TB infection to active disease include persons who have been infected recently (i.e., within the previous 2 years), children less than 4 years of age, persons with fibrotic lesions on chest radiographs, and persons with certain medical conditions (i.e., human immunodeficiency virus {HIV} infection, silicosis, gastrectomy or jejuno-ileal bypass, being greater than or equal to 10% below ideal body weight, chronic renal failure with renal dialysis, diabetes mellitus, immunosuppression resulting from receipt of high-dose corticosteroid or other immunosuppressive therapy, and some malignancies) (5).

   M. tuberculosis is carried in airborne particles, or droplet nuclei, that can be generated when persons who have pulmonary or laryngeal TB sneeze, cough, speak, or sing (6). The particles are an estimated 1-5 um in size, and normal air currents can keep them airborne for prolonged time periods and spread them throughout a room or building (7). Infection occurs when a susceptible person inhales droplet nuclei containing M. tuberculosis, and these droplet nuclei traverse the mouth or nasal passages, upper respiratory tract, and bronchi to reach the alveoli of the lungs. Once in the alveoli, the organisms are taken up by alveolar macrophages and spread throughout the body. Usually within 2-10 weeks after initial infection with M. tubercu- losis, the immune response limits further multiplication and spread of the tubercle bacilli; however, some of the bacilli remain dormant and viable for many years. This condition is referred to as latent TB infection. Persons with latent TB infection usually have positive purified protein derivative (PPD)-tuberculin skin-test results, but they do not have symptoms of active TB, and they are not infectious.

   In general, persons who become infected with M. tuberculosis have approximately a 10% risk for developing active TB during their lifetimes. This risk is greatest during the first 2 years after infection. Immunocompromised persons have a greater risk for the progression of latent TB infection to active TB disease; HIV infection is the strongest known risk factor for this progression. Persons with latent TB infection who become coinfected with HIV have approximately an 8%-10% risk per year for developing active TB (8). HIV-infected persons who are already severely immunosuppressed and who become newly infected with M. tuberculosis have an even greater risk for developing active TB (9-12).

   The probability that a person who is exposed to M. tuberculosis will become infected depends primarily on the concentration of infectious droplet nuclei in the air and the duration of exposure. Character- istics of the TB patient that enhance transmission include a) disease in the lungs, airways, or larynx; b) presence of cough or other forceful expiratory measures; c) presence of acid-fast bacilli (AFB) in the sputum; d) failure of the patient to cover the mouth and nose when coughing or sneezing; e) presence of cavitation on chest radiograph; f) inappropriate or short duration of chemotherapy; and g) administration of procedures that can induce coughing or cause aerosolization of M. tuberculosis (e.g., sputum induction). Environ- mental factors that enhance the likelihood of transmission include a) exposure in relatively small, enclosed spaces; b) inadequate local or general ventilation that results in insufficient dilution and/or removal of infectious droplet nuclei; and c) recirculation of air containing infectious droplet nuclei. Characteristics of the persons exposed to M. tuberculosis that may affect the risk for becoming infected are not as well defined. In general, persons who have been infected previously with M. tuberculosis may be less susceptible to subsequent infection. However, reinfection can occur among previously infected persons, especially if they are severely immunocompromised. Vaccination with Bacille of Calmette and Guerin (BCG) probably does not affect the risk for infection; rather, it decreases the risk for progressing from latent TB infection to active TB (13). Finally, although it is well established that HIV infection increases the likelihood of progressing from latent TB infection to active TB, it is unknown whether HIV infection increases the risk for becoming infected if exposed to M. tuberculosis.

3. Risk for Nosocomial Transmission of M. tuberculosis

   Transmission of M. tuberculosis is a recognized risk in health-care facilities (14-22). The magnitude of the risk varies considerably by the type of health-care facility, the prevalence of TB in the community, the patient population served, the HCW's occupational group, the area of the health-care facility in which the HCW works, and the effectiveness of TB infection-control interventions. The risk may be higher in areas where patients with TB are provided care before diagnosis and initiation of TB treatment and isolation precautions (e.g., in clinic waiting areas and emergency departments) or where diagnostic or treatment procedures that stimulate coughing are performed. Nosocomial transmission of M. tuberculosis has been associated with close contact with persons who have infectious TB and with the performance of certain procedures (e.g., bronchoscopy {17}, endotracheal intubation and suctioning {18}, open abscess irrigation {20}, and autopsy {21,22}). Sputum induction and aerosol treatments that induce coughing may also increase the potential for transmission of M. tuberculosis (23,24). Personnel of health-care facilities should be particularly alert to the need for preventing transmission of M. tuberculosis in those facilities in which immunocompromised persons (e.g., HIV-infected persons) work or receive care -- especially if cough-inducing procedures, such as sputum induction and aerosolized pentamidine treatments, are being performed.

   Several TB outbreaks among persons in health-care facilities have been reported recently (11,24-28; CDC, unpublished data). Many of these outbreaks involved transmission of multidrug-resistant strains of M. tuberculosis to both patients and HCWs. Most of the patients and some of the HCWs were HIV-infected persons in whom new infection progressed rapidly to active disease. Mortality associated with those outbreaks was high (range: 43%-93%). Furthermore, the interval between diagnosis and death was brief (range of median intervals: 4-16 weeks). Factors contributing to these outbreaks included delayed diagnosis of TB, delayed recognition of drug resistance, and delayed initiation of effective therapy -- all of which resulted in prolonged infectiousness, delayed initiation and inadequate duration of TB isolation, inadequate ventilation in TB isolation rooms, lapses in TB isolation practices and inadequate precautions for cough-inducing procedures, and lack of adequate respiratory protection. Analysis of data collected from three of the health-care facilities involved in the outbreaks indicates that transmission of M. tuberculosis decreased significantly or ceased entirely in areas where measures similar to those in the 1990 TB Guidelines were implemented (2,29- 32). However, several interventions were implemented simultaneously, and the effectiveness of the separate interventions could not be determined.

4. Fundamentals of TB Infection Control

   An effective TB infection-control program requires early identifi- cation, isolation, and effective treatment of persons who have active TB. The primary emphasis of the TB infection-control plan should be on achieving these three goals. In all health-care facilities, particularly those in which persons who are at high risk for TB work or receive care, policies and procedures for TB control should be developed, reviewed periodically, and evaluated for effectiveness to determine the actions necessary to minimize the risk for transmission of M. tuberculosis.

   The TB infection-control program should be based on a hierarchy of control measures. The first level of the hierarchy, and that which affects the largest number of persons, is using administrative measures intended primarily to reduce the risk for exposing uninfected persons to persons who have infectious TB. These measures include a) developing and implementing effective written policies and protocols to ensure the rapid identification, isolation, diagnostic evaluation, and treatment of persons likely to have TB; b) imple- menting effective work practices among HCWs in the health-care facility (e.g., correctly wearing respiratory protection and keeping doors to isolation rooms closed); c) educating, training, and counseling HCWs about TB; and d) screening HCWs for TB infection and disease.

   The second level of the hierarchy is the use of engineering controls to prevent the spread and reduce the concentration of infectious droplet nuclei. These controls include a) direct source control using local exhaust ventilation, b) controlling direction of airflow to prevent contamination of air in areas adjacent to the infectious source, c) diluting and removing contaminated air via general ventilation, and d) air cleaning via air filtration or ultraviolet germicidal irradiation (UVGI).

   The first two levels of the hierarchy minimize the number of areas in the health-care facility where exposure to infectious TB may occur, and they reduce, but do not eliminate, the risk in those few areas where exposure to M. tuberculosis can still occur (e.g., rooms in which patients with known or suspected infectious TB are being isolated and treatment rooms in which cough-inducing or aerosol- generating procedures are performed on such patients). Because persons entering such rooms may be exposed to M. tuberculosis, the third level of the hierarchy is the use of personal respiratory protective equipment in these and certain other situations in which the risk for infection with M. tuberculosis may be relatively higher.

   Specific measures to reduce the risk for transmission of M. tubercu- losis include the following:

• Assigning to specific persons in the health-care facility the supervisory responsibility for designing, implementing, evaluating, and maintaining the TB infection-control program (Section II.A).

• Conducting a risk assessment to evaluate the risk for trans- mission of M. tuberculosis in all areas of the health-care facility, developing a written TB infection-control program based on the risk assessment, and periodically repeating the risk assessment to evaluate the effectiveness of the TB infection- control program (Section II.B).

• Developing, implementing, and enforcing policies and protocols to ensure early identification, diagnostic evaluation, and effective treatment of patients who may have infectious TB (Section II.C; Suppl. 2).

• Providing prompt triage for and appropriate management of patients in the outpatient setting who may have infectious TB (Section II.D).

• Promptly initiating and maintaining TB isolation for persons who may have infectious TB and who are admitted to the inpatient setting (Section II.E; Suppl. 1).

• Effectively planning arrangements for discharge (Section II.E).

• Developing, installing, maintaining, and evaluating ventilation and other engineering controls to reduce the potential for airborne exposure to M. tuberculosis (Section II.F; Suppl. 3).

• Developing, implementing, maintaining, and evaluating a respir- atory protection program (Section II.G; Suppl. 4).

• Using precautions while performing cough-inducing procedures (Section II.H; Suppl. 3).

• Educating and training HCWs about TB, effective methods for preventing transmission of M. tuberculosis, and the benefits of medical screening programs (Section II.I).

• Developing and implementing a program for routine periodic counseling and screening of HCWs for active TB and latent TB infection (Section II.J; Suppl. 2).

• Promptly evaluating possible episodes of M. tuberculosis transmission in health-care facilities, including PPD skin-test conversions among HCWs, epidemiologically associated cases among HCWs or patients, and contacts of patients or HCWs who have TB and who were not promptly identified and isolated (Section II.K).

• Coordinating activities with the local public health department, emphasizing reporting, and ensuring adequate discharge follow-up and the continuation and completion of therapy (Section II.L).

2. Recommendations

1. Assignment of Responsibility

• Supervisory responsibility for the TB infection-control program should be assigned to a designated person or group of persons with expertise in infection control, occupational health, and engineering. These persons should be given the authority to implement and enforce TB infection-control policies.

• If supervisory responsibility is assigned to a committee, one person should be designated as the TB contact person. Questions and problems can then be addressed to this person.

2. Risk Assessment, Development of the TB Infection-Control Plan, and Periodic Reassessment

1. Risk assessment

1. General

• TB infection-control measures for each health-care facility should be based on a careful assessment of the risk for transmission of M. tuberculosis in that particular setting. The first step in developing the TB infection-control program should be to conduct a baseline risk assessment to evaluate the risk for transmission of M. tuberculosis in each area and occupational group in the facility (Table_1, Figure_1 Figure_1a Figure_1c). Appropriate infection-control inter- ventions can then be developed on the basis of actual risk. Risk assessments should be performed for all inpatient and outpatient settings (e.g., medical and dental offices).

• Regardless of risk level, the management of patients with known or suspected infectious TB should not vary. However, the index of suspicion for infectious TB among patients, the frequency of HCW PPD skin testing, the number of TB isolation rooms, and other factors will depend on whether the risk for transmission of M. tuberculosis in the facility, area, or occupational group is high, intermediate, low, very low, or minimal.

• The risk assessment should be conducted by a qualified person or group of persons (e.g., hospital epidemi- ologists, infectious disease specialists, pulmonary disease specialists, infection-control practitioners, health-care administrators, occupational health personnel, engineers, HCWs, or local public health personnel).

• The risk assessment should be conducted for the entire facility and for specific areas within the facility (e.g., medical, TB, pulmonary, or HIV wards; HIV, infectious disease, or pulmonary clinics; and emergency departments or other areas where TB patients might receive care or where cough-inducing procedures are performed). This should include both inpatient and outpatient areas. In addition, risk assessments should be conducted for groups of HCWs who work throughout the facility rather than in a specific area (e.g., respir- atory therapists; bronchoscopists; environmental services, dietary, and maintenance personnel; and students, interns, residents, and fellows).

• Classification of risk for a facility, for a specific area, and for a specific occupational group should be based on a) the profile of TB in the community; b) the number of infectious TB patients admitted to the area or ward, or the estimated number of infectious TB patients to whom HCWs in an occupational group may be exposed; and c) the results of analysis of HCW PPD test conversions (where applicable) and possible person-to-person transmission of M. tuberculosis (Figure_1 Figure_1a Figure_1c).

• All TB infection-control programs should include periodic reassessments of risk. The frequency of repeat risk assessments should be based on the results of the most recent risk assessment (Table_2, Figure_1 Figure_1a Figure_1c).

• The "minimal-risk" category applies only to an entire facility. A "minimal-risk" facility does not admit TB patients to inpatient or outpatient areas and is not located in a community with TB (i.e., counties or communities in which TB cases have not been reported during the previous year). Thus, there is essentially no risk for exposure to TB patients in the facility. This category may also apply to many outpatient settings (e.g., many medical and dental offices).

• The "very low-risk" category generally applies only to an entire facility. A very low-risk facility is one in which

1. patients with active TB are not admitted to inpatient areas but may receive initial assessment and diagnostic evaluation or outpatient management in outpatient areas (e.g., ambulatory-care and emergency departments) and b) patients who may have active TB and need inpatient care are promptly referred to a collaborating facility. In such facilities, the outpatient areas in which exposure to patients with active TB could occur should be assessed and assigned to the appropriate low-, intermediate-, or high-risk category. Categorical assignment will depend on the number of TB patients examined in the area during the preceding year and whether there is evidence of noso- comial transmission of M. tuberculosis in the area. If TB cases have been reported in the community, but no patients with active TB have been examined in the outpatient area during the preceding year, the area can be designated as very low risk (e.g., many medical offices).

   The referring and receiving facilities should establish a referral agreement to prevent inappropriate management and potential loss to follow-up of patients suspected of having TB during evaluation in the triage system of a very low-risk facility.

   In some facilities in which TB patients are admitted to inpatient areas, a very low-risk protocol may be appro- priate for areas (e.g., administrative areas) or occupational groups that have only a very remote possibility of exposure to M. tuberculosis.

   The very low-risk category may also be appropriate for outpatient facilities that do not provide initial assessment of persons who may have TB, but do screen patients for active TB as part of a limited medical screening before undertaking specialty care (e.g., dental settings).

• "Low-risk" areas or occupational groups are those in which a) the PPD test conversion rate is not greater than that for areas or groups in which occupational exposure to M. tuberculosis is unlikely or than previous conversion rates for the same area or group, b) no clusters *** of PPD test conversions have occurred, c) person-to-person transmission of M. tuberculosis has not been detected, and d) fewer than six TB patients are examined or treated per year.

• "Intermediate-risk" areas or occupational groups are those in which a) the PPD test conversion rate is not greater than that for areas or groups in which occupa- tional exposure to M. tuberculosis is unlikely or than previous conversion rates for the same area or group, b) no clusters of PPD test conversions have occurred, c) person-to-person transmission of M. tuberculosis has not been detected, and d) six or more patients with active TB are examined or treated each year. Survey data suggest that facilities in which six or more TB patients are examined or treated each year may have an increased risk for transmission of M. tuberculosis (CDC, unpublished data); thus, areas in which six or more patients with active TB are examined or treated each year (or occupa- tional groups in which HCWs are likely to be exposed to six or more TB patients per year) should be classified as "intermediate risk."

• "High-risk" areas or occupational groups are those in which a) the PPD test conversion rate is significantly greater than for areas or groups in which occupational exposure to M. tuberculosis is unlikely or than previous conversion rates for the same area or group, and epidemi- ologic evaluation suggests nosocomial transmission; or b) a cluster of PPD test conversions has occurred, and epidemiologic evaluation suggests nosocomial transmission of M. tuberculosis; or c) possible person-to-person transmission of M. tuberculosis has been detected.

• If no data or insufficient data for adequate determin- ation of risk have been collected, such data should be compiled, analyzed, and reviewed expeditiously.

2. Community TB profile

• A profile of TB in the community that is served by the facility should be obtained from the public health department. This profile should include, at a minimum, the incidence (and prevalence, if available) of active TB in the community and the drug-susceptibility patterns of M. tuberculosis isolates (i.e., the antituberculous agents to which each isolate is susceptible and those to which it is resistant) from patients in the community.

3. Case surveillance

• Data concerning the number of suspected and confirmed active TB cases among patients and HCWs in the facility should be systematically collected, reviewed, and used to estimate the number of TB isolation rooms needed, to recognize possible clusters of nosocomial transmission, and to assess the level of potential occupational risk. The number of TB patients in specific areas of a facility can be obtained from laboratory surveillance data on specimens positive for AFB smears or M. tuberculosis cultures, from infection-control records, and from databases containing information about hospital discharge diagnoses.

• Drug-susceptibility patterns of M. tuberculosis isolates from TB patients treated in the facility should be reviewed to identify the frequency and patterns of drug resistance. This information may indicate a need to modify the initial treatment regimen or may suggest possible nosocomial transmission or increased occupa- tional risk.

4. Analysis of HCW PPD test screening data

• Results of HCW PPD testing should be recorded in the individual HCW's employee health record and in a retrievable aggregate database of all HCW PPD test results. Personal identifying information should be handled confidentially. PPD test conversion rates should be calculated at appropriate intervals to estimate the risk for PPD test conversions for each area of the facility and for each specific occupational group not assigned to a specific area (Table_2). To calculate PPD test conversion rates, the total number of previously PPD-negative HCWs tested in each area or group (i.e., the denominator) and the number of PPD test conversions among HCWs in each area or group (the numerator) must be obtained.

• PPD test conversion rates for each area or occupational group should be compared with rates for areas or groups in which occupational exposure to M. tuberculosis is unlikely and with previous conversion rates in the same area or group to identify areas or groups where the risk for occupational PPD test conversions may be increased. A low number of HCWs in a specific area may result in a greatly increased rate of conversion for that area, although the actual risk may not be significantly greater than that for other areas. Testing for statistical significance (e.g., Fisher's exact test or chi square test) may assist interpretation; however, lack of statistical significance may not rule out a problem (i.e., if the number of HCWs tested is low, there may not be adequate statistical power to detect a significant difference). Thus, interpretation of individual situations is necessary.

• An epidemiologic investigation to evaluate the likelihood of nosocomial transmission should be conducted if PPD test conversions are noted (Section II.K.1).

• The frequency and comprehensiveness of the HCW PPD testing program should be evaluated periodically to ensure that all HCWs who should be included in the program are being tested at appropriate intervals. For surveillance purposes, earlier detection of transmission may be enhanced if HCWs in a given area or occupational group are tested on different scheduled dates rather than all being tested on the same date (Section II.J.3).

5. Review of TB patient medical records

• The medical records of a sample of TB patients examined at the facility can be reviewed periodically to evaluate infection-control parameters (Table_1). Parameters to examine may include the intervals from date of admission until a) TB was suspected, b) specimens for AFB smears were ordered, c) these specimens were collected, d) tests were performed, and e) results were reported. Moreover, the adequacy of the TB treatment regimens that were used should be evaluated.

• Medical record reviews should note previous hospital admissions of TB patients before the onset of TB symptoms. Patient-to-patient transmission may be suspected if active TB occurs in a patient who had a prior hospitalization during which exposure to another TB patient occurred or if isolates from two or more TB patients have identical characteristic drug-suscepti- bility or DNA fingerprint patterns.

• Data from the case review should be used to determine if there is a need to modify a) protocols for identifying and isolating patients who may have infectious TB, b) laboratory procedures, c) administrative policies and practices, or d) protocols for patient management.

6. Observation of TB infection-control practices

• Assessing adherence to the policies of the TB infection- control program should be part of the evaluation process. This assessment should be performed on a regular basis and whenever an increase occurs in the number of TB patients or HCW PPD test conversions. Areas at high risk for transmission of M. tuberculosis should be monitored more frequently than other areas. The review of patient medical records provides information on HCW adherence to some of the policies of the TB infection-control program. In addition, work practices related to TB isolation (e.g., keeping doors to isolation rooms closed) should be observed to determine if employers are enforcing, and HCWs are adhering to, these policies and if patient adherence is being enforced. If these policies are not being enforced or adhered to, appropriate education and other corrective action should be implemented.

7. Engineering evaluation

• Results of engineering maintenance measures should be reviewed at regular intervals (Table_3). Data from the most recent evaluation and from maintenance procedures and logs should be reviewed carefully as part of the risk assessment.

2. Development of the TB Infection-Control Plan

• Based on the results of the risk assessment, a written TB infection-control plan should be developed and implemented for each area of the facility and for each occupational group of HCWs not assigned to a specific area of the facility (Table_2; Table_3).

• The occurrence of drug-resistant TB in the facility or the community, or a relatively high prevalence of HIV infection among patients or HCWs in the community, may increase the concern about transmission of M. tuberculosis and may influence the decision regarding which protocol to follow (i.e., a higher-risk classification may be selected).

• Health-care facilities are likely to have a combination of low-, intermediate-, and high-risk areas or occupational groups during the same time period. The appropriate protocol should be implemented for each area or group.

• Areas in which cough-inducing procedures are performed on patients who may have active TB should, at the minimum, implement the intermediate-risk protocol.

3. Periodic Reassessment

• Follow-up risk assessment should be performed at the interval indicated by the most recent risk assessment (Figure_1 Figure_1a Figure_1c; Table_2). Based on the results of the follow-up assessment, problem evaluation may need to be conducted or the protocol may need to be modified to a higher- or lower-risk level.

• After each risk assessment, the staff responsible for TB control, in conjunction with other appropriate HCWs, should review all TB control policies to ensure that they are effective and meet current needs.

4. Examples of Risk Assessment

   Examples of six hypothetical situations and the means by which surveillance data are used to select a TB control protocol are described as follows:

   Hospital A. The overall HCW PPD test conversion rate in the facility is 1.6%. No areas or HCW occupational groups have a significantly greater PPD test conversion rate than areas or groups in which occupational exposure to M. tuberculosis is unlikely (or than previous rates for the same area or group). No clusters of PPD test conversions have occurred. Patient-to- patient transmission has not been detected. Patients who have TB are admitted to the facility, but no area admits six or more TB patients per year. The low-risk protocol will be followed in all areas.

   Hospital B. The overall HCW PPD test conversion rate in the facility is 1.8%. The PPD test conversion rate for the medical intensive-care unit rate is significantly higher than all other areas in the facility. The problem identification process is initiated (Section II.K). It is determined that all TB patients have been isolated appropriately. Other potential problems are then evaluated, and the cause for the higher rate is not identified. After consulting the public health department TB infection-control program, the high-risk protocol is followed in the unit until the PPD test conversion rate is similar to areas of the facility in which occupational exposure to TB patients is unlikely. If the rate remains significantly higher than other areas, further evaluation, including environmental and procedural studies, will be performed to identify possible reasons for the high conversion rate.

   Hospital C. The overall HCW PPD test conversion rate in the facility is 2.4%. Rates range from 0 to 2.6% for the individual areas and occupational groups. None of these rates is signifi- cantly higher than rates for areas in which occupational exposure to M. tuberculosis is unlikely. No particular HCW group has higher conversion rates than the other groups. No clusters of HCW PPD test conversions have occurred. In two of the areas, HCWs cared for more than six TB patients during the preceding year. These two areas will follow the intermediate-risk protocol, and all other areas will follow the low-risk protocol. This hospital is located in the southeastern United States, and these conversion rates may reflect cross-reactivity with nontuberculous mycobacteria.

   Hospital D. The overall HCW PPD test conversion rate in the facility is 1.2%. In no area did HCWs care for six or more TB patients during the preceding year. Three of the 20 respiratory therapists tested had PPD conversions, for a rate of 15%. The respiratory therapists who had PPD test conversions had spent all or part of their time in the pulmonary function laboratory, where induced sputum specimens were obtained. A low-risk protocol is maintained for all areas and occupational groups in the facility except for respiratory therapists. A problem evaluation is conducted in the pulmonary function laboratory (Section II.K). It is determined that the ventilation in this area is inadequate. Booths are installed for sputum induction. PPD testing and the risk assessment are repeated 3 months later. If the repeat testing at 3 months indicates that no more conversions have occurred, the respiratory therapists will return to the low-risk protocol.

   Hospital E. Hospital E is located in a community that has a relatively low incidence of TB. To optimize TB services in the community, the four hospitals in the community have developed an agreement that one of them (e.g., Hospital G) will provide all inpatient services to persons who have suspected or confirmed TB. The other hospitals have implemented protocols in their ambulatory-care clinics and emergency departments to identify patients who may have active TB. These patients are then transferred to Hospital G for inpatient care if such care is considered necessary. After discharge from Hospital G, they receive follow-up care in the public health department's TB clinic. During the preceding year, Hospital E has identified fewer than six TB patients in its ambulatory-care and emergency departments and has had no PPD test conversions or other evidence of M. tuberculosis transmission among HCWs or patients in these areas. These areas are classified as low risk, and all other areas are classified as very low risk.

   Hospital F. Hospital F is located in a county in which no TB cases have been reported during the preceding 2 years. A risk assessment conducted at the facility did not identify any patients who had suspected or confirmed TB during the preceding year. The facility is classified as minimal risk.

3. Identifying, Evaluating, and Initiating Treatment for Patients Who May Have Active TB

   The most important factors in preventing transmission of M. tuber- culosis are the early identification of patients who may have infectious TB, prompt implementation of TB precautions for such patients, and prompt initiation of effective treatment for those who are likely to have TB.

1. Identifying patients who may have active TB

• Health-care personnel who are assigned responsibility for TB infection control in ambulatory-care and inpatient settings should develop, implement, and enforce protocols for the early identification of patients who may have infectious TB.

• The criteria used in these protocols should be based on the prevalence and characteristics of TB in the population served by the specific facility. These protocols should be evaluated periodically and revised according to the results of the evaluation. Review of medical records of patients who were examined in the facility and diagnosed as having TB may serve as a guide for developing or revising these protocols.

• A diagnosis of TB may be considered for any patient who has a persistent cough (i.e., a cough lasting for greater than or equal to 3 weeks) or other signs or symptoms compatible with active TB (e.g., bloody sputum, night sweats, weight loss, anorexia, or fever). However, the index of suspicion for TB will vary in different geographic areas and will depend on the prevalence of TB and other characteristics of the population served by the facility. The index of suspicion for TB should be very high in geographic areas or among groups of patients in which the prevalence of TB is high (Section I.B). Appropriate diagnostic measures should be conducted and TB precautions implemented for patients in whom active TB is suspected.

2. Diagnostic evaluation for active TB

• Diagnostic measures for identifying TB should be conducted for patients in whom active TB is being considered. These measures include obtaining a medical history and performing a physical examination, PPD skin test, chest radiograph, and microscopic examination and culture of sputum or other appropriate specimens (6,34,35). Other diagnostic procedures (e.g., bronchoscopy or biopsy) may be indicated for some patients (36,37).

• Prompt laboratory results are crucial to the proper treatment of the TB patient and to early initiation of infection control. To ensure timely results, laboratories performing mycobacteriologic tests should be proficient at both the laboratory and administrative aspects of specimen processing. Laboratories should use the most rapid methods available (e.g., fluorescent microscopy for AFB smears; radiometric culture methods for isolation of mycobacteria; r-nitro-a- acetylamino-b-hydroxy-proprophenone {NAP} test, nucleic acid probes, or high-pressure liquid chromatography {HPLC} for species identification; and radiometric methods for drug- susceptibility testing). As other more rapid or sensitive tests become available, practical, and affordable, such tests should be incorporated promptly into the mycobacteriology laboratory. Laboratories that rarely receive specimens for mycobacteriologic analysis should refer the specimens to a laboratory that more frequently performs these tests.

• Results of AFB sputum smears should be available within 24 hours of specimen collection (38).

• The probability of TB is greater among patients who have positive PPD test results or a history of positive PPD test results, who have previously had TB or have been exposed to M. tuberculosis, or who belong to a group at high risk for TB (Section I.B). Active TB is strongly suggested if the diagnostic evaluation reveals AFB in sputum, a chest radiograph suggestive of TB, or symptoms highly suggestive of TB. TB can occur simultaneously in immunosuppressed persons who have pulmonary infections caused by other organisms (e.g., Pneumocystis carinii or Mycobacterium avium complex) and should be considered in the diagnostic evaluation of all patients who have symptoms compatible with TB (Suppl. 1; Suppl. 2).

• TB may be more difficult to diagnose among persons who have HIV infection (or other conditions associated with severe suppression of cell-mediated immunity) because of a nonclassical clinical or radiographic presentation and/or the simultaneous occurrence of other pulmonary infections (e.g., P. carinii pneumonia and M. avium complex). The difficulty in diagnosing TB in HIV-infected persons may be further compounded by impaired responses to PPD skin tests (39,40), the possibly lower sensitivity of sputum smears for detecting AFB (41), or the overgrowth of cultures with M. avium complex in specimens from patients infected with both M. avium complex and M. tuberculosis (42).

• Immunosuppressed patients who have pulmonary signs or symptoms that are ascribed initially to infections or conditions other than TB should be evaluated initially for coexisting TB. The evaluation for TB should be repeated if the patient does not respond to appropriate therapy for the presumed cause(s) of the pulmonary abnormalities (Suppl. 1; Suppl. 2).

• Patients with suspected or confirmed TB should be reported immediately to the appropriate public health department so that standard procedures for identifying and evaluating TB contacts can be initiated.

3. Initiation of treatment for suspected or confirmed TB

• Patients who have confirmed active TB or who are considered highly likely to have active TB should be started promptly on appropriate treatment in accordance with current guidelines (Suppl. 2) (43). In geographic areas or facilities that have a high prevalence of MDR-TB, the initial regimen used may need to be enhanced while the results of drug-susceptibility tests are pending. The decision should be based on analysis of surveillance data.

• While the patient is in the health-care facility, anti-TB drugs should be administered by directly observed therapy (DOT), the process by which an HCW observes the patient swallowing the medications. Continuing DOT after the patient is discharged should be strongly considered. This decision and the arrangements for providing outpatient DOT should be made in collaboration with the public health department.

4. Management of Patients Who May Have Active TB in Ambulatory-Care Settings and Emergency Departments

• Triage of patients in ambulatory-care settings and emergency departments should include vigorous efforts to promptly identify patients who have active TB. HCWs who are the first points of contact in facilities that serve populations at risk for TB should be trained to ask questions that will facilitate identi- fication of patients with signs and symptoms suggestive of TB.

• Patients with signs or symptoms suggestive of TB should be evaluated promptly to minimize the amount of time they are in ambulatory-care areas. TB precautions should be followed while the diagnostic evaluation is being conducted for these patients.

• TB precautions in the ambulatory-care setting should include a) placing these patients in a separate area apart from other patients, and not in open waiting areas (ideally, in a room or enclosure meeting TB isolation requirements); b) giving these patients surgical masks **** to wear and instructing them to keep their masks on; and c) giving these patients tissues and instructing them to cover their mouths and noses with the tissues when coughing or sneezing.

• TB precautions should be followed for patients who are known to have active TB and who have not completed therapy until a determination has been made that they are noninfectious (Suppl. 1).

• Patients with active TB who need to attend a health-care clinic should have appointments scheduled to avoid exposing HIV-infected or otherwise severely immunocompromised persons to M. tubercu- losis. This recommendation could be accomplished by designating certain times of the day for appointments for these patients or by treating them in areas where immunocompromised persons are not treated.

• Ventilation in ambulatory-care areas where patients at high risk for TB are treated should be designed and maintained to reduce the risk for transmission of M. tuberculosis. General-use areas (e.g., waiting rooms) and special areas (e.g., treatment or TB isolation rooms in ambulatory areas) should be ventilated in the same manner as described for similar inpatient areas (Sections II.E.3, II.F; Suppl. 3). Enhanced general ventilation or the use of air-disinfection techniques (e.g., UVGI or recirculation of air within the room through high-efficiency particulate air {HEPA} filters) may be useful in general-use areas of facilities where many infectious TB patients receive care (Section II.F; Suppl. 3).

• Ideally, ambulatory-care settings in which patients with TB are frequently examined or treated should have a TB isolation room(s) available. Such rooms are not necessary in ambulatory-care settings in which patients who have confirmed or suspected TB are seen infrequently. However, these facilities should have a written protocol for early identification of patients with TB symptoms and referral to an area or a collaborating facility where the patient can be evaluated and managed appropriately. These protocols should be reviewed on a regular basis and revised as necessary. The additional guidelines in Section II.H should be followed in ambulatory-care settings where cough-inducing procedures are performed on patients who may have active TB.

5. Management of Hospitalized Patients Who Have Confirmed or Suspected TB

1. Initiation of isolation for TB

• In hospitals and other inpatient facilities, any patient suspected of having or known to have infectious TB should be placed in a TB isolation room that has currently recommended ventilation characteristics (Section II.E.3; Suppl. 3). Written policies for initiating isolation should specify a) the indications for isolation, b) the person(s) authorized to initiate and discontinue isolation, c) the isolation practices to follow, d) the monitoring of isolation, e) the management of patients who do not adhere to isolation practices, and f) the criteria for discontinuing isolation.

• In rare circumstances, placing more than one TB patient together in the same room may be acceptable. This practice is sometimes referred to as "cohorting." Because of the risk for patients becoming superinfected with drug-resistant organisms, patients with TB should be placed in the same room only if all patients involved a) have culture-confirmed TB, b) have drug-susceptibility test results available on a current specimen obtained during the present hospitalization, c) have identical drug-susceptibility patterns on these specimens, and d) are on effective therapy. Having isolates with identical DNA fingerprint patterns is not adequate evidence for placing two TB patients together in the same room, because isolates with the same DNA fingerprint pattern can have different drug-susceptibility patterns.

• Pediatric patients with suspected or confirmed TB should be evaluated for potential infectiousness according to the same criteria as are adults (i.e., on the basis of symptoms, sputum AFB smears, radiologic findings, and other criteria) (Suppl. 1). Children who may be infectious should be placed in isolation until they are determined to be noninfectious. Pediatric patients who may be infectious include those who have laryngeal or extensive pulmonary involvement, pronounced cough, positive sputum AFB smears, or cavitary TB or those for whom cough-inducing procedures are performed (44).

• The source of infection for a child with TB is often a member of the child's family (45). Therefore, parents and other visitors of all pediatric TB patients should be evaluated for TB as soon as possible. Until they have been evaluated, or the source case is identified, they should wear surgical masks when in areas of the facility outside of the child's room, and they should refrain from visiting common areas in the facility (e.g., the cafeteria or lounge areas).

• TB patients in intensive-care units should be treated the same as patients in noncritical-care settings. They should be placed in TB isolation and have respiratory secretions submitted for AFB smear and culture if they have undiagnosed pulmonary symptoms suggestive of TB.

• If readmitted to a health-care facility, patients who are known to have active TB and who have not completed therapy should have TB precautions applied until a determination has been made that they are noninfectious (Suppl. 1).

2. TB isolation practices

• Patients who are placed in TB isolation should be educated about the mechanisms of M. tuberculosis transmission and the reasons for their being placed in isolation. They should be taught to cover their mouths and noses with a tissue when coughing or sneezing, even while in the isolation room, to contain liquid drops and droplets before they are expelled into the air (46).

• Efforts should be made to facilitate patient adherence to isolation measures (e.g., staying in the TB isolation room). Such efforts might include the use of incentives (e.g., providing them with telephones, televisions, or radios in their rooms or allowing special dietary requests). Efforts should also be made to address other problems that could interfere with adherence to isolation (e.g., management of the patient's withdrawal from addictive substances {including tobacco}).

• Patients placed in isolation should remain in their isolation rooms with the door closed. If possible, diagnostic and treatment procedures should be performed in the isolation rooms to avoid transporting patients through other areas of the facility. If patients who may have infectious TB must be transported outside their isolation rooms for medically essential procedures that cannot be performed in the isolation rooms, they should wear surgical masks that cover their mouths and noses during transport. Persons transporting the patients do not need to wear respiratory protection outside the TB isolation rooms. Procedures for these patients should be scheduled at times when they can be performed rapidly and when waiting areas are less crowded.

• Treatment and procedure rooms in which patients who have infectious TB or who have an undiagnosed pulmonary disease and are at high risk for active TB receive care should meet the ventilation recommendations for isolation rooms (Section II.E.3; Suppl. 3). Ideally, facilities in which TB patients are frequently treated should have an area in the radiology department that is ventilated separately for TB patients. If this is not possible, TB patients should wear surgical masks and should stay in the radiology suite the minimum amount of time possible, then be returned promptly to their isolation rooms.

• The number of persons entering an isolation room should be minimal. All persons who enter an isolation room should wear respiratory protection (Section II.G; Suppl. 4). The patient's visitors should be given respirators to wear while in the isolation room, and they should be given general instructions on how to use their respirators.

• Disposable items contaminated with respiratory secretions are not associated with transmission of M. tuberculosis. However, for general infection-control purposes, these items should be handled and transported in a manner that reduces the risk for transmitting other microorganisms to patients, HCWs, and visitors and that decreases environmental contamination in the health-care facility. Such items should be disposed of in accordance with hospital policy and applicable regulations (Suppl. 5).

3. The TB isolation room

• TB isolation rooms should be single-patient rooms with special ventilation characteristics appropriate for the purposes of isolation (Suppl. 3). The primary purposes of TB isolation rooms are to a) separate patients who are likely to have infectious TB from other persons; b) provide an environ- ment that will allow reduction of the concentration of droplet nuclei through various engineering methods; and c) prevent the escape of droplet nuclei from the TB isolation room and treatment room, thus preventing entry of M. tuber- culosis into the corridor and other areas of the facility.

• To prevent the escape of droplet nuclei, the TB isolation room should be maintained under negative pressure (Suppl. 3). Doors to isolation rooms should be kept closed, except when patients or personnel must enter or exit the room, so that negative pressure can be maintained.

• Negative pressure in the room should be monitored daily while the room is being used for TB isolation.

• The American Society of Heating, Refrigerating and Air- Conditioning Engineers, Inc. (ASHRAE) (47), the American Institute of Architects (AIA) (48), and the Health Resources and Services Administration (49) recommend a minimum of 6 air changes per hour (ACH) for TB isolation and treatment rooms. This ventilation rate is based on comfort and odor control considerations. The effectiveness of this level of airflow in reducing the concentration of droplet nuclei in the room, thus reducing the transmission of airborne pathogens, has not been evaluated directly or adequately.

  Ventilation rates of greater than 6 ACH are likely to produce an incrementally greater reduction in the concentration of bacteria in a room than are lower rates (50-52). However, accurate quantitation of decreases in risk that would result from specific increases in general ventilation levels has not been performed and may not be possible.

  For the purposes of reducing the concentration of droplet nuclei, TB isolation and treatment rooms in existing health- care facilities should have an airflow of greater than or equal to 6 ACH. Where feasible, this airflow rate should be increased to greater than or equal to 12 ACH by adjusting or modifying the ventilation system or by using auxiliary means (e.g., recirculation of air through fixed HEPA filtration systems or portable air cleaners) (Suppl. 3, Section II.B.5.a) (53). New construction or renovation of existing health-care facilities should be designed so that TB isolation rooms achieve an airflow of greater than or equal to 12 ACH.

• Air from TB isolation rooms and treatment rooms used to treat patients who have known or suspected infectious TB should be exhausted to the outside in accordance with applicable federal, state, and local regulations. The air should not be recirculated into the general ventilation. In some instances, recirculation of air into the general ventilation system from such rooms is unavoidable (i.e., in existing facilities in which the ventilation system or facility configuration makes venting the exhaust to the outside impossible). In such cases, HEPA filters should be installed in the exhaust duct leading from the room to the general ventilation system to remove infectious organisms and particulates the size of droplet nuclei from the air before it is returned to the general ventilation system (Section II.F; Suppl. 3). Air from TB isolation and treatment rooms in new or renovated facilities should not be recirculated into the general ventilation system.

• Although not required, an anteroom may increase the effec- tiveness of the isolation room by minimizing the potential escape of droplet nuclei into the corridor when the door is opened. To work effectively, the anteroom should have positive air pressure in relation to the isolation room. The pressure relationship between the anteroom and the corridor may vary according to ventilation design.

• Upper-room air UVGI may be used as an adjunct to general ventilation in the isolation room (Section II.F; Suppl. 3). Air in the isolation room may be recirculated within the room through HEPA filters or UVGI devices to increase the effective ACH and to increase thermal efficiency.

• Health-care facilities should have enough isolation rooms to appropriately isolate all patients who have suspected or confirmed active TB. This number should be estimated using the results of the risk assessment of the health-care facility. Except for minimal- and very low-risk health-care facilities, all acute-care inpatient facilities should have at least one TB isolation room (Section II.B).

• Grouping isolation rooms together in one area of the facility may reduce the possibility of transmitting M. tuberculosis to other patients and may facilitate care of TB patients and the installation and maintenance of optimal engineering (parti- cularly ventilation) controls.

4. Discontinuation of TB isolation

• TB isolation can be discontinued if the diagnosis of TB is ruled out. For some patients, TB can be ruled out when another diagnosis is confirmed. If a diagnosis of TB cannot be ruled out, the patient should remain in isolation until a determination has been made that the patient is noninfec- tious. However, patients can be discharged from the health- care facility while still potentially infectious if appro- priate postdischarge arrangements can be ensured (Section II.E.5).

• The length of time required for a TB patient to become noninfectious after starting anti-TB therapy varies consid- erably (Suppl. 1). Isolation should be discontinued only when the patient is on effective therapy, is improving clinically, and has had three consecutive negative sputum AFB smears collected on different days.

• Hospitalized patients who have active TB should be monitored for relapse by having sputum AFB smears examined regularly (e.g., every 2 weeks). Nonadherence to therapy (i.e., failure to take medications as prescribed) and the presence of drug- resistant organisms are the two most common reasons why patients remain infectious despite treatment. These reasons should be considered if a patient does not respond clinically to therapy within 2-3 weeks.

• Continued isolation throughout the hospitalization should be strongly considered for patients who have MDR-TB because of the tendency for treatment failure or relapse (i.e., difficulty in maintaining noninfectiousness) that has been observed in such cases.

5. Discharge planning

• Before a TB patient is discharged from the health-care facility, the facility's staff and public health authorities should collaborate to ensure continuation of therapy. Discharge planning in the health-care facility should include, at a minimum, a) a confirmed outpatient appointment with the provider who will manage the patient until the patient is cured, b) sufficient medication to take until the outpatient appointment, and c) placement into case management (e.g., DOT) or outreach programs of the public health department. These plans should be initiated and in place before the patient's discharge.

• Patients who may be infectious at the time of discharge should only be discharged to facilities that have isolation capability or to their homes. Plans for discharging a patient who will return home must consider whether all the household members were infected previously and whether any uninfected household members are at very high risk for active TB if infected (e.g., children less than 4 years of age or persons infected with HIV or otherwise severely immunocompromised). If the household does include such persons, arrangements should be made to prevent them from being exposed to the TB patient until a determination has been made that the patient is noninfectious.

6. Engineering Control Recommendations

1. General ventilation

   This section deals only with engineering controls for general-use areas of health-care facilities (e.g., waiting-room areas and emergency departments). Recommendations for engineering controls for specific areas of the facility (e.g., TB isolation rooms) are contained in the sections encompassing those areas. Details regarding ventilation design, evaluation, and supplemental approaches are described in Supplement 3.

• Health-care facilities should either a) include as part of their staff an engineer or other professional with expertise in ventilation or b) have this expertise available from a consultant who is an expert in ventilation engineering and who also has hospital experience. These persons should work closely with infection-control staff to assist in controlling airborne infections.

• Ventilation system designs in health-care facilities should meet any applicable federal, state, and local requirements.

• The direction of airflow in health-care facilities should be designed, constructed, and maintained so that air flows from clean areas to less-clean areas.

• Health-care facilities serving populations that have a high prevalence of TB may need to supplement the general ventil- ation or use additional engineering approaches (i.e., HEPA filtration or UVGI) in general-use areas where TB patients are likely to go (e.g., waiting-room areas, emergency depart- ments, and radiology suites). A single-pass, nonrecirculating system that exhausts air to the outside, a recirculation system that passes air through HEPA filters before recir- culating it to the general ventilation system, or upper air UVGI may be used in such areas.

2. Additional engineering control approaches

1. HEPA filtration

   HEPA filters may be used in a number of ways to reduce or eliminate infectious droplet nuclei from room air or exhaust (Suppl. 3). These methods include placement of HEPA filters

1. in exhaust ducts discharging air from booths or enclosures into the surrounding room; b) in ducts or in ceiling- or wall-mounted units, for recirculation of air within an individual room (fixed recirculation systems); c) in portable air cleaners; d) in exhaust ducts to remove droplet nuclei from air being discharged to the outside, either directly or through ventilation equipment; and e) in ducts discharging air from the TB isolation room into the general ventilation system. In any application, HEPA filters should be installed carefully and maintained meticulously to ensure adequate functioning.

   The manufacturers of in-room air cleaning equipment should provide documentation of the HEPA filter efficiency and the efficiency of the device in lowering room air contaminant levels.

2. UVGI

   For general-use areas in which the risk for transmission of M. tuberculosis is relatively high, UVGI lamps may be used as an adjunct to ventilation for reducing the concentration of infectious droplet nuclei (Suppl. 3), although the effective- ness of such units has not been evaluated adequately. Ultra- violet (UV) units can be installed in a room or corridor to irradiate the air in the upper portion of the room (i.e., upper-room air irradiation), or they can be installed in ducts to irradiate air passing through the ducts. UV units installed in ducts should not be substituted for HEPA filters in ducts that discharge air from TB isolation rooms into the general ventilation system. However, UV units can be used in ducts that recirculate air back into the same room.

   To function properly and decrease hazards to HCWs and others in the health-care facility, UV lamps should be installed properly and maintained adequately, which includes the monitoring of irradiance levels. UV tubes should be changed according to the manufacturer's instructions or when meter readings indicate tube failure. An employee trained in the use and handling of UV lamps should be responsible for these measures and for keeping maintenance records. Applicable safety guidelines should be followed. Caution should be exercised to protect HCWs, patients, visitors, and others from excessive exposure to UV radiation.

7. Respiratory Protection

• Personal respiratory protection should be used by a) persons entering rooms in which patients with known or suspected infectious TB are being isolated, b) persons present during cough-inducing or aerosol-generating procedures performed on such patients, and c) persons in other settings where administrative and engineering controls are not likely to protect them from inhaling infectious airborne droplet nuclei (Suppl. 4). These other settings include transporting patients who may have infectious TB in emergency transport vehicles and providing urgent surgical or dental care to patients who may have infectious TB before a determination has been made that the patient is noninfectious (Suppl. 1).

• Respiratory protective devices used in health-care settings for protection against M. tuberculosis should meet the following standard performance criteria:

1. The ability to filter particles 1 um in size in the unloaded ***** state with a filter efficiency of greater than or equal to 95% (i.e., filter leakage of less than or equal to 5%), given flow rates of up to 50 L per minute.

2. The ability to be qualitatively or quantitatively fit tested in a reliable way to obtain a face-seal leakage of less than or equal to 10% (54,55).

3. The ability to fit the different facial sizes and character- istics of HCWs, which can usually be met by making the respirators available in at least three sizes.

4. The ability to be checked for facepiece fit, in accordance with standards established by the Occupational Safety and Health Administration (OSHA) and good industrial hygiene practice, by HCWs each time they put on their respirators (54,55).

• The facility's risk assessment may identify a limited number of selected settings (e.g., bronchoscopy performed on patients suspected of having TB or autopsy performed on deceased persons suspected of having had active TB at the time of death) where the estimated risk for transmission of M. tuberculosis may be such that a level of respiratory protection exceeding the standard performance criteria is appropriate. In such circumstances, a level of respiratory protection exceeding the standard criteria and compatible with patient-care delivery (e.g., more protective negative-pressure respirators; powered air-purifying particulate respirators {PAPRs}; or positive-pressure air-line, half-mask respirators) should be provided by employers to HCWs who are exposed to M. tuberculosis. Information on these and other respirators is in the NIOSH Guide to Industrial Respiratory Protection (55) and in Supplement 4 of this document.

• In some settings, HCWs may be at risk for two types of exposure:

1. inhalation of M. tuberculosis and b) mucous membrane exposure to fluids that may contain bloodborne pathogens. In these settings, protection against both types of exposure should be used.

• When operative procedures (or other procedures requiring a sterile field) are performed on patients who may have infectious TB, respiratory protection worn by the HCW should serve two functions: a) it should protect the surgical field from the respiratory secretions of the HCW, and b) it should protect the HCW from infectious droplet nuclei that may be expelled by the patient or generated by the procedure. Respirators with exhalation valves and most positive-pressure respirators do not protect the sterile field.

• Health-care facilities in which respiratory protection is used to prevent inhalation of M. tuberculosis are required by OSHA to develop, implement, and maintain a respiratory protection program (Suppl. 4). All HCWs who use respiratory protection should be included in this program. Visitors to TB patients should be given respirators to wear while in isolation rooms, and they should be given general instructions on how to use their respirators.

• Facilities that do not have isolation rooms and do not perform cough-inducing procedures on patients who may have TB may not need to have a respiratory protection program for TB. However, such facilities should have written protocols for the early identification of patients who have signs or symptoms of TB and procedures for referring these patients to a facility where they can be evaluated and managed appropriately. These protocols should be evaluated regularly and revised as needed.

• Surgical masks are designed to prevent the respiratory secretions of the person wearing the mask from entering the air. To reduce the expulsion of droplet nuclei into the air, patients suspected of having TB should wear surgical masks when not in TB isolation rooms. These patients do not need to wear particulate respir- ators, which are designed to filter the air before it is inhaled by the person wearing the respirator. Patients suspected of having or known to have TB should never wear a respirator that has an exhalation valve, because this type of respirator does not prevent expulsion of droplet nuclei into the air.

8. Cough-Inducing and Aerosol-Generating Procedures

1. General guidelines

   Procedures that involve instrumentation of the lower respiratory tract or induce coughing can increase the likelihood of droplet nuclei being expelled into the air. These cough-inducing procedures include endotracheal intubation and suctioning, diagnostic sputum induction, aerosol treatments (e.g., penta- midine therapy), and bronchoscopy. Other procedures that can generate aerosols (e.g., irrigation of tuberculous abscesses, homogenizing or lyophilizing tissue, or other processing of tissue that may contain tubercle bacilli) are also covered by these recommendations.

• Cough-inducing procedures should not be performed on patients who may have infectious TB unless the procedures are absolutely necessary and can be performed with appropriate precautions.

• All cough-inducing procedures performed on patients who may have infectious TB should be performed using local exhaust ventilation devices (e.g., booths or special enclosures) or, if this is not feasible, in a room that meets the ventilation requirements for TB isolation.

• HCWs should wear respiratory protection when present in rooms or enclosures in which cough-inducing procedures are being performed on patients who may have infectious TB.

• After completion of cough-inducing procedures, patients who may have infectious TB should remain in their isolation rooms or enclosures and not return to common waiting areas until coughing subsides. They should be given tissues and instructed to cover their mouths and noses with the tissues when coughing. If TB patients must recover from sedatives or anesthesia after a procedure (e.g, after a bronchoscopy), they should be placed in separate isolation rooms (and not in recovery rooms with other patients) while they are being monitored.

• Before the booth, enclosure, or room is used for another patient, enough time should be allowed to pass for at least 99% of airborne contaminants to be removed. This time will vary according to the efficiency of the ventilation or filtration used (Suppl. 3, Table_S31).

2. Special considerations for bronchoscopy

• If performing bronchoscopy in positive-pressure rooms (e.g., operating rooms) is unavoidable, TB should be ruled out as a diagnosis before the procedure is performed. If the broncho- scopy is being performed for the purpose of diagnosing pulmonary disease and that diagnosis could include TB, the procedure should be performed in a room that meets TB isolation ventilation requirements.

3. Special considerations for the administration of aerosolized pentamidine

• Patients should be screened for active TB before prophylactic therapy with aerosolized pentamidine is initiated. Screening should include obtaining a medical history and performing skin testing and chest radiography.

• Before each subsequent treatment with aerosolized penta- midine, patients should be screened for symptoms suggestive of TB (e.g., development of a productive cough). If such symptoms are elicited, a diagnostic evaluation for TB should be initiated.

• Patients who have suspected or confirmed active TB should take, if clinically practical, oral prophylaxis for P. carinii pneumonia.

9. Education and Training of HCWs

   All HCWs, including physicians, should receive education regarding TB that is relevant to persons in their particular occupational group. Ideally, training should be conducted before initial assignment, and the need for additional training should be reevaluated periodically (e.g., once a year). The level and detail of this education will vary according to the HCW's work responsibilities and the level of risk in the facility (or area of the facility) in which the HCW works. However, the program may include the following elements:

• The basic concepts of M. tuberculosis transmission, pathogenesis, and diagnosis, including information concerning the difference between latent TB infection and active TB disease, the signs and symptoms of TB, and the possibility of reinfection.

• The potential for occupational exposure to persons who have infectious TB in the health-care facility, including information concerning the prevalence of TB in the community and facility, the ability of the facility to properly isolate patients who have active TB, and situations with increased risk for exposure to M. tuberculosis.

• The principles and practices of infection control that reduce the risk for transmission of M. tuberculosis, including information concerning the hierarchy of TB infection-control measures and the written policies and procedures of the facility. Site-specific control measures should be provided to HCWs working in areas that require control measures in addition to those of the basic TB infection-control program.

• The purpose of PPD skin testing, the significance of a positive PPD test result, and the importance of participating in the skin- test program.

• The principles of preventive therapy for latent TB infection. These principles include the indications, use, effectiveness, and the potential adverse effects of the drugs (Suppl. 2).

• The HCW's responsibility to seek prompt medical evaluation if a PPD test conversion occurs or if symptoms develop that could be caused by TB. Medical evaluation will enable HCWs who have TB to receive appropriate therapy and will help to prevent transmission of M. tuberculosis to patients and other HCWs.

• The principles of drug therapy for active TB.

• The importance of notifying the facility if the HCW is diagnosed with active TB so that contact investigation procedures can be initiated.

• The responsibilities of the facility to maintain the confiden- tiality of the HCW while ensuring that the HCW who has TB receives appropriate therapy and is noninfectious before returning to duty.

• The higher risks associated with TB infection in persons who have HIV infection or other causes of severely impaired cell-mediated immunity, including a) the more frequent and rapid development of clinical TB after infection with M. tuberculosis, b) the differences in the clinical presentation of disease, and c) the high mortality rate associated with MDR-TB in such persons.

• The potential development of cutaneous anergy as immune function (as measured by CD4+ T-lymphocyte counts) declines.

• Information regarding the efficacy and safety of BCG vaccination and the principles of PPD screening among BCG recipients.

• The facility's policy on voluntary work reassignment options for immunocompromised HCWs.

10. HCW Counseling, Screening, and Evaluation

    A TB counseling, screening, and prevention program for HCWs should be established to protect both HCWs and patients. HCWs who have positive PPD test results, PPD test conversions, or symptoms suggestive of TB should be identified, evaluated to rule out a diagnosis of active TB, and started on therapy or preventive therapy if indicated (5). In addition, the results of the HCW PPD screening program will contribute to evaluation of the effectiveness of current infection- control practices.

1. Counseling HCWs regarding TB

• Because of the increased risk for rapid progression from latent TB infection to active TB in HIV-infected or otherwise severely immunocompromised persons, all HCWs should know if they have a medical condition or are receiving a medical treatment that may lead to severely impaired cell-mediated immunity. HCWs who may be at risk for HIV infection should know their HIV status (i.e., they should be encouraged to voluntarily seek counseling and testing for HIV antibody status). Existing guidelines for counseling and testing should be followed routinely (56). Knowledge of these conditions allows the HCW to seek the appropriate preventive measures outlined in this document and to consider voluntary work reassignments. Of particular importance is that HCWs need to know their HIV status if they are at risk for HIV infection and they work in settings where patients who have drug-resistant TB may be encountered.

• All HCWs should be informed about the need to follow existing recommendations for infection control to minimize the risk for exposure to infectious agents; implementation of these recommendations will greatly reduce the risk for occupational infections among HCWs (57). All HCWs should also be informed about the potential risks to severely immunocompromised persons associated with caring for patients who have some infectious diseases, including TB. It should be emphasized that limiting exposure to TB patients is the most protective measure that severely immunosuppressed HCWs can take to avoid becoming infected with M. tuberculosis. HCWs who have severely impaired cell-mediated immunity and who may be exposed to M. tuberculosis may consider a change in job setting to avoid such exposure. HCWs should be advised of the option that severely immunocompromised HCWs can choose to transfer voluntarily to areas and work activities in which there is the lowest possible risk for exposure to M. tuber- culosis. This choice should be a personal decision for HCWs after they have been informed of the risks to their health.

• Employers should make reasonable accommodations (e.g., alternative job assignments) for employees who have a health condition that compromises cell-mediated immunity and who work in settings where they may be exposed to M. tuber- culosis. HCWs who are known to be immunocompromised should be referred to employee health professionals who can indivi- dually counsel the employees regarding their risk for TB. Upon the request of the immunocompromised HCW, employers should offer, but not compel, a work setting in which the HCW would have the lowest possible risk for occupational exposure to M. tuberculosis. Evaluation of these situations should also include consideration of the provisions of the Americans With Disabilities Act of 1990 ****** and other applicable federal, state, and local laws.

• All HCWs should be informed that immunosuppressed HCWs should have appropriate follow-up and screening for infectious diseases, including TB, provided by their medical practi- tioner. HCWs who are known to be HIV-infected or otherwise severely immunosuppressed should be tested for cutaneous anergy at the time of PPD testing (Suppl. 2). Consideration should be given to retesting, at least every 6 months, those immunocompromised HCWs who are potentially exposed to M. tuberculosis because of the high risk for rapid progression to active TB if they become infected.

• Information provided by HCWs regarding their immune status should be treated confidentially. If the HCW requests voluntary job reassignment, the confidentiality of the HCW should be maintained. Facilities should have written procedures on confidential handling of such information.

2. Screening HCWs for active TB

• Any HCW who has a persistent cough (i.e., a cough lasting greater than or equal to 3 weeks), especially in the presence of other signs or symptoms compatible with active TB (e.g., weight loss, night sweats, bloody sputum, anorexia, or fever), should be evaluated promptly for TB. The HCW should not return to the workplace until a diagnosis of TB has been excluded or until the HCW is on therapy and a determination has been made that the HCW is noninfectious.

3. Screening HCWs for latent TB infection

• The risk assessment should identify which HCWs have potential for exposure to M. tuberculosis and the frequency with which the exposure may occur. This information is used to determine which HCWs to include in the skin-testing program and the frequency with which they should be tested (Table_2).

• If HCWs are from risks groups with increased prevalence of TB, consideration may be given to including them in the skin- testing program, even if they do not have potential occupa- tional e