Back to the future: a brief history of contact precautions

In response to “hundreds of requests for information about effective methods of isolating hospitalized patients,” the Centers for Disease Control (CDC) published the first edition of Isolation Techniques for Use in Hospitals in 1970. ¹⁶ This document would undergo multiple iterations over the years before becoming the current Guideline for Isolation Precautions, ^{Reference Siegel, Rhinehart, Jackson and Chiarello1} published and updated by the Healthcare Infection Control Practices Advisory Committee (HICPAC) since the 1990s. ^{Reference Garner17} In the original recommendations, PPE guidance was based on the specific organism causing infection (ie, rubella, group A Streptococcus). These infectious syndromes were then grouped into isolation types based on the known epidemiology of transmission. Gowns and gloves featured prominently in enteric, wound/skin, and strict precautions. There was a distinction made between wearing gowns/gloves at all times when in the vicinity of the patient, versus with direct contact only. ¹⁶

Major changes to these strategies were made in the 1983 update ^{Reference Garner and Simmons18} in an attempt to simplify. Hospitals could continue to choose isolation strategy by the specific organism or choose a category based on transmission route (ie, “Respiratory,” “Drainage/Secretion,” “Enteric”). The authors noted that using transmission categories simplified overall recommendations, at the cost of using more isolation than strictly necessary for some infections. ^{Reference Garner and Simmons18} Hospitals were encouraged to choose between the two strategies, or “design their own system.” The 1983 Guideline first introduced “Contact Isolation” to decrease transmission of multidrug-resistant organisms (MDROs). However, even in this framework, gowns and gloves were indicated if soiling or touching infective material was likely, rather than every entry into a room. ^{Reference Garner and Simmons18}

The suggestion for hospitals to create their own isolation system was taken literally by Harborview Medical Center, as described by Lynch et al and named “body substance isolation (BSI)”. ^{Reference Lynch, Cummings, Roberts, Herriott, Yates and Stamm19} BSI addressed transmission of healthcare-associated infections (HAI) as well as blood and body fluid (BBF)-derived viruses. It was also a response to the observation that MDROs passed from non-isolated patients to others in the absence of clinical infectious disease. In BSI, gowns and gloves were used for all patients, regardless of infection status, whenever contact with “moist body substances” was anticipated. ^{Reference Lynch, Cummings, Roberts, Herriott, Yates and Stamm19} Universal precautions soon followed BSI in treating every patient as potentially infected. ²⁰

In 1996, HICPAC grouped precautions into standard and transmission-based (TBP), ^{Reference Garner21} distinguishing foundational infection control practices to be applied to all patient care and extra interventions for patients with known infections, such as CP for MDROs. ^{Reference Garner21,Reference Soule22}

While CP continue to be recommended for MDROs in the 2007 HICPAC Guideline ^{Reference Siegel, Rhinehart, Jackson and Chiarello1} and the Management of MDROs in Healthcare Settings, ^{Reference Siegel, Rhinehart, Jackson and Chiarello23} the recommendation includes a bundle of additional guidance around patient placement, staffing, hand hygiene (HH), and cleaning and disinfection. ^{Reference Siegel, Rhinehart, Jackson and Chiarello1,Reference Siegel, Rhinehart, Jackson and Chiarello23} Disentangling CP from IPC bundles has been a major challenge in assessing effectiveness of CP.

Methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant enterococci (VRE)

The controversy around CP for MRSA and VRE spans decades. ^{Reference Boyce, Havill, Kohan, Dumigan and Ligi5,Reference Lai24} It remains a topic of intense sparring between experts in the field, both at professional society meetings and in published medical literature. The crux of the debate is that (1) these organisms are endemic, or widely colonizing both the sick and non-sick, in hospitals as well as the community, so the practice of gowning and gloving for a subset of patients is going to miss much of the potential transmission; (2) there are other parts of the prevention bundle likely more effective as they better address the overall transmission pathways (eg. decolonization, HH); and (3) HCP nonadherence limits benefit of CP. The overall published data on CP for MRSA and VRE are diverse in design and methodology, as well as results, leaving enough uncertainty to continue the discussion.

Many of the studies purported to be addressing efficacy of CP in reducing transmission of MDROs were neither designed nor executed to answer this question. Morgan et al. ^{Reference Morgan, Murthy and Silvia Munoz-Price3} acknowledge this in undertaking their literature review; the initial intent was to include studies that specifically addressed CP as the intervention in endemic settings, but in fact could only find 3 studies in the MRSA literature review and zero for VRE; they pivoted to inclusion of studies bundling CP with active surveillance (AS) and/or universal use of gowns and gloves. ^{Reference Morgan, Murthy and Silvia Munoz-Price3}

The best conclusion we can draw from cluster-randomized studies ^{Reference Huang, Septimus and Kleinman25–Reference Huskins, Huckabee and O’Grady27} looking at the combination of AS and subsequent “extra” CPs, is that putting a few more patients on CP as a result of their screening results does not make much difference. ^{Reference Huang, Septimus and Kleinman25–Reference Huskins, Huckabee and O’Grady27} Any potential benefit of gowns/gloves is overshadowed by more effective parts of the bundle in REDUCE MRSA and MOSAR trials (ie, chlorhexidine gluconate (CHG) skin treatments), ^{Reference Huang, Septimus and Kleinman25} and by little overall difference in CP coverage across the arms of the study groups. ^{Reference Huang, Septimus and Kleinman25–Reference Huskins, Huckabee and O’Grady27}

The multicenter, cluster-randomized, universal gowning and gloving BUGG study was designed specifically to look at CP as a sole intervention for reduction in MRSA and VRE transmission. ^{Reference Harris, Pineles and Belton28} This study showed a significant decrease in MRSA acquisitions when CP were universally applied in the ICU, at nearly 3 acquisitions prevented per 1,000 patient days. ^{Reference Harris, Pineles and Belton28} As many facilities were also using universal CHG bathing, the group performed subsequent analysis to separate the CP intervention from CHG bathing effects, ^{Reference Morgan, Pineles and Shardell29} demonstrating benefit of universal CP beyond that which could be obtained with CHG bathing alone. ^{Reference Morgan, Pineles and Shardell29} No effects on VRE or g organism acquisition were noted. ^{Reference Morgan, Pineles and Shardell29,Reference Harris, Morgan, Pineles, Magder, O’Hara and Johnson30}

Further support for an incremental impact of CP for MRSA comes from the Veterans Affairs system of hospitals during the COVID-19 pandemic, when hospitals were given the option of escalate CP in efforts to conserve PPE. ^{Reference Evans, Simbartl and McCauley31} This natural experiment saw similar hospitals act as controls enabling a comparison of MRSA HAI rates between hospitals keeping CP for all MRSA infected or colonized patients versus de-escalation of CP. Facilities discontinuing CP for MRSA experienced MRSA HAI rates double to triple the rates of facilities using CP, or CP plus AS, respectively. ^{Reference Evans, Simbartl and McCauley31} This study does have limitations, as it was not randomized and there was no assessment for readiness to remove CP. On the other side of the argument, many facilities have published their experience with cessation of CP for MRSA/ VRE, with no observed increase in HAI rates. ^{Reference Kleyman, Cupril-Nilson and Robinson32,Reference Marra, Edmond, Schweizer, Ryan and Diekema33} These facilities focused on horizontal infection prevention strategies, such as CHG bathing, improving HH, HAI prevention bundles, environmental cleaning. ^{Reference Haessler, Martin and Scales34,Reference Martin, Colaianne and Bridge35} HAI rates not only remained stable in many of these facilities, but actually decreased in the postCP period. ^{Reference Haessler, Martin and Scales34,Reference Martin, Colaianne and Bridge35} Epidemiologic data suggests overall MRSA transmission was falling nationwide concurrently, calling into question whether rates would have fallen even further had CP been continued. ^{Reference Kourtis, Hatfield and Baggs36}

Recently, some of the centers involved in initial de-escalation studies have published subsequent data demonstrating stable declines in rates over 10 years after de-escalation of CP, suggesting the data is not just following larger epidemiologic shifts but is positively impacted by the focus on horizonal over vertical methods. ^{Reference Haessler, Martin and Scales34,Reference Martin, Colaianne and Bridge35} One additional study notably performed whole genome sequencing on MRSA isolates in the periods before and after de-implementation of CP, finding no change in transmission as defined by WGS. ^{Reference Karunakaran, Pless and Ayres37}

Another recent study examined the characteristics associated with successful de-escalation of CP. ^{Reference Martin, Colaianne and Bridge35} Twelve hospitals discontinued CP and 3 continued as controls. Successful hospitals had low baseline HAI rates and high HH compliance. The authors note that while this intervention was successful in a variety of settings, hospitals considering changing CP policy should assess their readiness; discontinuation is not recommended in settings with poor HH or high baseline HAI rates. ^{Reference Martin, Colaianne and Bridge35}

Increasingly, we are seeing the declaration, “CP are supported by low quality evidence,” or more bluntly, “CP don’t work.” This is overly simplistic. There are examples of studies evaluating CP that are robust in methodology and expertly executed. If we were honestly interpreting the body of evidence over the last several decades, for both outbreak and endemic settings, including observational and the cluster-randomized studies, we should acknowledge that CP, when used with sufficient patient coverage and fidelity, will likely prevent some transmission of MRSA. ^{Reference Boyce, Havill, Kohan, Dumigan and Ligi5,Reference Harris, Pineles and Belton28,Reference Morgan, Pineles and Shardell29,Reference Evans, Simbartl and McCauley31} The problem is the amount of patients we need to have on CP, and the amount of effort it takes to consistently implement such a strategy across a facility, is increasingly impractical under routine/endemic circumstances. The conclusion is that CP do not prevent enough MRSA transmission to justify the amount of effort and waste that CP generates. This is major paradigm shift from the rhetoric of “getting to zero.” IPC increasingly prioritizes some interventions over others in the context of finite resources.

Martin et al. ^{Reference Martin, Morgan, Pryor and Bearman38} recently published an updated survey of the Society for Healthcare Epidemiology of America (SHEA) Research Network facilities, finding that the number of hospitals that have de-escalated CP in 2021 rose to 35% from 7% in 2015 when a similar survey had been performed. Most facilities still using CP were interested in changing their policies, noting the precedent set by peer organizations. ^{Reference Martin, Morgan, Pryor and Bearman38}

Extended-spectrum beta-lactamase-producing enterobacterales (ESBL-E)

Similarly to the MRSA/VRE experience, many healthcare facilities are questioning the use of CP for ESBL-E infected and colonized patients. IPC guidelines continue to recommend CP for ESBL-E, ^{Reference Siegel, Rhinehart, Jackson and Chiarello1,Reference Tacconelli, Cataldo and Dancer39} given the difficulty in treating infections caused by ESBL-E, the propensity of ESBL-E to cause outbreaks (particularly in acute ^{Reference Harris, Perencevich and Johnson40} and long-term care ^{Reference Burgess, Johnson and Porter41} facilities), and the successful resolution of outbreaks using CP. ^{Reference Podnos, Cinat, Wilson, Cooke, Gornick and Thrupp42–Reference Laurent, Rodriguez-Villalobos and Rost44} Patients at highest risk for acquisition of ESBL-E in healthcare settings include those receiving extensive nursing care at the bedside, further emphasizing a potential role for PPE to prevent transmission. ^{Reference Harris, Perencevich and Johnson40–Reference Langer, Lafaro, Genese, McDonough, Nahass and Robertson43}

Of note, during outbreaks, the use of PPE was frequently bundled with enhanced cleaning, AS, and cohort isolation of patients and/or staff. ^{Reference Podnos, Cinat, Wilson, Cooke, Gornick and Thrupp42–Reference Laurent, Rodriguez-Villalobos and Rost44} In endemic settings, many observational studies fail to show a change in ESBL-E rates following discontinuation of CP, ^{Reference Hagiya and Otsuka45,Reference Maechler, Schwab and Hansen46} In a systematic scoping review ^{Reference Hagiya and Otsuka45} of 9 studies including 1 multicenter cluster-randomized crossover trial ^{Reference Maechler, Schwab and Hansen46} and 8 observational studies, ^{Reference Domenech de Cellès, Zahar, Abadie and Guillemot47–Reference AbiGhosn, AlAsmar, Abboud, Bailey and Haddad54} discontinuation of CP for patients with ESBL-E did not result in increased transmission. While patient-to-patient transmission of ESBL-E clearly occurs in healthcare settings, ^{Reference Harris, Perencevich and Johnson40–Reference Laurent, Rodriguez-Villalobos and Rost44} CP do not seem to prevent it from happening. ^{Reference Tschudin-Sutter, Lucet, Mutters, Tacconelli, Zahar and Harbarth4,Reference Hagiya and Otsuka45}

The rationale for discontinuation of CP for ESBL-E include the following. Firstly, ESBL-E are acquired in the community as well as healthcare settings, ^{Reference Tschudin-Sutter, Lucet, Mutters, Tacconelli, Zahar and Harbarth4,Reference Ling, Peri, Furuya-Kanamori, Harris and Paterson55} with food, ^{Reference Ribeiro, Nespolo, Rossi and Fairbrother56} farm animals, ^{Reference Leverstein-van Hall, Dierikx and Cohen Stuart57,Reference Hasib, Magouras and St-Hilaire58} and international travelers ^{Reference Woerther, Andremont and Kantele59,Reference Ruppé, Armand-Lefèvre and Estellat60} found to have high rates of ESBL-E. Secondly, Enterobacterales are generally not viable in dry conditions compared to Gram-positives like MRSA and VRE, ^{Reference Hirai61,Reference Hota62} and therefore have low rates of cross transmission from index to contact patient. ^{Reference Tschudin-Sutter, Frei, Dangel, Stranden and Widmer63–Reference Valenza, Schulze and Friedrich65} Thirdly, many ESBL-E infections are related to a prior colonizing strain and facilitated in acute care by antimicrobial selective pressures, not direct transmission. ^{Reference Zahar, Poirel, Dupont, Fortineau, Nassif and Nordmann66–Reference Tosh and McDonald68}

Emerging and extensively resistant organisms

As the controversy around CP for MRSA/VRE and ESBL-E expands into the awareness of non-IPC HCPs, front-line providers increasingly question if CP are necessary for other MDROs. IPC guidelines nearly uniformly recommend CP for emerging and non-endemic MDROs such as Candida auris, carbapenem-resistant Enterobacterales (CRE), vancomycin-intermediate or -resistant Staphylococcus aureus (VISA/VRSA), and other extensively drug-resistant organisms. ^{Reference Siegel, Rhinehart, Jackson and Chiarello1,Reference Siegel, Rhinehart, Jackson and Chiarello23,Reference Tacconelli, Cataldo and Dancer39} The rationale is that these organisms are more prevalent in healthcare settings than the community (non-endemic), are difficult to treat given paucity of available effective antimicrobials, and have clear propensity to cause healthcare-related outbreaks.

While CP are widely accepted as the standard for control of these MDROs, given healthcare’s role in facilitating spread. However, the steady propagation of these organisms globally suggests CP is not very effective at the population level. Outbreaks are often multifactorial, have diverse reservoirs including the hospital environment and ongoing influx of colonized patients. The use of PPE for patient care undoubtedly prevents some person-to-person transmission, but interventions that better address all routes of transmission (ie endogenous colonizing flora, environmental reservoirs, fomites) are needed. Importantly, CP represents an easy recommendation to make but hard on HCP and patients. Calls for adequate healthcare facility staffing, resources to ensure a clean healthcare environment, ^{Reference Zuberi80–82} and equipment that is designed with cleaning and infection control issues in mind ^{82–Reference Tomlinson83} are more difficult to recommend and enforce but would be more valuable in addressing issues of transmission in hospitals. In this context, we need to identify expiration dates for CP for these organisms. While it is true that colonization with these organisms can persist for prolonged periods of time, ^{Reference Banach, Bearman and Barnden84} such that colonization is presumed indefinite, we do not know if this translates into ongoing transmission risks, particularly in the context of a contaminated environment, or if the patient recovers function and requires lower acuity of care. More data is needed to identify scenarios in which the healthcare environment itself poses greater transmission risk than the remotely colonized patient’s potential flora.

Overuse of CPs in special settings: low-hanging fruits?

Healthcare facilities caring for complex patients have at times extrapolated from acute care to employ CP for MDROs in ambulatory and home-care environments. Yet the focus of CDC/HICPAC ^{Reference Siegel, Rhinehart, Jackson and Chiarello1,Reference Siegel, Rhinehart, Jackson and Chiarello23} and AAP guidelines ^{Reference Rathore and Jackson85} for IPC in ambulatory settings is on transmission of acute, communicable illnesses, such as measles and tuberculosis. Regarding CP for MDROs, CDC/HICPAC acknowledge there is a lack of data on MDRO transmission in ambulatory settings, and it is presumed to be low, with the exceptions of chronic hemodialysis and cystic fibrosis clinics (where MDRO transmission is well-documented). The CDC’s MDRO Guideline ^{Reference Siegel, Rhinehart, Jackson and Chiarello23} explicitly recommends standard precautions for most ambulatory settings.

CP for visitors in acute care hospitals is recommended by SHEA Expert Guidance ^{Reference Munoz-Price, Banach and Bearman86} for some MDROs, such as extensively drug-resistant g organisms (ie, CRE), but not for endemic MRSA/VRE.Authors acknowledge limited evidence exists to guide recommendations.

Gloves and, occasionally, gowns are used for vaccination. Although PPE is not necessary unless there is risk of contact with body fluids or non-intact skin, ^{Reference Hutin, Hauri and Chiarello87,88} their use remain significant. In a 2018 universitywide flu vaccination campaign, about 12,000 gloves were used. ^{Reference Biederman, Hartman, Amarasekara, Schneider, Alvarez-Loayza and Brigman89} It was estimated that, if 50% of SARS-CoV-2 vaccinations were carried out using gloves, 3–5 billion pairs of gloves would be used. ⁹⁰

During the initial COVID-19 outbreak, there was uncertainty about transmission risks via direct contact and, in abundance of caution, CP were widely recommended along with droplet and/or airborne precautions. However, a growing accumulation of data suggests gowns are unnecessary, ^{Reference Rodriguez-Nava, Diekema and Salinas91,Reference Rabin, Marr and Blumberg92} citing a lack of data that the virus recovered from surfaces or fomites is viable for transmission, the rarity of documented PPE contamination with virus, and clear evidence of respiratory route spread, particularly in super-spreader events in the setting of closed spaces with stagnant air. ^{Reference Rodriguez-Nava, Diekema and Salinas91,Reference Rabin, Marr and Blumberg92} The IPC community has opportunities to be more nimble in changing PPE recommendations when new data becomes available.

Harmonizing prevention and sustainability

CP can prevent transmission of infectious disease when used optimally. However, CP come with significant costs to patients, providers, the healthcare system, and ultimately our goals of global public health (Table 1). ^{Reference Morgan, Diekema, Sepkowitz and Perencevich8–Reference Clapp98} We argue that it is specifically the gown/glove components of the CP bundle driving these costs. Could the PPE component of CP be employed more thoughtfully to maximize benefit-to-cost ratios of MDRO prevention?

Table 1. Costs of PPE use by stakeholder type

Bearman et al. ^{Reference Bearman, Harris and Tacconelli99} suggest a “precision-based approach” to CP for endemic pathogens (MRSA/VRE, ESBL-E), based on facility, host, and pathogen factors. This strategy uses CP for vulnerable patients, with organisms of high epidemiologic concern, in high acuity settings. ^{Reference Bearman, Harris and Tacconelli99} This strategy provides facilities not comfortable with complete discontinuation of CP a “middle ground” to relax CP in lower risk scenarios. ^{Reference Bearman, Harris and Tacconelli99} Recent HCP focus groups discussed using a “risk-tailored” approach to CPs, focusing on high-risk activities and settings. Most participants were open to a risk-tailored approach, but some felt it may be too complicated or confusing to implement. ^{Reference Harris and Calfee100} This was further supported by a multisite survey of a diverse group HCP of varying roles. ^{Reference Calfee and Snyder101} The downside of risk-based strategies are the increased complexity it adds to IP recommendations, creating potential for confusion on the front lines.

CDC recommends a strategy of enhanced-barrier precautions (EBP) for residents of long-term care facilities with certain MDROs in the context of contaminating activities since 2019. ^{Reference Mody, Gontjes and Cassone102,Reference Mayoryk, O’Hara and Robinson103} While there has been confusion about when and how to apply these recommendations, ^{Reference Mayoryk, O’Hara and Robinson103} EBP allows for some balance between the goal of decreasing MDRO transmission and allowing patients to participate in therapy activities. The characteristics of patients selected for EBP are similar to the high-risk patients described in precision-based CP. ^{Reference Bearman, Harris and Tacconelli99}

Beyond prioritization, CP are too frequently studied and discussed as an all-or-nothing intervention. Yet, the specific elements of CP include private rooms, dedicated medical equipment, and potentially enhanced cleaning processes, in addition to PPE. Thus, the benefit we do see from CP could be related to the co-occurring parts of the bundle and not necessary gowns/gloves. Improper use of PPE and self-contamination when doffing ^{Reference Osei-Bonsu, Masroor and Cooper104,Reference Tomas, Kundrapu and Thota105} likely further dilutes benefit. If CP were redefined to include gowns/gloves only for contaminating care (similar to EBP), along with the preservation of a private room, enhanced cleaning, etc, it is unlikely that much IPC power would be lost, and PPE costs would be greatly reduced.

Lastly, we need to reconsider how adherence to CP is measured. For simplicity, adherence is generally defined as the use of gowns/gloves for any room entry. Of note, this was not the intent of CP as first developed. ^{Reference Garner and Simmons18,Reference Garner21,} In practice, many entries and exits from the patient room are exceptionally low risk for self-contamination with MDRO from surfaces or the patient, particularly in the context of appropriate HH. If CP adherence were redefined as gowns/gloves for actual body contact with the patient and/or environment, this too would prevent overuse of low-value PPE. Many facilities have an “observation zone” just inside the room, where providers can observe the patient or check equipment settings, without interacting physically, and where gowns/gloves are not required.

The next generation of HCPs are calling loudly for new approaches to prevention of transmission of MDROs, ^{Reference Doshi, Lee, Hymes, Guzman-Cottrill and Jaggi13} that balance the costs and benefits, as is the goal of all clinicians in an effort to “First Do No Harm.” The IP community must respond with creative solutions to maximize the benefit of the PPE we use, while aggressively removing the excess. Lastly, as PPE addresses only the tip of transmission iceberg, we need to refocus efforts on standard precautions adherence, foundational IP practices of HH, and advocating for clean environments and devices for our patients.