Article Text
PDF
Abstract
Background Relatively little is known about the cognitive processes of healthcare workers that mediate between performance-shaping factors (eg, workload, time pressure) and adherence to infection prevention and control (IPC) practices. We taxonomised the cognitive work involved in IPC practices and assessed its role in how pathogens spread.
Methods Forty-two registered nurses performed patient care tasks in a standardised high-fidelity simulation. Afterwards, participants watched a video of their simulation and described what they were thinking, which we analysed to obtain frequencies of macrocognitive functions (MCFs) in the context of different IPC practices. Performance in the simulation was the frequency at which participants spread harmless surrogates for pathogens (bacteriophages). Using a tertiary split, participants were categorised into a performance group: high, medium or low. To identify associations between the three variables—performance groups, MCFs and IPC practices—we used multiblock discriminant correspondence analysis (MUDICA).
Results MUDICA extracted two factors discriminating between performance groups. Factor 1 captured differences between high and medium performers. High performers monitored the situation for contamination events and mitigated risks by applying formal and informal rules or managing their uncertainty, particularly for sterile technique and cleaning. Medium performers engaged more in future-oriented cognition, anticipating contamination events and planning their workflow, across many IPC practices. Factor 2 distinguished the low performers from the medium and high performers who mitigated risks with informal rules and sacrificed IPC practices when managing tradeoffs, all in the context of minimising cross-contamination from physical touch.
Conclusions To reduce pathogen transmission, new approaches to training IPC (eg, cognitive skills training) and system design are needed. Interventions should help nurses apply their knowledge of IPC fluidly during patient care, prioritising and monitoring situations for risks and deciding how to mitigate risks. Planning IPC into one’s workflow is beneficial but may not account for the unpredictability of patient care.
- infection control
- human factors
- simulation
- decision making
Data availability statement
Data are available on reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Factors internal and external to healthcare workers affect adherence to infection prevention and control practices.
WHAT THIS STUDY ADDS
We described the cognitive work of practising infection prevention and control, which transforms these factors into behaviours.
We then identified the forms of cognitive work that were associated with the spread of surrogate pathogens during simulated patient care.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
The current research can be used to better support healthcare workers practising infection prevention and control by redesigning healthcare systems as well as developing a cognitive skills training to improve macrocognition, such as monitoring and decision making.
Introduction
Healthcare-associated infections affect approximately 1 in 31 hospitalised patients,1 leading to 90 000 deaths annually in the USA alone and placing a significant financial burden on healthcare systems.2 Common infection prevention and control (IPC) practices, such as hand hygiene or using personal protective equipment (PPE), could prevent up to 70% of these infections if applied effectively by healthcare workers (HCWs).3 As breaches in IPC practices are common,4 5 attention has been devoted to identifying factors, both internal and external to HCWs, that influence adherence to IPC. Internal factors include HCWs’ knowledge,6 7 perceptions6 7 and attitudes towards IPC practices.6 7 External factors include system characteristics, such as the design of the built environment,6 tasks,6 8 supplies or equipment6 and the organisation.6
Relatively less is known about the cognitive processes of HCWs that transform these factors into behaviours.9 10 For example, decision making in complex settings like healthcare often involves making tradeoffs between competing priorities, which are compelled by external factors (eg, time pressure) and resolved, in part, by internal factors (eg, perceptions of risk) to produce behaviour (eg, not performing hand hygiene). Describing these processes, or the cognitive work involved in applying IPC practices, is useful for improving adherence to IPC practices as it may help explain both positive and negative HCW behaviours. Understanding how HCWs integrate IPC into the complexities of patient care could identify forms of cognitive work that should be supported through better system design6 11 or strengthened with training.11 12
In the present study, we described the cognitive work of IPC in terms of macrocognition, which refers to the high-level cognitive processes by which people manage complex situations.10 11 Macrocognition identifies several key functions,11 13 14 which include decision making,13 15 planning,13 sensemaking14 and monitoring.13 14 Because macrocognition is a generic framework, however, we sought to first describe these macrocognitive functions (MCFs) in the context of IPC. Second, we sought to understand how these forms of cognitive work relate to the spread of pathogens during simulated patient care.
Methods
As the goal of IPC is to prevent the spread of pathogens, we measured HCW performance in terms of the frequency at which they spread surrogate pathogens to surfaces in a challenging high-fidelity simulation. To connect performance with cognitive processes, we performed a cognitive task analysis, which is a systematic process for capturing and eliciting the cognitive processes underlying task performance.11 We elicited these processes by asking HCWs to report their thoughts while watching a video recording of their simulation (ie, a think-aloud) to avoid interfering with task performance.11
Participants
As nurses comprise the largest group of HCWs and have frequent contact with patients, we recruited 45 registered nurses from a tertiary/quaternary academic medical centre, using flyers and unit announcements. Inclusion criteria were being at least 18 years of age and working as a registered nurse in an emergency department, intensive care unit or medical/surgical unit. We excluded two participants because of equipment issues, and one participant was excluded for not completing the simulation. The remaining 42 participants (median=4.89 years since first licensure, IQR 2.32–18.23) came from the intensive care unit (n=13), emergency department (n=12) and medical/surgical unit (n=17). For 2.5 hours, participants were compensated with US$125.
Simulation design
Simulations were conducted at Emory University School of Medicine's simulation centre. We developed a standardised simulation scenario involving two male adult inpatients (Laerdal SimMan high-fidelity patient manikins), in separate rooms, who required only standard precautions (see online supplemental material for patient descriptions and simulation layout). Based on prior research,16 we selected tasks that were high or low in the perceived risk of infection to the patient and high or low in the perceived risk of HCW contamination.17 Patient 1’s tasks were changing the dressing on a stage 4 sacral pressure injury, toileting with a bedpan, inserting a nasogastric tube and inserting a peripheral intravenous line (PIV). Patient 2’s tasks were inserting an indwelling urinary catheter, collecting a stool specimen, auscultating heart and breath sounds and administering an intravenous medication. Manikins could converse with participants but could not move themselves.
Supplemental material
Tasks were documented in a simulated electronic medical record that was accessed from a mobile workstation-on-wheels. Additionally, simulations contained common challenges to applying IPC practices8 18: participants did not receive assistance from other staff (short-staffed), were instructed they had 1 hour to complete and document tasks (time-pressure) and were interrupted twice in each patient room at specific moments during tasks (see online supplemental material for details). PPE and products for cleaning patients and environmental disinfection were available in the patient’s supply cart. Alcohol-based hand rub dispensers were located on entry and exit of patient rooms.
Procedure
Before simulations, four variants of bacteriophage λ were applied to unique target sites19: λTemp on patient 1’s stage 4 pressure injury, λKan on patient 1’s stool in a bedpan beneath the patient, λChl on patient 2’s genitalia and λVir on patient 2’s stool on a disposable underpad beneath the patient.20 Two of the four tasks for each patient harboured a different source of contamination: changing the dressing on pressure injury (λTemp), toileting in bed (λKan), inserting a Foley catheter (λChl) and collecting a stool specimen (λVir).
Before the start of the simulation, a researcher consented participants and collected demographic data, such as experience and area of practice. Participants understood that there may be contamination in each room but were unaware of its locations, sampling sites and that we were only interested in IPC. Participants donned a pair of disposable scrubs, and the researcher then oriented them to the simulation space, equipment and supplies, high-fidelity manikins and tasks by walking through the supply room, two patient rooms and hallway between them. To minimise observation effects, we explained that the simulation was an opportunity to learn how to improve clinical practices based on the aggregate performance of participants. A researcher then placed a GoPro HERO8 mobile camera on the participant’s forehead outside of the patient rooms, after which a clinical-researcher gave a scripted handover for each patient at the bedside. Participants then performed the eight tasks in any order they wished. Although participants were told they had 1 hour to complete the simulation, we allowed participants to exceed 1 hour to complete all tasks.
After the simulation, participants performed a cued-retrospective think-aloud in a separate debriefing room,21 22 in which they described what they were thinking about in the simulation while watching a video playback from their point of view. The think-aloud and playback were recorded synchronously. While the participant performed the think-aloud, we sampled 16 sites from the simulation grouped into 4 types of surfaces: high-touch surfaces, workstation-on-wheels surfaces, nurse surfaces and patient critical sites. High-touch surfaces included the bedside table, bedrails, bedrail controls, vital signs monitor and supply cart in each patient room.23 Workstation-on-wheels surfaces included the keyboard and table.24 Nurse surfaces included the nurse’s bare hands and disposable scrubs.25 26 Patient critical sites included patient 1’s sacral pressure injury and patient 2’s genitalia. Each surface was sampled for the presence of each λ-variant. After sampling, surfaces were disinfected with 70% ethanol, DNA Away, RNase Away and Lysol wipes. Phage identification of positive surfaces samples was determined by direct PCR testing. Online supplemental figures S1 and S2 show a flow diagram of the procedure and layout of the simulation space, respectively.
Analysis
Measuring performance
For each of the four surface types—(1) high-touch surfaces, (2) surfaces on the workstation-on-wheels, (3) surfaces on the nurse and (4) patient critical sites—we calculated the percentage of surfaces that became contaminated with bacteriophage during the simulation. We then calculated an overall performance score that was a weighted average of the four surface types with weights reflecting the severity of contaminating each type of surface. We obtained weights using the Analytical Hierarchy Process,27 where subject matter experts, including two infection preventionists, two infectious disease physicians and two nurses from a biocontainment unit, rated in a pairwise fashion the relative importance of each surface type on a standard 9-point scale in contributing to the likelihood of a patient becoming colonised with a bacterial pathogen while in the hospital. Ratings were aggregated using the geometric mean, and weights were derived using the eigenvalue method28 29 in the ahpsurvey package in R.30 31 Contaminating patient critical sites had the greatest weight (50.49%), followed by self-contamination (26.09%), high-touch surfaces (14.88%) and surfaces on the workstation-on-wheels (8.54%).
Segmenting think-alouds
Think-alouds were transcribed into verbalisations, which were complete thoughts separated by a significant pause. Two nurses independently coded verbalisations for whether the participant was clearly thinking about IPC, which was defined as whether a verbalisation either described recognising a risk related to IPC or referred to an IPC practice. A third nurse resolved disagreements. Inter-rater reliability (IRR) was assessed with Gwet’s agreement coefficient 1 (AC1).32 IRR was nearly perfect (AC1=0.91).33
Content analyses of think-alouds
Our goals were, first, to verify the involvement of MCFs in IPC practices and, second, to describe MCFs in the context of IPC. Thus, analyses of the think-alouds comprised two phases: a deductive content analysis, in which we categorised verbalisations in terms of MCFs,11 followed by an inductive content analysis, in which we identified and assigned verbalisations to subcategories of MCFs derived from the think-aloud data.34 Frequencies of subcategories were obtained for subsequent analysis.
In the deductive content analysis, two human factors experts independently coded verbalisations for whether they involved each of the four MCFs (table 1). IRR for each MCF was based on 14 randomly selected transcripts; the remaining transcripts were divided equally between the two coders. Final codes for dually coded transcripts came equally from both coders. In the inductive content analysis, verbalisations involving any MCF were written on separate cards and grouped according to MCF. To identify subcategories of each MCF, two human factors experts and three nurses independently performed open card sorts of verbalisations,35 36 where each sorter grouped similar verbalisations together and described their groupings. Sorters then discussed their groupings to arrive at a consensual set of subcategories and their definitions. To measure IRR, and obtain frequencies of these subcategories, five nurses (two of whom did not develop the coding scheme) performed closed card sorts,35 where verbalisations were sorted into the subcategories. The final assignment of a verbalisation to a subcategory reflected the simple majority of the five sorters’ judgements for a verbalisation. Open and closed card sorts were performed with the Proven By Users online card-sorting tool.37
Macrocognitive functions (MCFs)’ definitions, inter-rater reliability (IRR) and prevalence
Lastly, verbalisations involving any MCF subcategory were coded for the IPC practices they involved: hand hygiene; cleaning the patient, environment or equipment; achieving or maintaining sterility; PPE; disposing or containing potentially contaminated material; or minimising contamination via contact (see online supplemental material Table S1 for definitions). One human factors expert and one nurse independently performed a closed card sort of verbalisations (see ‘IPC practices’ section for agreement statistics). The final assignment of a verbalisation to an IPC practice(s) reflected the agreement of the two coders; a second human factors expert resolved disagreements.
Multiblock discriminant correspondence analysis
To identify associations between performance groups, MCFs subcategories and IPC practices, we used multiblock discriminant correspondence analysis (MUDICA).38 39 MUDICA is a multivariate statistical technique for identifying variables that are important for discriminating between groups of individuals. MUDICA is non-parametric and is appropriate when the number of variables exceeds the number of observations.39
In MUDICA, the extent to which the co-occurrences of groups and variables deviate from independence is akin to the total variance of the data,40 which MUDICA decomposes into a set of orthogonal factors along which groups and variables are represented as factor scores; the first factor explains the greatest amount of variance (its eigenvalue, λ), the second factor explains the greatest amount of residual variance and so forth. Each factor maximises between group differences, capturing important and unique associations between these groups and variables. We performed MUDICA using R.38 41
MUDICA analyses a group by variable contingency table, where groups were the three performance groups and variables were formed by crossing the 19 MCF subcategories (table 1) with the 6 IPC practices. Variables contained the frequencies at which each MCF subcategory occurred in the context of each IPC practice for each performance group. After removing variables occurring less than twice, 83 variables remained for analysis. MUDICA analyses the contributions of groups and variables, and the contributions of blocks of related variables, which collectively offer more information than individual variables.38 39 Blocks were the 19 MCF subcategories. To ensure that the contribution of a block was unrelated to the number of variables comprising it (median=5 variables; range=1–6), we used a normalisation procedure in which the frequencies in a block were divided by the number of variables in that block prior to analysis (Krishnan, personal communication).
Results
HCW performance
Of all contaminated sampling sites, high-touch surfaces were most common (47%), followed by surfaces on the nurse (26%), patient critical sites (16%) and the workstation-on-wheels (11%). Overall performance scores ranged from 38% to 100% (M=77%, SD=15.3%), with higher scores indicating better performance (figure 1). We rank ordered participants according to their overall performance score and categorised them into high (n=13), medium (n=16) or low (n=13) performers using a tertiary split. Group size varied slightly due to ties. For most surface types, the average percentage of contaminated sampling sites per participant was lowest for the highest performers: contaminating a critical site on a patient (high performers=0%, medium=22%, low=42%), self-contamination (high=19%, medium=31%, low=58%) and the workstation-on-wheels (high=4%, medium=9%, low=31%). Contaminating a high-touch surface, however, was similar between groups (high=11%, medium=14%, low=14%). Overall performance was not significantly associated with experience (ie, years since licensure; rs
=0.07, p=0.64), unit type (F=0.89, p=0.42, 
=0.04) nor time to complete the simulation (one participant removed due to technical difficulties; M=65 min, SD=8, rs
=0.16, p=0.33).
The overall performance score for each nurse is shown in green. The other coloured bars reflect the weighted percentage of contaminated sampling sites for each surface type, which detract from one’s overall performance score. ED, emergency department; ICU, intensive care unit; MS, medical/surgical; WoW, workstation-on-wheels.
Content analyses of think-alouds
In the deductive content analysis, IRR of coding MCFs ranged from moderate to nearly perfect (AC1 range=0.59–0.88, median=0.74; table 1). From a total of 973 verbalisations, 598 (61%) described at least one MCF; the remaining verbalisations did not reflect participants’ thinking during the simulation (eg, comments based on hindsight or digressions). Of the 598 verbalisations, 49% described decision making, 44% monitoring, 17% sensemaking and 13% planning. Across the four MCFs, inductive content analysis produced 19 subcategories of MCFs, with IRR ranging from fair to nearly perfect (AC1 range=0.30–0.96, median=0.71; table 1). Table 1 shows percentage of verbalisations categorised into MCF subcategories. For each MCF, nearly all verbalisations involved at least one subcategory of that MCF (verbalisations could be categorised as ‘other’; see table 1 note): monitoring=99% of verbalisations, planning=99%, sensemaking=94% and decision making=90%.
IPC practices
IRR of coding IPC practices (see online supplemental material for definitions) ranged from substantial to nearly perfect33: minimising contamination from contact (AC1=0.62; 16% of verbalisations); cleaning the patient, equipment or environment (AC1=0.85; 18%); disposing or containing potentially contaminated material (AC1=0.88; 9%); achieving or maintaining sterility (AC1=0.89; 33%); using PPE (AC1=0.89; 19%) and hand hygiene (AC1=0.95; 7%). Ninety-five per cent of MCF subcategories occurred (at least twice) in the context of achieving or maintaining sterility, 84% minimising contamination from contact (‘contact contamination’), 79% cleaning the patient, equipment or environment, 79% using PPE, 58% disposing or containing potentially contaminated material and 42% hand hygiene.
Multiblock discriminant correspondence analysis
MUDICA extracted two factors, which explained 52.1% (λ=0.15) and 47.9% (λ=0.14) of the variance in the contingency table, respectively. A permutation test with 10 000 permutations38 showed that a significant proportion of the variance was driven by differences between performance groups, R2=0.66 (pperm<0.001); all groups differed significantly from one another in the factor space (figure 2) as indicated by non-overlapping 95% confidence ellipses based on 1000 bootstrapped resamples.38 We interpret each factor based on the groups, blocks and the variables comprising the blocks that occupy the positive or negative side of a factor and contribute substantially to the variance of that factor (see figures 3 and 4). A substantial contribution is one that is larger than the average contribution (ie, 1 divided by the number of elements38 42): 
for groups, 
(5.26%) for blocks and 
(1.20%) for variables.
Performance group centres, 95% confidence ellipses for performance group centres and individuals in the factor space. Asterisks are the centres of the performance groups, which are encompassed by 95% confidence ellipses. Points are individual participants projected into the factor space and then coloured according to their performance group. Because no confidence ellipses overlap, all groups differed significantly on each factor at p<0.05 level.
Important signed contributions to factor 1 from groups, blocks and variables. Contributions are the percentage of the variance of a factor accounted for by a given group, block or variable. The sign of a contribution indicates whether an element contributes to the positive or negative side of a factor. Dashed lines indicate the threshold for a substantial contribution; subthreshold blocks and variables not shown. The colours of variables in C correspond to the colours of the blocks in B to which they belong; grey bars in C indicate variables that did not belong to a block in B. IPC, infection prevention and control; PPE, personal protective equipment.
Important signed contributions to factor 2 from groups, blocks and variables. Contributions are the percentage of the variance of a factor accounted for by a given group, block or variable. The sign of a contribution indicates whether an element contributes to the positive or negative side of a factor. Dashed lines indicate the threshold for a substantial contribution; subthreshold blocks and variables not shown. The colours of variables in C correspond to the colours of the blocks in B to which they belong; gray bars in C indicate variables that did not belong to a block in B. IPC, infection prevention and control; PPE, personal protective equipment.
Factor 1 reflected differences between the high and medium performance groups (figure 3A). Contrary to medium performers, high performers were distinguished by how they monitored and made decisions during patient care (figure 3B), primarily in the context of achieving or maintaining sterility and when cleaning the patient, environment or equipment (figure 3C). Specifically, high performers were more often recognising risks as situations unfolded, such as detecting contamination events during sterile technique. When deciding how to mitigate risks, high performers more often applied both formal rules (eg, cleaning stethoscope before use, cleaning PIV catheter hub before medication administration) as well as informal rules (eg, putting an extra underpad beneath patient during wound care to facilitate cleaning, bringing an extra pair of sterile gloves for Foley catheter insertion). Lastly, high performers were more often aware of a lack of knowledge, and when unable to resolve their uncertainty about what to do, they devised a cautious approach (eg, treating wound care as a sterile procedure when uncertain, avoiding using a potentially contaminated bedside table for PIV insertion because of insufficient time to clean).
Contrary to high performers, medium performers were distinguished by how they engaged more in forms of future-oriented cognition: anticipating contamination events (eg, pre-emptively double-gloving for tasks involving bodily fluids) and preparing for tasks (eg, placing Foley catheter insertion kit where it will not become contaminated during procedure), planning the sequence of tasks (eg, performing clean before dirty tasks) and planning the sequence of steps of tasks (eg, mentally reviewing the steps for inserting a Foley catheter; figure 3B). Medium performers engaged in these forms of sensemaking and planning across a variety of IPC practices, including achieving or maintaining sterility, minimising contamination by contact and using PPE (figure 3C).
Factor 2 reflects associations that distinguish the low performers from the medium and high performers (figure 4A). Low performers were characterised by how they monitored and responded to the ongoing situation (figure 4B), all of which was done in the context of minimising contamination via contact (figure 4C). Specifically, low performers were more often tracking what they touched and detecting contamination events. They mitigated risks by more often applying informal rules (eg, avoiding touching objects after touching patients) and deciding to sacrifice IPC practices when constraints or competing priorities arose (eg, not cleaning a bedside table because of fatigue, not cleaning the PIV insertion site because of time pressure). Moreover, low performers were also characterised by how they neglected to assess certain aspects of situations (ie, whether a surface is sterile, clean or dirty and identifying appropriate IPC practices in a situation) as well as manage their uncertainty (ie, proceed cautiously when uncertain; figure 4B), either in the context of achieving or maintaining sterility or using PPE (figure 4B).
Discussion
Healthcare-associated infections have devastating effects2 yet are largely preventable with common IPC practices.3 In the present study, we measured how successfully HCWs prevented the spread of surrogate pathogens to surfaces during simulated patient care. Contamination spread in simulation resembled that during actual patient care, with high-touch surfaces43 and the HCW44 being the most commonly contaminated types of surfaces. Importantly, how HCWs spread contamination was related to how they engaged in forms of macrocognition.
High performers actively monitored the unfolding situation for risks, like contamination events, and mitigated risks by applying formal and informal rules or proceeding conservatively when uncertain, mainly in the context of sterile technique and cleaning. Concurrently, no high performers contaminated a critical site on a patient, which was the most influential outcome determining one’s overall performance. Critical sites comprised one patient’s genitalia, with which participants interacted in a sterile manner during Foley catheter insertion, and another patient’s pressure injury, which participants cleaned during wound care. Both critical sites were also proximal to contaminated stool, which HCWs cleaned after collecting a specimen and toileting the patient with a bedpan, respectively. Contrary to high performers, medium performers engaged less in present-oriented and more in future-oriented cognition, trying to anticipate contamination events and plan their workflow accordingly. Although planning one’s workflow can reduce opportunities for cross-contamination,5 45 46 the complexity and unpredictability of the patient care we simulated created risks that were difficult to foresee and mitigate.47
The opposing cognitive styles of medium and high performers align with the ‘competent’ and ‘proficient’ stages of Benner’s model of skill development in nursing, respectively.47–49 In the competent stage (cf. medium performers), nurses manage the complexity of patient care by limiting the unexpected through anticipation and planning.48 However, dutifully following plans can hamper the ability to notice and respond to changes in clinical situations.50 Progressing from competent to proficient requires a qualitative leap to a flexible, response-based approach to delivering patient care.48 49 In the proficient stage (cf. high performers), anticipation and planning occurs more automatically for nurses,51 who now direct their attention to noticing important changes in the situation and deciding how to respond (eg, with rules).49 Although experience is necessary for achieving proficiency, Benner48 (and others52) argue that it is the amount of experiential learning, not simply the sheer amount of experience, that drives progression.
Low performers were distinguished from medium and high performers along a different dimension entirely. Although low performers were present-oriented, they mitigated risks ineffectively by often applying informal rules and sacrificing IPC practices when competing priorities or constraints arose. Moreover, low performers primarily monitored and responded to situations to avoid spreading contamination by physical touch (cf. high performers’ focus on sterile technique and cleaning). Minimising contamination from contact depends on accurately assessing whether a surface is contaminated. However, low performers notably under-reported assessing whether surfaces were sterile, clean or dirty. Moreover, physical contact with the patient and their environment often occurs inadvertently, outside of the HCW’s awareness.4 The cognitive style of low performers seems to reflect a misguided belief system that deprioritises IPC (eg, sacrificing IPC practices) and holds that simply avoiding spreading contamination through touch4 and following informal rules will achieve acceptable performance.
Differences between performance groups suggest that tailored training approaches are needed; low performers appear to have a belief system that should first be challenged before being restructured around fundamentals of IPC. Providing visual feedback with fluorescent lotion or powder during simulated patient care is one approach for revealing trainee misconceptions about and lapses in IPC practices.53 54 Rather than addressing a knowledge deficit, training for medium performers should instead focus on applying knowledge more fluidly during patient care via cognitive skills training, such as decision-making games.47 The ShadowBox method is one such approach for improving the monitoring and decision-making skills of less proficient practitioners.12 In this approach, trainees are given a scenario involving a work-related challenge, which stops at predetermined decision points to allow the trainee to prioritise goals, possible actions or cues to monitor. The trainee is then provided an expert panel’s rank ordering and rationale with which to compare with their own thinking.12 Prevailing approaches to training IPC focus on imparting factual and procedural knowledge via lectures, demonstrations or written materials.55 Cognitive skills training complements these approaches and provides the kind of experiential learning that nurses need to progress from the ‘competent’ to the ‘proficient’ stage of performance.47
Complementary to training efforts, understanding the cognitive work of IPC can also inform interventions for allocating MCFs, such as monitoring, across people or technology during IPC practices.11 For example, high performers were focused on detecting breaches in their sterile technique and, simultaneously, did not contaminate the patient’s genitalia during Foley catheter insertion. Monitoring, therefore, appears to be an important form of cognitive work for achieving and maintaining sterility. In the context of sterile procedures such as central line or Foley catheter insertion, the use of two-person teams has reduced central line-associated bloodstream infections56 57 and catheter-associated urinary tract infection58 in part by distributing the cognitive work of monitoring across multiple HCWs; by having one person monitor the person inserting the catheter and stop the procedure if sterility is broken, some of the cognitive work of the person inserting the catheter is offloaded.
The present study has several limitations. Regarding our measures of cognition, participants described their thought processes, which does not capture aspects of cognition that occur outside one’s awareness, such as executing highly practiced behaviours.59 Even so, the think-aloud method is useful for capturing aspects of cognition that require effortful, conscious processing (eg, decision making),60–62 which aligns with our goal of identifying forms of cognitive work. Our performance measure is also limited; although the surfaces we contaminated or sampled were grounded in the literature,19 23–26 we could not exhaustively sample surfaces comprising each type of surface nor could we exhaustively contaminate surfaces given the limited number of bacteriophage λ variants.20 Additionally, participants who performed highly (or poorly) may have done so at the expense (or benefit) of other facets of performance, such as patient stability, patient satisfaction or efficiency.17
The present study taxonomised the cognitive work of performing common IPC practices and demonstrated that the forms of work in which one engages is associated with how pathogens disseminate during simulated patient care. Differences between performance groups suggest new targets and approaches for training HCWs in IPC as well as system design. Interventions should help nurses apply their knowledge of IPC fluidly during patient care (cf. high performers) by prioritising and monitoring situations for risks (eg, during sterile technique) and deciding how to mitigate risks with formal rules or principles. The benefits of taking a primarily proactive approach to practising IPC are limited as plans may need to be adapted or abandoned to account for the dynamics of patient care.50 Future research should confirm our results in different institutions and patient populations and investigate the origins of cognitive differences, particularly with respect to the role of internal factors (eg, HCW beliefs and types of experience).7
Data availability statement
Data are available on reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
This study was approved by Emory University Institutional Review Board (#00117289). Participants gave informed consent to participate in the study before taking part.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors JMM: guarantor, funding acquisition, supervision, conceptualisation, methodology, writing—original draft, review and editing. BW: formal analysis, writing—original draft, review and editing, visualisation. JM: conceptualisation, methodology, writing—review and editing. GG: conceptualisation, methodology, writing—review and editing. PRG: conceptualisation, methodology, writing—review and editing. KBB: methodology, investigation, writing—review and editing. BAB: methodology, investigation, writing—review and editing. REMK: project administration, supervision, writing—review and editing. LL: conceptualisation, methodology. CSK: funding acquisition, conceptualisation, writing—review and editing.
Funding This study was funded by Centers for Disease Control and Prevention (grant number: NU38CK000481).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.