Introduction
Practice recommendations from the Society for Healthcare Epidemiology of America (SHEA), the Infectious Diseases Society of America, and the Association for Professionals in Infection Control and Epidemiology consider contact precautions for patients colonized or infected with methicillin-resistant Staphylococcus aureus (MRSA) to be an “essential practice” for prevention of MRSA transmission in acute care hospitals. Reference Popovich, Aureden, Ham, Harris, Hessels, Huang, Maragakis, Milstone, Moody, Yokoe and Calfee1 Despite this, a recent SHEA Research Network study estimated that 35% of acute care hospitals no longer use contact precautions for MRSA. Reference Martin, Morgan, Pryor and Bearman2 Several studies have assessed the impact of discontinuing contact precautions for MRSA-colonized or infected patients, Reference Marra, Edmond, Schweizer, Ryan and Diekema3,Reference Kleyman, Cupril-Nilson, Robinson, Thakore, Haq, Chen, Oyesanmi, Browning, Pino and Mhaskar4 and many reported no change in MRSA infection rates. Reference Karunakaran, Pless, Ayres, Ciccone, Penzelik, Sundermann, Martin, Griffith, Waggle, Hodges and Harrison5–Reference Haessler, Martin, Scales, Kang, Doll, Stevens, Uslan, Pryor, Edmond, Godbout, Abbas and Bearman9 However, many of these studies were single-center, under-powered, non-randomized designs that did not assess MRSA acquisition or transmission and/or did not control for factors associated with transmission at the study population or individual patient level yet these studies are cited as rationale for discontinuing contact precautions.
There is also ongoing debate about the negative impacts of contact precautions. Some may cite possible harms of contact precautions, Reference Morgan, Diekema, Sepkowitz and Perencevich10 such as fewer visits by healthcare personnel (HCP), delays in transitions of care, patient dissatisfaction, increase in noninfectious adverse events Reference Martin, Morgan, Pryor and Bearman2 , negative impacts on mental health, and the physical discomfort and the time required for donning and doffing. However, one cluster-randomized trial showed no increase in adverse events with the use of universal contact precautions in the intensive care unit (ICU) setting, Reference Harris, Pineles, Belton, Johnson, Shardell, Loeb, Newhouse, Dembry, Braun, Perencevich and Hall11 while a single-center quasi-experimental study showed a reduction in noninfectious adverse events among patients with MRSA or vancomycin-resistant Enterococcus after contact precautions were discontinued. Reference Martin, Bryant and Grogan12 Few studies systematically documented patient or frontline HCP perspectives about contact precautions specifically. Although most hospitals still use contact precautions for MRSA, there is a growing interest in using contact precautions differently to minimize patient harms and improve staff comfort. Reference Martin, Morgan, Pryor and Bearman2,Reference O’Hara, Harris, Calfee, Snyder, Cottam, O’Hara and Martin13,Reference Bearman, Harris and Tacconelli14
To estimate the patient-level risk of MRSA transfer from patient to HCP, including whether specific HCP-patient interactions are more likely to lead to glove or gown contamination with MRSA, a previous study enrolled 402 MRSA-colonized patients in 23 ICUs at 4 hospitals and observed 3,982 HCP interactions. Reference O’Hara, Calfee, Miller, Pineles, Magder, Johnson, Morgan and Harris15 This study found that HCP gloves and gowns were contaminated with MRSA after 14.3% of interactions and that contamination was highest among occupational/physical therapists (odds ratio [OR], 6.96; 95% confidence interval [CI], 3.51, 13.79), respiratory therapists (OR, 5.34; 95% CI, 3.04, 9.39), and when any HCP touched the patient (OR, 2.59; 95% CI, 1.04, 6.51). These results suggest that the risk of MRSA transmission differs by HCP type and that transmission is higher when touching the patient when compared to touching only the environment Reference O’Hara, Harris, Calfee, Snyder, Cottam, O’Hara and Martin13 .
As compared to an “all or none” approach to contact precautions, a targeted or “risk-tailored” evidence-based implementation of contact precautions for patients with MRSA may more effectively reduce the risk of potential adverse effects while retaining the greatest potential benefits. The implementation of a risk-tailored approach to contact precautions (eg, using gloves and gowns when caring for patients with MRSA carriage only for certain high-risk patient care activities, inpatient locations, or HCP roles) would be more difficult to implement than current practice. Therefore, before trialing the effectiveness of a risk-tailored approach, it is important to assess the feasibility of such an approach by better understanding HCP risk perception, motivations for adherence to contact precautions, and policy preferences related to implementing a risk-tailored approach to contact precautions.
The “Discrete Choice Experiment” (DCE) methodology is an innovative approach to quantify individuals’ attitudes toward different policies. The DCE methodology also quantifies perceived barriers and facilitators to implementing an intervention or policy change, such as targeted contact precautions. Originally developed for economics and market research, DCEs are becoming increasingly common in healthcare evaluations. Reference Ryan and Gerard16,Reference Bridges, Kinter, Kidane, Heinzen and McCormick17 DCEs are implemented in surveys or questionnaires consisting of a series of choice sets. Each choice set (a contact precaution policy in this study) consists of two or more realistic, but hypothetical, scenarios between which respondents are asked to choose. These scenarios are described by a discrete number of attributes and each attribute takes one of several levels. Respondent choices provide information on the relative importance of the attributes and the trade-offs they are willing to make between them (eg, in this study, HCP acceptance or preference of different approaches to contact precautions based on risk factors for MRSA transmission). The exact influence of each level on their choices can be quantified through statistical modeling. Reference Ryan and Farrar18 The DCE approach allows one to assign quantitative values to preferences while assessing trade-offs and to statistically analyze these components of decision-making. Thus, the DCE provides a quantitative technique to assess different HCP choices relative to contact precautions. In this investigation, we aimed to use DCE methodology to characterize HCP opinion about a risk-tailored approach to contact precautions for MRSA based upon three attributes: Healthcare personnel, setting, and activity (Figure 1).
Attributes and levels included in choice sets in the discrete choice experiment.
Methods
Survey development and distribution
We conducted three focus groups with occupationally diverse groups of HCP to inform and refine the choice sets used in this discrete choice survey Reference Harris, Pineles, Belton, Johnson, Shardell, Loeb, Newhouse, Dembry, Braun, Perencevich and Hall11 . We then distributed the discrete choice experiment survey to HCPs at three medical centers in New York, Pittsburgh, and Baltimore. Emails with study information and the link to the survey were sent to leaders of key clinical groups and these leaders were asked to distribute the information more broadly. Paper fliers were also posted in staff lounges, conference rooms, and hallways. Study team members attended existing meetings to discuss the study and encourage participation. Of the three participating sites, two use contact precautions for patients colonized or infection with MRSA and one discontinued routine use of contact precautions for MRSA colonization and infection. This study was approved by the Institutional Review Boards at Weill Cornell Medicine, the University of Pittsburgh, and the University of Maryland, Baltimore.
The final survey was administered using Qualtrics and respondents were presented with eight questions (or choice sets) where they were asked to choose between two hypothetical policy options for glove and gown use to prevent MRSA transmission (See supplemental material for full survey). In each comparison, respondents selected their preferred policy option. As each attribute had two levels, there were 2 x 2 x 2 = 8 possible combinations and 28 possible two-alternative choice sets [C(8,2) = 8!/([8-2]! X 2!) = 28). To limit respondent burden, we constructed a partial factorial design that included 8 choice sets. This experimental design optimized for efficiency and quality, measured by D-efficiency. The survey also included several demographic questions in addition to questions assessing risk perception, motivation for compliance with infection control policies, and self-reported compliance.
Sample size
Our target sample size was based on the formula proposed by de Bekker-Grob et al., which multiplies 500 by the number of levels per attribute (n = 2) divided by the product of the number of choice sets (n = 8) and the alternatives in each choice set (n = 2). Reference de Bekker-Grob, Donkers, Jonker and Stolk19 Based on this heuristic, our design required at least 63 respondents for sufficient statistical precision for main effects estimates.
Statistical analysis
We analyzed the study participant responses using a mixed logit model. The advantage of this approach over other common options is that it allows for subject-level heterogeneity and the integration of both between- and within-subject-level variables. The base model included only the main effects of the attributes. Additional models included interaction terms for occupation and risk perception to assess heterogeneity in preferences. The model coefficients are centered on zero, reported on an absolute scale with no upper or lower bound, and represent the average utility associated with a given attribute level. We also calculated the choice probability of key policy options as the exponentiated utility of that alternative divided by the sum of exponentiated utilities across all alternatives present in the design.
Results
The characteristics of the 326 survey respondents are detailed in Table 1. Respondents’ occupational roles were physicians or advanced practice providers (80, 25%), nurses (72, 22%), environmental services personnel (49, 15%), respiratory therapists (49, 15%), occupational or physical therapists (36, 11%), and “other” (40, 12%). Occupations in the “other” category included social workers, radiology technicians, dieticians, patient service representatives, food and nutrition employees, and discharge planners. The majority of respondents (212, 65%) reported wearing gloves and gowns “all the time” when required.
Respondent characteristics (N = 326)
APP, advanced practice provider; IQR, interquartile range.
† One respondent did not report.
* Occupations in the “other” category included social workers, radiology technicians, dieticians, patient service representatives, food and nutrition employees, and discharge planners.
Risk perception and motivations for adherence
When we asked respondents whether they perceived their own MRSA contamination risk while at work as ‘Low’ (contamination occurs in 5% of interactions), ‘Moderate’ (15%), or ‘High’ (30%), occupational and physical therapists (24, 67%) respiratory therapists (27, 55%), and environmental services personnel (26, 53%) were more likely than not to perceive their risk to be “high.” Most nurses (41, 57%) and physicians or other advance practice providers (58, 73%) viewed their contamination risk as “low or moderate” (Figure 2).
Healthcare personnel perceived risk of MRSA self-contamination during care of a patient with MRSA carriage, by role.
HCPs who reported wearing gowns and gloves “all of the time” when required were highly motivated to protect themselves, their patients, and their families. There was greater variability in levels of motivation among those who reported wearing contact precautions “Most of the time” or “Some of the time.” Physicians and advance practice providers were least motivated to protect themselves, while all 49 environmental services personnel who responded to the survey ranked their motivation level as 10 (highest) to wear gloves and a gown when entering the room of a patient colonized or infected with MRSA to protect themselves, other patients, and their families (Figure 3).
Healthcare personnel self-reported motivation to protect self, patients, and own family stratified by self-reported compliance with contact precaution policies.
Acceptability and preferences for a risk-tailored policy
Respondents’ preferred policy with the highest utility score was to use gloves and gown for all HCPs roles (utility, 0.17; 95% CI, 0.12 to 0.23), in high-risk settings (utility, 0.12; 95% CI 0.07–0.18), when touching the patient (utility, 0.11; 95% CI 0.06–0.17) (Figure 4). Sixty-three percent (95% CI 60%–66%) of respondents overall would support a risk-tailored approach over an approach where contact precautions are used by all HCPs in all settings and for all activities. The support for a risk-tailored policy varied by HCP role (P < 0.02), with the strongest probability of support from physicians and advanced practice providers (77%, 95% CI 72%–82%) and the least support from environmental services personnel (45%, 95% CI 37%–53%) (Figure 5).
Estimated utility with 95% confidence intervals for each potential policy component.
Probability of support for two potential risk-tailored policies stratified by healthcare personnel role.
Discussion
The findings from this discrete choice experiment suggest that most HCPs surveyed prefer a risk-tailored approach to contact precautions when caring for patients with MRSA, rather than contact precautions for all types of patient care for patients with MRSA carriage in any inpatient care setting. Specifically, most selected a policy where gloves and a gown would be used by all HCP, but only in high-risk settings such as the ICU, emergency department, or bone marrow transplant unit, and only when touching the patient or their immediate surroundings. This approach would mean that if HCP enter a patient’s room and plans to interact with only the environment (eg, use the computer), gloves and gown would not be necessary, and that contact precautions would not be routinely used by any providers for any care activity outside of high-risk settings. A risk-tailored approach such as this has the potential to decrease costs and mitigate environmental waste while maintaining this transmission-prevention measure during interactions where risk of MRSA contamination of HCP hands and attire is highest. The predominance of HCP self-reporting of frequent contact precaution adherence when indicated (95% report “most” or “all” of the time), and the high motivation scores in this group, suggest against a compelling need to abandon contact precautions for the care of patients with MRSA carriage due to HCP preference.
This study also confirms that the vast majority of HCP are strongly motivated to protect their patients, themselves, and their families. This is consistent with previously published studies that focused on HCP motivations for adherence to hand hygiene recommendations and suggested that most HCP want to protect themselves and feel a strong responsibility to patients to prevent infections. Reference Erasmus, Brouwer, Van Beeck, Oenema, Daha, Richardus, Vos and Brug20,Reference Eveillard, Bruna, Kouatchet, Dubé, Poiroux, Dabin, Hitoto, Mercat, Joly-Guillou and Desfontaines21 HCP self-perceived risk was generally consistent with data from previous studies Reference Martin, Colaianne, Bridge, Bilderback, Tanner, Wagester, Yassin, Pontzer and Snyder8 - except among environmental services personnel respondents. Environmental services personnel perceived their MRSA contamination risk as “high” when data demonstrates that they are one of the lowest-risk occupational groups Reference Martin, Colaianne, Bridge, Bilderback, Tanner, Wagester, Yassin, Pontzer and Snyder8 . Perhaps because of this perception that their risk is high, environmental services personnel were least supportive of a risk-tailored approach to contact precautions and instead favored wearing gloves and a gown at all times when entering the room of a patient known to be colonized or infected with MRSA. Physicians and advanced practice providers may have a better baseline understanding of risk stratification and perhaps were therefore most supportive of a more complex, risk-tailored approach. If a risk-tailored approach is considered at an institution, education on risk for environmental service will likely be necessary for success as the perceived and actual risk are currently not in alignment.
This study is subject to several limitations. We sampled a large number of HCP in diverse roles in several geographically distinct settings, however, this was only a small proportion of the overall workforce and the results may not be the same in other healthcare organizations where factors such as organizational culture may vary. We compare the use of contact precautions with the risk-tailored application of contact precautions but do not directly compare HCP assessment of risk-tailored application of contact precautions to no use of contact precautions. A DCE is a validated tool to assess complex alternatives, but a survey may not match with HCP practicing the surveyed behaviors; in this case, a trial of risk-tailored approach should include re-assessment of HCP motivation and preferences.
Conclusion
This discrete choice survey demonstrates that a risk-tailored approach to contact precautions when caring for patients with MRSA may be acceptable to a majority of HCP, suggesting that this approach may be a feasible alternative to the current all-or-none approach to contact precautions for MRSA. Further research is necessary to determine optimal implementation and impact of a risk-tailored approach, including impact on transmission and on mitigating potential harms, such waste, cost, HCP visits, adverse events, and patient and HCP satisfaction.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/ice.2025.65
Acknowledgments
The authors thank Lisa Pineles for supporting this study.
Financial support
The project described was supported by cooperative agreement 5U54CK000615 from CDC. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of CDC.
Competing interests
All authors report no conflicts of interest relevant to this study.
