The last decade increasingly has recognised the contribution of iatrogenic injury to morbidity, mortality and healthcare costs.1^–3 Risk of such injury is particularly high in the intensive care unit (ICU) where the patient’s health is already compromised.4 This increased risk is, in part, a function of the complexity of the care where multiple decisions and frequent changes in therapy, patients, conditions, and personnel interact.356 Injury prevention in complex environments such as the ICU relies heavily on the presence of a safety culture.357^–10

Safety culture is “the product of individual and group values, attitudes, perceptions, competencies and patterns of behaviour that determine the commitment to an organisation’s safety management.11 The safety climate, while occasionally used interchangeably with culture, is more of a “snapshot” of the culture and focuses on attitudes of members.1112 The medical literature approaches safety culture via different conceptual frameworks. Some approaches examine specific, task-oriented behaviours such as properly applying patient restraints and locking medication carts,5 while others use a conceptual framework of general statements about patient, task, team and working conditions.3 Working or functional definitions of safety culture come from the “subculture” of a unit, and have more to do with job requirements, training and experience than with the espoused ideal safety culture.13

Interventions implemented to improve the safety of ICUs2357 generally use for outcomes either a specific clinical outcome (eg, nosocomial infection rates) or results from a standardised safety climate survey. But how ICU nurses characterise patient safety within their subculture is really unknown,814^–16 as is the effect of “successful” safety interventions on nurses’ perceptions of safety. In order to better understand these issues, we conducted focus-group interviews following implementation of a project that reduced catheter-related bloodstream infections in ICUs at four community hospitals. We triangulated the focus-group data with additional data from safety climate surveys and review of existing hospital checklists used by management leadership on walk rounds in the four hospitals. Using this qualitative approach, we explored how nursing staff in a unit that participated in an intensive, but highly specific safety project, now perceive patient safety in their unit.

METHODS

Overview of project to decrease hospital-acquired infections

The project has previously been described.17 Briefly, 10 Midwestern hospitals enlisted through a local health council participated in a 2-year project to implement evidence-based practices to reduce hospital acquired catheter-related bloodstream infections (CRBSI) in the intensive care unit and train teams in strategies to facilitate change. Through a process of team meetings and work/learning/reporting cycles, hospital teams used (1) forcing functions (a central line cart), (2) a checklist completed by ICU nurses during insertion of the central line by physicians and (3) feedback of process adherence and infection rates to significantly reduce CRBSI.17 Approval by local institutional review boards (IRB) at each hospital was obtained.

Focus groups

In the second year of the project, after the successful reduction in CRBSI, ICU nurses from a convenience sample of four hospitals participated in focus groups to discuss patient safety. We chose a focus-group format, since group discussion and dynamics increase the ability to access multiple perspectives and can lead to a broader and richer description of a topic than a series of individual interviews. In addition, focus groups have been found to be especially useful when investigating complex behaviours and motivations.18^–20 We held two focus groups at each of the four participating hospitals.

Participants

We recruited ICU nurses at the participating hospitals by posters, announcements and personal appeals by nursing leadership. Participation was voluntary for the nurses, and on their own time before and after shifts. In order to facilitate frank discussion about each organisation, we included only staff and charge nurses in the groups. (Charge nurses are experienced nurses who rotate responsibility for managing workload staffing and each shift’s patient care needs)

Interview guide

The study team developed the interview guide based on the medical literature and existing safety climate surveys2122 and then pilot-tested the guide for clarity and understanding with ICU nurses at a non-participating hospital. The interview guide posed a number of discussion probes, including (1) what is patient safety on this unit—give examples; (2) tell a story of an error, adverse event or near miss; (3) what parts of your job are patient safety; and (4) how well does administration prioritise safety on the unit? The interviews were moderated by NCE, an experienced focus-group researcher, and observed by at least one other research team member, who took notes on non-verbal communication. All interviews were audiotaped.

Analysis

Audiotapes were transcribed and reviewed for accuracy, and all identifying information deleted. All three qualitative analysts (NCE, SMB, IK) coded one transcript using the editing method,2324 in which the categories were derived from the data themselves, and then met to review the initial coding categories and come to agreement. Then, one author (NCE) coded all the transcripts using NVivo 2.0 software, into categories derived from the data, explicitly checking them against other categories and the original data, and searching for patterns. Each transcript was also read and reviewed by one additional research analyst (either SMB or IK); during a series of discussions, the coding for the additional seven transcripts was discussed and agreement reached by consensus.

Additional data sources

“Safety checklists” from management leadership walk rounds

An emerging practice in hospitals to highlight the importance of patient safety is the use of management walk rounds by senior leadership.11 These rounds are one method by which leadership conveys to front-line staff the importance of safety and the commitment of management to safety. A written checklist guides these walk rounds at the four study hospitals, but since two of the study hospitals used the same list, we obtained a total of three checklists. Using the categories derived from the focus-group analysis, we assigned the content areas of these safety forms to an existing category, and added new categories as needed, if not previously mentioned by the nurses in the focus groups. The findings from these data were compared with nurses’ perceptions of safety from the focus groups.

Safety climate survey

In the second year of the project, a random sample of nurses, ancillary personal and hospital leadership at nine of the 10 study hospitals completed two previously published and validated surveys (the Institute for Healthcare Improvement safety climate survey;22 and the Agency for Healthcare Research and Quality hospital survey on patient safety culture21). We report here only findings from staff and charge nurses from the four hospitals that also participated in the focus-group interviews. The return rate for this group was 40%. Given the staggered nature of the intervention, survey 1 was completed 5–15 months and survey 2, 15–24 months after beginning the project. Samples were not matched but rather independent for each survey administration. Due to a non-significant independent groups t test, we combined survey results from these two time periods. Neither of these surveys addresses specific components of safety, but they both have several questions on respondents’ perceptions of management’s commitment and role in safety. We used these findings to augment the focus-group discussion on this topic.

Overall analysis

The entire research team read, reviewed and discussed the original and summary data from all three sources (survey, focus group and safety checklists). The research team included two physicians (primary care (NCE) and intensivist (MLR)) and a nurse (SMB) with research interests in patient safety, an organisational psychologist (MN) and a psychometrician (IK)—this diversity of expertise adds to the “trustworthiness” of our analysis.2526 Over a series of meetings, we looked for themes and models of interactions between nurses’ definitions of safety, stories of error, and their perception of managements’ commitment to safety. The data were also reviewed within the context of adult learning theories to form a cohesive explanatory model

RESULTS

Participants

All the nurses in this study worked in ICUs that had recently participated in a successful project to decrease CRBSI, and all the ICUs had permanently incorporated the projects’ changes into the daily work of central line placement.

Focus groups

A total of 33 nurses (two men, 31 women) attended the eight focus groups, and each group had between two and five people attending. Three groups contained charge and staff nurses, while five groups had only staff nurses.

Surveys

Of the 92 completed surveys from the four hospitals, staff nurses completed 78 and nurse managers 14. Completed surveys were nearly equally distributed between the two survey periods (48% first administration, 52% from the second). The survey respondents were largely middle-aged (75% between 35 and 55 years) and had worked in their positions for more than a decade (53% for >12 years) while a minority (14%) had been in their positions less than 3 years.

Main findings

From the focus groups, we identified three main categories related to nurses’ current perceptions of safety: characterisations of patient safety, safety tasks associated with work and perceptions of management’s prioritisation of safety. We found that the perception of nurses regarding patient safety problems match that of the checklists used by management leadership during walk rounds, where patient safety related to environmental issues such as alarms and restraints and the safe administration of medications is stressed. Nurses are ambivalent about managements’ commitment to safety, giving higher marks for group education and training, but lower marks for staffing decisions.

Characterisation of patient safety

Despite the recent intervention to reduce infection and improve safety of central line placement, of the seven categories that emerged from nurses’ stories and definitions of safety, infection was one of the most rarely mentioned categories (table 1); in fact, only environmental safety, medication and general safety comments were mentioned in every group. Most of the comments described environmental safety (which included the use of bed rails, alarms, patient falls, extubation, patient visibility, restraints, equipment, and patient self-harm) and medication safety. For example, one nurse described patient safety as:

It is medication injuries, or physical injuries that could have been prevented, or anything of that nature. Or even if a patient gets injured at their own hand; like if they’re pulling out lines and bleeding all over the place, or extubating themselves, stuff like that. That’s what I think about with patient safety.

General comments about safety, such as “It’s about making sure that they (patients) get the right care,” were also mentioned.

Nurses infrequently mentioned issues related to laboratory, staffing or infection. A nurse hesitatingly noted, “I guess infection would fall under that too. I don’t know. I don’t jump to that immediately.” In fact, one participant, when asked specifically about safety related to the CRBSI project noted:

I think we look at safety two different ways, though. When you look at some safety like sepsis…that’s kind of a passive safety you’re not really seeing things until maybe after the fact, months down the road and you get your results. Okay the patient didn’t get something, versus someone who is combative or in DTs (delirium tremens) …You (or the patient) fall … one minute they’re fine, the next minute they get up and they fall from the bed. I think its two different safety issues or ideas, you know, maybe mild safety to severe safety.

One way nurses learn about how their hospital management prioritises patient safety is from what their leaders look for during leadership walk rounds. Categories of safety from checklists used during the administration walk rounds mirrored those expressed as important in the focus groups (table 2). Hospitals 3 and 4 used the same brief list of questions, but hospitals 1 and 2 each had their own checklists. Almost half (49% and 47%) of those checklist items were about environmental and medication safety.

Safety tasks associated with work

The safety tasks identified by nurses matched the perception of patient safety elicited by the focus groups. The chief safety task mentioned involved individual self initiated work to double check and monitor (table 3). Of the 28 comments made about this safety task, 25 (89%) concerned physical aspects of environmental safety (equipment, restraints, alarms, etc) or medications. For example, a nurse described her safety tasks as:

When we get on shift and going in there and double checking my drugs—making the calculations … Making sure that everything is labelled, and I know what’s getting to this person; Checking the bag and making sure it’s the right medication that’s being hung; Checking my restraints making sure they’re tied down and nobody’s going to extubate themselves.

Nurses also noted, but to a lesser extent, that their roles as patient advocate and educator served a patient safety function.

Nurses in the focus groups frequently noted that the patient’s well-being was their highest priority in their job, and maintaining this focus led to better patient safety. Participants believed that the abilities of the staff and their commitment to the job of nursing lessened errors and patient harm. Nurses expressed positive emotions around the safety successes they attributed to their personal abilities and talents. For example, a nurse described job satisfaction as, “If you catch something and feel like, ‘I made a difference in that patient’s outcome,’ you know before they coded …” Here again is an example of patient safety driven by individual knowledge and motivation.

Perceptions of management’s commitment to safety

Focus-group participants expressed no consensus regarding the commitment of nursing and hospital management to patient safety. Although all focus-group participants were asked to rate management’s commitment on a 1–10 scale, only 20 of the 33 nurses did so—many participants refused to commit to a response. While there were two ratings of 2 (low) and one of 10 (high), the other 17 were between 5 and 8. This lukewarm assessment was seen in comments like, “But mostly I think a 5, kind of like we do all the work to do it but then we don’t see any follow up.” Other nurses were more positive, “It isn’t constantly in our face but they certainly make sure that we have mechanisms and policies and everything in place that we are expected to follow.”

The nurses felt that management communicated safety to the unit in four ways: group activities, like in-service classes and talks (at four (100%) of the hospitals, 12 total times mentioned), staffing decisions (at four (100%) of the hospitals, seven total times mentioned), written notices and policies (at three (75%) of the hospitals, 8 total times mentioned) and management checks (at one (25%) of the hospitals, three total times mentioned). Most of the comments about the group activities and written notices were either positive or neutral. There were three stories (from two hospitals) of unitwide changes being made to equipment based on problems experienced by nurses—two of these were initiated by nursing communication to management, and one by management specifically asking for nursing input. Staffing decisions, however, were as likely to be seen as a lack of management’s commitment to safety as a positive—for example, “In our unit there is no give for patient safety. By that I mean when your patient, all of a sudden turns around from being stable to being totally unstable requiring more than maybe two nurses, is there a give for us to make that happen? No.”

The safety climate surveys had a total of 12 questions specifically asking about nursing and hospital management and administration (table 4). In contrast with the ambivalence of the focus-group findings, the mean scores were positive towards managements’ role and actions in patient safety. Consistent with the focus-group discussion, however, is the fact that the lowest score was given to the statement, “we have enough staff to handle the workload,” This mean score of 3.29 also had the widest standard deviation (1.19), indicating a wide range of responses to this statement.

DISCUSSION

The ICU nurses at these four community hospitals all worked in units that were involved in an intervention that decreased catheter-related bloodstream infections. They received education and protocol changes from physician and nurse leaders in the unit, yet when questioned broadly about patient safety in the months after this intervention, nurses rarely mentioned any of these activities or ideas. Participant nurses describe patient safety as double-checking medications and environmental threats, along with advocating for the well-being of the patient. Nurses believe they can protect patients because of their commitment to their job and their personal abilities—the traditional medical model of performance where educated, motivated staff are expected to deliver perfect performance. Nurses also do not express receiving focused cohesive messages about safety from their management, other than the acknowledgement that safety is important. Using adult learning models, we propose three possible explanatory factors for the inability of a specific safety initiative to generalise to broader concepts of safety: self-initiated learning is stronger than management-initiated learning; delayed outcomes, even when positive, have less effect than immediate outcomes; and first-degree problem-solving (fixing the immediate problem) tends to maintain the status quo.

There are many theories about how adults learn. While some of them focus exclusively on the key role of experience,27 almost all note the importance of experience and personal relevance in successful learning.2829 When asked to define and describe patient safety during the focus groups, the nurses almost exclusively relate those areas where they personally discover problems. The primary importance of monitoring and double-checking medications and physical elements of the environment to nurses is validated by the walk-round checklists which are strongly slanted to the same issues. In juxtaposition to the strength of self-initiated problem-solving are the passive messages about safety they receive from nursing and hospital management. While the nurses acknowledge that management cares about safety when answering survey questions, the focus-group participants describe management’s safety message as occurring predominantly through educational activities (classes, posters and notices). Consistent with adult learning theory, studies show that these types of interventions generally have little or no impact on professional practices.3031

The nurses in our study were proud of their dedication to their patients and their job. Their diligence in finding immediate fixes for problems they encountered did lead to immediate benefit for their patients, were gratifying to the nurses and allowed them to continue caring for their patients. Problem-solving coupled with immediate feedback is another desirable component of adult learning.28 When a nurse finds a loose restraint and fixes it or removes an incorrect medication from an IV line, she sees an immediate value. Safety programs like the CRBSI project offer no immediate feedback on patient safety—the patient immediately looks no differently whether the line was inserted following the protocols or not. Even if process and outcome feedback occurs monthly, it is relatively distant compared with the immediate rescue feedback from double-checking.

A study of nurses in other hospital settings found that 92% of problems were solved by first-order problem-solving, described as dealing with problems only by focusing on doing what it takes to continue the care of the patient and involving only those with who the nurse is most comfortable (not necessarily those best able to solve the problem).3233 This local success leads to less organisational learning and fewer system improvements. Figure 1 shows how this model of first-order problem-solving occurs within the ICU. Most of the patient safety problems described by the participants were medication errors and environmental threats that were dealt with by double-checking and patient advocacy. This first-order problem-solving usually fixed the problem for the patient and gave the nurse a sense of satisfaction. However, these problems keep recurring—the focus-group participants gave multiple examples of medication and environmental errors—but could relate few examples from years of experience of “second order” problem-solving—system changes that eliminated the problem.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Most problems and errors discovered by nurses are dealt with within the constraints of continuing patient care and asking for help within comfort zones. These “first order”32 problem-solving skills usually leave the status quo unchanged, allowing the same or similar problems or error to continue occurring.

There are limitations to this study. While the CRBSI project was performed at 10 hospitals, this study reports on four of them. The willingness of the ICUs at these four hospitals to participate in the focus groups may mean that the management at these ICUs possesses different qualities than those that did not participate. Focus groups were not performed prior to the central line project. While we are describing nurses’ perceptions after an intervention, and do not claim to assess change, future research should consider exploring what types of interventions do change nurses’ functional descriptions of safety. And while focus groups allow us to gain a broad understanding of a topic, they do not allow us to assess the importance of issues to individuals or sub-groups. We do not know if the same individuals completed surveys and also attended a focus group, and the response rate to the survey was low (40%), but since this research was exploratory and analysed qualitatively, the triangulation of data from these sources still adds to the trustworthiness of our conclusions.2526

The institution of evidence-based practice is a concrete approach to systematically improve the safety of an ICU.17 However, despite the success of this project in reducing the CRBSI rate, participants in the project did not generalise their experience or learning to other aspects of patient safety. Instead, the nurses defined their primary contribution to patient safety as double-checking environmental and medication issues. Despite “education” from management, these nurses continue to define patient safety primarily from personal experiences. We speculate from these results that substantive changes in true patient safety culture require multiple interventions and continued forces. Broadening administration checklists or rotating topics addressed might be an initial approach. Future research should continue to explore methods for improving the entire range of nursing safety within the ICU.