The blueprint for prevention of nosocomial tuberculosis transmission is clear, but why don’t we have the will to follow it?

Over the last 30 years, healthcare settings in resource-rich countries have enjoyed tremendous success in the prevention of healthcare-associated tuberculosis (TB) transmission. This success has been driven by implementation of effective infection control measures as well as public health efforts which have reduced the risk of active TB disease in the community and resulted in fewer patients with TB in the hospital. In the late 1980s and early 1990s, there were outbreaks of nosocomial transmission of TB in the USA because of ineffective infection control measures and the emergence of the HIV/AIDS epidemic [1]. Introduction of a hierarchy of TB infection control measures that focused on enhanced administrative controls played a critical role in terminating an outbreak at our own institution in the early 1990s [2]. As part of enhanced administrative controls, airborne infection isolation (AII) precautions at the time of hospital admission became mandatory for patients in whom pulmonary TB was in the differential diagnosis, persons living with HIV with an abnormal chest radiograph (‘atypical’ chest radiograph findings can be seen in persons with HIV and pulmonary TB), and any patient in whom a sputum or other respiratory specimen was ordered for TB diagnostic testing including acid-fast bacilli (AFB) smear and culture. These relatively straightforward interventions led to prevention of nosocomial transmission as well as a nearly 10-fold decline in TB exposure episodes and days patients with pulmonary TB were not placed in AII precautions [2]. In the years following implementation of these administrative controls, new TB infections among healthcare workers (HCWs) due to occupational exposures have been extremely rare [3].

Thus, the blueprint for implementation of effective TB infection control measures has been known for several decades. Our experience and that of others led the US Centers for Disease Control (CDC), the WHO and other groups to recommend effective TB infection control plans based on a hierarchy of control measures (administrative controls, environmental controls, and personal respiratory protection) [4,5]. Administrative controls, which are the most important component of the hierarchy of controls, consist of policies and procedures that allow for early identification of patients with possible pulmonary TB disease and timely initiation of AII precautions. This ensures patients who may have infectious TB are separated from other patients, and those found to have active TB disease are promptly diagnosed and started on effective treatment [6].

Unfortunately, like so many advances in TB treatment and prevention, these successes have not been implemented in those countries most affected by the disease. Uptake of the hierarchy of control measures, particularly administrative controls, has generally been poor in the 30 high-TB-burden countries which account for >85% of global TB cases [7]. The devastating nature of nosocomial transmission of TB to both patients and healthcare providers in these settings (all of which are low- and middle-income countries (LMICs)), was highlighted by nosocomial transmission of extensively drug-resistant (XDR)-TB in South Africa [8]. In this issue of CMI, Kim et al. show persistent shortcomings in TB infection control measures at an academic medical centre located in a high-income country, South Korea [9]. The authors found that over 100 patients with pulmonary TB experienced a delay in being placed in AII precautions lasting 4 days or longer during the 3-year study period and that these delays may have resulted in 17 new TB infections among HCWs [9]. While concerning, the results offer an opportunity to revisit proven TB infection control interventions that would probably have prevented most delays in isolation and prevented nosocomial transmission.

First and foremost is the need for administrative controls. Over 85% of persons diagnosed with pulmonary TB had a sputum smear and culture for AFB collected within 3 days of hospital admission and even in those with a positive result, 22% experienced a delay in the initiation of AII precautions [9]. Although the authors do not report the infection control policies at their institution, these data suggest only persons with a confirmed diagnosis of pulmonary TB are placed on AII precautions and that doing so is at the discretion of the treating clinician. The problem with this approach is that it results in significant delays in appropriate institution of AII precautions in patients with pulmonary TB precisely at the time when they are maximally infectious: immediately prior to diagnosis and initiation of treatment [10]. By the time patients are identified and isolated, they may have already spent several days on an inpatient ward exposing HCWs and other patients (many of whom are highly susceptible because of HIV co-infection or other immunocompromising conditions) to TB while awaiting test results. Thus, it is critical to isolate at the time of admission not only patients with a confirmed diagnosis pulmonary TB, but also those in whom pulmonary TB is considered a diagnostic possibility. Requiring AII precautions of those who may have active TB disease serves as a further control by taking the decision to isolate out of the hands of the clinician. At our institution, ordering a work-up for possible pulmonary TB (e.g., ordering an AFB sputum smear and culture and/or molecular diagnostic test for M. tuberculosis) triggers mandatory AII precautions at admission until either TB has been ruled out (e.g., two negative AFB smears and/or negative TB PCR test) or an alternative diagnosis has been established. A trained infection preventionist reviews isolation orders and microbiology results to ensure potentially infectious patients are isolated in a timely fashion.

Rather than simply leaving the decision of whom to isolate and evaluate for pulmonary TB to the judgement of clinicians, it is important to have administrative controls in place which provide policies, procedures and guidance on which types of patients require AII precautions. To this end, education regarding signs and symptoms of pulmonary TB, local TB epidemiology, and hospital policies and procedures for AII precautions are an important component of administrative controls. Infection control personnel can also help to educate clinicians by providing training on the need for a high index of suspicion for TB and offering feedback when persons with pulmonary TB are not properly isolated. The disparity in timely respiratory isolation of pulmonary TB cases across specialties described by Kim and colleagues underscores why such programmes are important [9].

Thus, if a hierarchy of effective infection control measures are proven to prevent nosocomial TB transmission, why has global uptake been poor? A reason most often cited for this inaction, particularly in LMICs, is limited resources and capacity to implement effective administrative controls. Such concerns are not without reason but political will is needed to find solutions. AII precautions with negative pressure isolation are not available in most LMIC healthcare facilities [5]. LMICs also suffer from shortages of HCWs [11], which may limit available medical personnel to effectively implement infection control protocols.

However, with the introduction of rapid molecular diagnostics such as Xpert MTB/RIF, geographic separation of hospitalized patients under evaluation for pulmonary TB has never been more feasible, including in LMIC settings. One promising strategy that employs many of the proven infection control strategies and can be implemented in resource-limited settings is FAST (Find cases Actively, Separate temporarily, and Treat effectively based on molecular diagnostics) [10,12]. While access to negative pressure rooms will remain challenging in many low-resource settings, geographic separation of patients should be achievable. Sputum testing with the Xpert MTB/RIF can be performed in less than 2 h and has a comparable or higher sensitivity than three AFB sputum smears [13]. Patients under evaluation for pulmonary TB need only be separated for brief periods of time pending Xpert results, greatly limiting the amount of isolation space required. Those diagnosed with pulmonary TB can simultaneously be assessed for rifampin resistance and rapidly started on appropriate therapy, thereby limiting infectiousness. In patients requiring hospitalization, this also allows for geographic cohorting within the hospital based on TB status: no TB, probable drug-susceptible TB, and MDR-TB. Thus, the largest barrier to global implementation of effective administrative controls is a lack of political will, both at the local level and sadly at the international level as well, where TB infection prevention has not always been a high priority.

The important work of Kim and colleagues highlights persistent gaps in global implementation of effective TB infection control measures based on a hierarchy of controls. There remains an urgent need to improve uptake of administrative controls and rapid molecular diagnostics to limit nosocomial TB transmission. The blueprint for success is known. We must now find the political will to implement it.