Patient safety has been a highly researched topic in health care since the year 2000. One approach for improving patient safety has been to encourage patients to take an active role in their safety during their health care experiences (Entwistle, Mello, & Brennan, 2005; Schwappach, 2010). However, there is little evidence that sheds light on how patients view their roles, particularly in the acute care setting. Some scholars have theorized about what patients may be capable of doing and why they should be good at reducing their risk of adverse events (Koutantji, Davis, Vincent, & Coulter, 2005; Lyons, 2007; Vincent & Coulter, 2002). However, little research has examined patient perceptions of what they believe they should be doing or are willing to do. To fully engage patients in their care experiences, it is important to understand their perceptions. This study used an inductive methodology to qualitatively explore what patients believe the patient role in patient safety should be.
Background
Patient safety has been inconsistently conceptualized and defined. This may be because scholars have not agreed on the extent to which patient safety is different from quality (Emanuel et al., n.d.). Emanuel et al. (n.d.) defined safety in part as an attribute that emerges from a health care system that "minimizes the incidence and impact of, and maximizes recovery from, adverse events" (p. 6). During the past decade, there has been a concerted effort in the United States to encourage patients to become more involved in assuring patient safety. In 2000, the Joint Commission began focusing on patient safety and has regularly launched new campaigns with the objective of engaging patients and their families in the hospital and other care settings. Such campaigns began in 2002 with the Speak Up initiative, in collaboration with the Centers for Medicare and Medicaid Services (http://www.jointcommission.org). As of January 2010, the Joint Commission made available 13 Speak Up publications for various health services. A survey conducted in 2008 found that 80% of 1,900 hospitals were utilizing Speak Up materials, making many of these available in patient rooms (http://www.jointcommission.org). The Agency for Healthcare Research and Quality (AHRQ) has made several patient publications available, the most widely known is its 20 Tips to Help Prevent Medical Errors. Similarly, the National Patient Safety Foundation includes patient publications on its Web site, and several of the AHRQ documents are posted there. Most of the patient materials encourage patients to be observant, to ask questions, and to talk with their care providers if they believe anything is wrong with their care. Similar efforts have been made in Europe as well (Schwappach, 2010). Clearly, U.S. and European health policy leaders believe that patients should play a role in their safety when receiving health care.
Rationale for Involving Patients in Safety
Involving patients in efforts to improve safety makes sense in many ways. Importantly, given the fragmentation of care and the large numbers of care providers with whom patients interact, patients are usually the only participants who are involved in the entire process of care (Koutantji et al., 2005; Lyons, 2007; Schwappach, 2010). As a central customer of health care, patients' views on patient safety should be a primary concern for any quality improvement efforts (Vincent & Coulter, 2002). It could be argued as well that engaging patients in patient safety should be one aspect of patient-centered care (Entwistle, 2007). There is some evidence that patient-centered care is positively related to patient safety (Burroughs et al., 2007; Rathert & May, 2007; Vincent & Coulter, 2002).
Patients can play numerous important roles in their safety, several of which were examined by Koutantji et al. (2005). First, patients can be involved in monitoring and reporting adverse events, which is commonly accepted as necessary for improving the quality of care delivery. Adverse events are chronically underreported by staff (Rowin et al., 2008), so patient reporting could provide another lens with which to view care delivery systems. Second, when patients are educated about their medications and potential side effects, they may be empowered to double check that they are being given the correct medication and the correct dose, at the correct time. Third, patients can be informed about what to expect with surgeries and encouraged to report any adverse events or complications immediately. Fourth, some evidence suggests that patients can play a role in reducing hospital-acquired infections by observing and asking about staff hand-washing; and finally, patients can assist in being properly identified prior to treatment. They can keep a list of their medical histories, allergies, and current medications. As use of electronic personal health records increases, patients can make sure these are up to date and that their current care providers have access to them (Koutantji et al., 2005).
An examination of large-scale patient education efforts found that most of the United States' large-scale patient-directed advisories (AHRQ, Joint Commission, U.S. Department of Health and Human Services [HHS], and National Patient Safety Foundation) had not been developed using patient input, and none had been tested on patients prior to rolling them out nationally (Entwistle et al., 2005; Schwappach, 2010). One study examined patient perceived effectiveness of 14 of the AHRQ's 20 tips for avoiding medical errors and asked patients to indicate their likelihood of actually taking such actions (Hibbard, Peters, Slovic, & Tusler, 2005). Although patients perceived most of the actions to be effective, most indicated that they will not be likely to actually question their care providers (e.g., "Asking health care workers who come in contact with you if they have washed their hands" and "Confirming whether you are getting the right medication and dose"). Given that large-scale educational efforts are resource intensive, empirical evidence should support the effectiveness of patient education strategies.
Patient Empowerment
The patient empowerment literature may add some insight into assumptions about patient involvement in patient safety. Patient empowerment has mostly been studied in the context of chronic illness as an alternative paradigm to patient compliance (Anderson, 1985; Anderson & Funnell, 2005). Although inconsistently defined, a recent literature review found that patient empowerment tends to be defined in terms of patients having the rights and abilities to make choices and taking responsibility for the consequences of their choices (Aujoulat, Marcolong, Bonadiman, & Deccache, 2008). Generally, self-determination is a guiding principle. Empowerment seems a logical direction given recent emphasis on patient-centered care. However, several scholars have identified complexities with patient empowerment that warrant further research and evidence, possibly prior to expecting patients to be engaged in their safety. Anderson and Funnell (2005) argued that patient empowerment essentially requires a paradigm shift based on how medicine has been traditionally practiced and that health care paradigm shifts are strongly resisted by practitioners and patients. For example, they pointed out that in acute care, it is typical for patients to surrender control to care providers, who use their expertise to solve patient health problems. Furthermore, their research found that although clinicians agreed intellectually with the notion of patient empowerment, their behavior in practice still reflected a traditional acute care paradigm. Aujoulat, d'Hoore, and Deccache (2007) found that there are two key dimensions necessary for patient empowerment: a personal transformation for the patient and an interpersonal cocreation with the care provider. In other words, empowering patients does not mean "giving" power to them; it means collaborating with them through a caring relationship and enabling empowerment to occur. Thus, engaging patients in their safety may require more than providing educational materials that suggest they ask questions; it may require a transformation of interpersonal dynamics between patients and clinicians.
There are additional complexities to consider when expecting patients to be engaged in their safety. Salmon and Hall (2003) argued that assuming that patients are the primary agents in managing their care serves to locate the responsibility of problem areas (such as unexplained symptoms or pain) within the patient and thus removes some responsibility from clinicians. Although the patient empowerment literature has focused more on management of chronic disease than on acute care patient safety, such findings should be kept in mind before assuming that patients should be expected to take an active role in their safety. Indeed, Tomes (2007) argued that patient empowerment has produced positive changes, particularly in some treatment protocols. But demands for cost containment and other market issues have limited the self-determination of patients and care providers alike, and there continue to be "asymmetries in the knowledge and power bases" in the care provider-patient relationship (Tomes, 2007, p. 698). Such complex issues may make it difficult for patients to reliably be active agents in their safety.
Efficacy of the Patient Role
Although there are many ways patients could assist in ensuring their safety, few empirical studies have actually examined the efficacy of the patient role. In one randomized clinical trial (n = 209), it was hypothesized that medical inpatients who were provided personalized drug information and medication lists during their hospital stay will experience fewer adverse drug events compared with a control group (Weingart et al., 2004). The researchers proposed that if patients have possession of current drug information, they or their surrogates could prevent errors in prescribing. This is because they should have knowledge about allergies and should be able to confirm dispensed medications with their lists in terms of the drug name, dosing, form, route, and administration time. Patients in both groups received drug safety information upon admission to the hospital, but only patients in the intervention group (n = 107) received current personalized drug information. Overall, 5.3% of patients experienced an adverse drug event, and 7.7% of patients experienced a close call. There were no significant differences between the groups for adverse events, close calls, severity levels of events, or the number of intercepted close calls. Of three serious adverse events, all were considered preventable, and one was determined preventable due to the patient's actions. This case occurred in the intervention group, so at least in that situation, the intervention did not prevent the serious adverse event. Schwappach (2010) conducted a systematic review of studies that attempted to engage patients in safety efforts in the United States and Europe. The study found that although survey responses indicate that patients have favorable attitudes toward the idea of engaging in safety, their intentions and comfort levels in participating are less clear. In addition, this review found little evidence that efforts to increase patient involvement have been successful.
Although experts will continue to debate the role of the patient, policy leaders will continue to encourage patients to be vigilant about their care. An important area of research is to gain a better understanding of how patients define and view their roles. Indeed, research has found that patients often view their roles very differently from how clinicians view the patient role (Anderson, 1985). Understanding patient perceptions of their role should help policy makers determine the most effective ways to engage patients in their safety. This study was set out to inductively answer the following research question: What do hospitalized patients believe their role should be in patient safety? This question arose after a focus group study in which patients discussed how they defined patient safety and what types of incidents happen in the acute care environment that make them concerned for their safety. Given that this was exploratory research into patient perceptions, we made no hypotheses about how patients will define their roles.
Method
Participants
This study was part of a larger study designed to develop a survey measure of patient safety. A random sample was drawn from a sampling frame of all medical and surgical patients who had had an overnight acute care visit in the previous 90 days at one of three participating acute care hospitals. The hospitals were part of the same health system. Patients were excluded from sampling if they had a psychiatric visit, were discharged to a facility other than home, or were not discharged alive. The sample was drawn based on the historical patient response rates at these hospitals of around 30%, with a target of 300 returns per hospital. Overall, 1,040 questionnaires were returned, for a response rate of 33%.
Procedure
Surveys were mailed to patient homes in packets that included a questionnaire, a cover letter, and a postage-paid return envelope. A cover letter explained that the hospital was working with researchers to pilot test questions that were related to "patient safety and medical mistakes." Nonresponders were mailed two reminder survey packets at 2-week intervals. Questionnaires included six global patient satisfaction items typically asked by the hospitals, as well as 38 new close-ended items that inquired about specific events that had occurred during the patient's visit. At the end of the questionnaire an open-ended question asked, "Please tell us what you believe is the patient's role in patient safety." Responses were keyed verbatim into a spreadsheet.
Analysis
Analysis was conducted in several phases. Because the study utilized an inductive, exploratory approach, the first phase proceeded according to steps outlined by Creswell (2009) in which raw data are continually reflected upon as the analysis occurs. Creswell's strategy includes the steps of organizing the data, reading through them, coding, describing, identifying themes, interrelating themes, and interpreting meanings. Three investigators participated in this phase. Initially, one researcher who had extensive experience with qualitative methods and patient perceptions read through all of the responses and listed general reoccurring themes. Next, the other two investigators independently reviewed each comment and noted whether responses fit into the categories identified by the first investigator. Responses that were obviously unrelated to the question (e.g., "patient is deceased") were removed from the analysis. The investigators then discussed how they had categorized each response. On the basis of this discussion, some of the original categories were revised slightly, and some new ones were added. Numeric codes were assigned for each category, and responses were coded. The researchers again discussed the assignments until consensus was reached.
For the second phase, categories were combined into higher levels of abstraction, into two overarching themes. The first theme was whether the patient's role should be active or passive. The investigators coded a variable for comments regarding whether they were active (1) or passive (0). The second theme emerged because many comments implied that patients should be able to trust that they are getting safe care. The researchers examined the comments in terms of definitions and dimensions of trust identified by Hall, Dugan, Zheng, and Mishra (2001) and Mechanic and Meyer (2000). Trust was defined as "optimistic acceptance of a vulnerable situation in which the truster believes the trustee cares for the truster's interests" (Hall et al., 2001, p. 615). According to Hall et al. (2001, 2002), patient trust in providers involves expectations of (a) fidelity (caring, respecting, and avoiding conflicts of interest), (b) competence (technical: avoiding mistakes, providing good results; interpersonal communication; and bedside manner), (c) honesty (telling the truth, avoiding falsehoods, gathering accurate medical history, obtaining informed consent, and avoiding outright lies, half-truths, or deception by silence), (d) confidentiality (properly using sensitive information), and (e) global trust (content that may cover more than one dimension, or general implications of trust). We independently assigned each comment a code regarding whether the response implied that patients should be able to trust or not (patient role is to trust = 1, comment that did not indicate that a patient should trust = 0). The researchers then assigned comments that implied trust to one of the five dimensions.
Finally, the third analysis phase included an examination of several patient variables that could be related to active/passive and trust perceptions. Data on patient age (in years) and length of hospital stay (in days) were available from the hospital. In addition, two self-report measures asked about patient general health status and the patients' overall ratings of the care they received. Patients rated both of these items on an ordinal scale where 1 = poor and 5 = excellent. For statistical analysis, we used the Wilcoxon rank-sum test because all data were nonparametric (active/passive, trust), were ordinal (health status, overall rating), or had nonnormal distributions (age and length of stay). We compared respondents who answered the patient role question with those who did not to see if the groups differed on any of these variables. We then examined these variables based on whether respondents had indicated that patients should be active/passive or be able to trust. Finally, we examined how the qualitative variables were related to overall ratings of care.
Results
Descriptives
Forty-seven percent of the respondents (n = 491) answered the open-ended question, with roughly half of the respondents from each hospital included. We compared general responses for each hospital separately to see if there were differences in response categories based on the hospital where the patient received care. Patient response rates, demographics, and frequency of themes mentioned were similar at each of the hospitals, and the rank order of the top categories was the same (Table 1). Therefore, to increase the sample size and possibly reveal less frequently mentioned themes, we combined the data into one set for further analysis.
As might be expected, respondents who answered the patient role question were slightly different from those who did not. Those who answered the question were younger and had a shorter average length of stay than those who did not, indicating that they may have had less severe health concerns while in the hospital. Variable means and standard deviations for both groups are presented in Table 2. Correlations among study variables for those who answered the patient role question are depicted in Table 3.
First-Phase Qualitative Analysis
Overall, patients most frequently said that the patient role is to follow instructions (23%). Related to this, and sometimes in combination, many patients said patients should listen to their care providers (7%). As one patient put it, the patient's role is "To listen carefully and follow directions of your care givers." Additional subcategories were identified in which following instructions was implied (e.g., being cooperative). The second most frequent comment category was that patients should expect competent care (8%). Care competence was indicated with comments such as "To be monitored and watched closely" and "The patient should feel confident that care will be given quickly[horizontal ellipsis]" and a comment from the parent of a newborn: "The parent should be involved in the safe care of the infant, but not have to make sure it is happening." Many patients said that the patient role is to proactively pay attention to what is going on and ask a lot of questions of clinicians (7%). One patient put it this way: "To be aware of what is happening and to ask questions when in doubt of what is happening. To challenge anything that doesn't seem RIGHT."
Three other categories were mentioned relatively frequently. These were (1) knowing and being informed about their conditions and treatments (6%), (2) monitoring what is going on around them (4%), and (3) being active participants in the care being provided. For "knowing," we included comments such as "Patients should be aware of personal needs and reactions to medications" and "If possible, a patient or their family member should make sure that their meds or treatments are the correct ones." Overall, most patients appeared to be suggesting that patients should listen, trust, and be cooperative. One patient pointed out that, often, patients are not in the position to ask questions: "Assuming the patient is able to, they should play an active part in safe care - but how many of those people are in the hospital?" Response percentages are presented in Table 4.
Examination of the data broken into age groups showed a nonsignificant trend that older patients were more likely to provide comments indicating passivity. The most striking difference in responses by group was for respondents who were identified as parents of newborns (their ages were unknown, as the newborn was the patient of record). For this group, only 9% of comments indicated that patients should follow instructions. This group was much more likely than any of the others to indicate that patients should be active participants in the care experience (15%) and should know and be informed about patient conditions and treatment plans (18%). Interestingly, a much greater proportion of this group indicated that patients should expect to be treated by competent staff (18%). All of the other patient groups indicated that following instructions was the most important role for the patient. However, for patients under age 45 years, just as many as the parents of newborns indicated that patients should ask questions about their care (18%). Table 5 depicts theme frequencies by patient age group.
Second-Phase Qualitative Analysis
At a higher level, analysis suggested that many of the patients expected the patient role to be passive. Patients overwhelmingly implied that patient safety was not their responsibility; they wanted to trust their clinicians and providers to provide the care they needed when they were sick. The comments indicated that respondents felt vulnerable as patients, and they expected or needed to be able to trust that they were safe; in fact, half indicated this (57%). Recall that the comments were recoded if they indicated one of the dimensions of trust articulated by Hall et al. (2001). Over 60% of the trust comments fell into the category of competence. As noted earlier, many comments directly indicated that patients should expect competent staff. We included variations of "follow instructions" in this category because we reasoned that if patients are to passively follow instructions, this implies that care providers and the care delivery systems are competent. Example comments included the following: "Patient has no other role but to follow orders and/or instructions given by doctors and/or nurses" and "As a patient I place my life in the hands of the doctors and nurses and expect their knowledge and experience to justify my confidence."
The second largest proportion of comments fell into the honesty dimension (20%). In this dimension, we included comments indicating that patients should be kept informed and be made aware of their treatment plans and prognoses. An example comment was, "To be made aware of what effects your treatment should cause[horizontal ellipsis]" The next most frequent dimension included comments related to global trust (14%). Here, we included comments that mentioned trust directly, or having confidence. "The patient should feel completely safe and assured that the staff is competent" and "Believe that doctors are doing their best to zero in on what is or isn't the problem[horizontal ellipsis]" articulated this dimension. The fidelity dimension received 7% of comments. This included comments such as "I believe and have faith that all health care professionals truly care and respect their fields[horizontal ellipsis]" Finally, no comments fit into the dimension of confidentiality. Trust counts and percentages appear in Table 6.
Comparative Statistical Analysis
Comparison of the qualitative variables on the patient and outcome variables revealed several significant differences. The Wilcoxon rank-sum test indicated that respondents who answered the patient role question were significantly younger than those who did not answer it (p = .04) and had significantly better health status (p = .02). However, there were no significant differences among those who answered the question in terms of their overall quality of care ratings or length of stay. The only other significant difference was that those who were younger were more likely to say the patient should be able to trust (p = .04). Therefore, length of stay, health status, or overall quality of care ratings were not related to whether patients perceived the patient role as active/passive, or expected to be able to trust. Complete results of the Wilcoxon rank-sum analysis appear in Table 7.
Discussion
Results from this study were somewhat surprising. Most notably, in contrast to what patient safety policy makers have assumed, patients in this study tended to believe that patients should simply be able to trust that they are getting safe care in the hospital, rather than needing to play an active role in ensuring safe care. Additional analysis on some hospital-reported variables (age and length of stay) and patient-reported outcome variables (general health status and rating of the overall quality of care received) indicated that patients who said that the patient should trust were not very different from those who did not say this. In fact, those who were younger and reported slightly better health status were more likely to say that they should be able to trust. Given that younger people with less time spent in the hospital answered the question, it could be that people who did not answer the question would be sicker and less likely to believe that they should play a role in safety.
The most frequently mentioned content theme in the qualitative analysis was that patients should follow instructions, with roughly one quarter of patients indicating this. This suggests that patients in the acute care setting generally may be inclined to be passive as opposed to more proactive in patient safety. When these comments were broken down further, many of the respondents who had indicated that patients should follow instructions also indicated that patients should listen to their care providers and be cooperative. These results are consistent with the traditional acute care paradigm in which patients give their control over to clinicians to gain expertise, technology, and caring compassion (Anderson & Funnell, 2005). The acute care paradigm is inconsistent with patients taking charge and managing their care, let alone their safety. Furthermore, Anderson and Funnell (2005) argued that this paradigm is embedded in the minds of health care professionals and is the basis of most policies and procedures of health care organizations. Thus, even patients who want to be proactive in their safety will likely face difficulties in the acute care setting due to this paradigm which is resistant to change.
Another frequent theme was that patients should ask questions. Asking questions is consistent with what health care leaders are recommending and is a good first step in getting patients to be more proactive. On the other hand, some research suggests that patients often feel concerned and even take personal actions to help facilitate their safety but do not communicate this to their care providers (Schwappach, 2008). Thus, getting patients to ask questions may require a consistent and concerted effort by care providers to enable patients to feel comfortable in this role.
Although the top themes seemed prominent, it is notable that there was a variety of responses. Clearly, there is not a patient consensus on what the patient role should be. This could be due to patient characteristics such as education, socioeconomic status, and even previous experience with errors. Schwappach (2010) noted that patients who had experienced errors in the past, or who knew people who had experienced errors, were more likely to be vigilant observers while in the care setting. In addition, different contexts, even within the acute care setting, could result in differences in patient perceptions of their roles. Although our data did not indicate any differences in patient perceptions based on length of hospital stay, it is possible that patients who are in the hospital with a traumatic injury, for example, might have different ideas about the patient role than those who are there for a medical ailment. Future research should examine this possibility, as differing care needs may require different types of patient engagement strategies.
Patient Trust
As with that of Mechanic and Meyer (2000), our analysis found that most respondents who implied that patients should trust were concerned with technical or interpersonal competence, and this was followed by honesty and disclosure. Trust is an issue that appears to have been overlooked in the zeal to improve patient safety. Patients are inherently vulnerable in the health care system, and it appears that many of them may simply believe that they should surrender their trust to competent care providers. Indeed, trust can be a coping mechanism that arises in the context of the distress of physical illness or disability (Hall et al., 2001). It could be that asking patients to be more proactively involved in preventing adverse medical events will undermine the trust they have in their care providers and the general health care system. In one study that examined patient involvement in decision making, Kraetschmer, Sharpe, Urowitz, and Deber (2004) found that trust was significantly related to the extent to which patients wanted to be involved in decision making. The study found that the more trust participants had in their physicians, the less they wanted to be involved in decision making. Similarly, it could be that patients who trust more do not think that they should be questioning their safety. According to Hall et al. (2001), patient trust goes beyond satisfaction in that it is concerned with assessing the character and motivation of physicians, other clinicians, and organizations. When trust is violated, it can lead to moral outrage, rather than simply disappointment that expected results are not achieved. Thus, scholars and practitioners need to consider the impact on patient trust that some of their proposed patient involvement activities may pose. Indeed, Lyons (2007) argued that rather than relying on patients to help improve safety, we should be developing more reliable and transparent systems to foster among patients a greater sense of confidence and trust in their health care.
Practical Implications
If patients do indeed believe that they should listen and follow instructions, strategies that require clinicians to provide patient engagement instructions might be effective. Hospitals trying to engage patients in safety should focus first on getting patients to simply ask more questions about their care, possibly giving each patient specific questions they should ask and instructions about when they should ask. Aujoulat et al. (2007) proposed that getting patients involved requires staff to develop empowering relationships, in which there is patient centeredness, mutual acknowledgement, relatedness, and continuity. Developing such relationships will require hospital management to provide an appropriate working environment, one that gives staff the incentives, or at least the latitude, to spend the time they need to facilitate such activities. This may require a cultural transformation for many organizations. This issue should be studied further to determine if hospitals really are ready for the interpersonal efforts that will be needed for clinicians to assist in empowering patients.
If patients are willing to accept that their role is to know about themselves and their medical conditions and treatments, this seems like another topic on which organizations and care providers could become involved in educating patients. Encouraging patients to learn about themselves could be coupled with encouraging them to ask questions. Indeed, many of the comments that stated that patients should know about themselves were made in combination with asking questions and keeping informed. However, care providers and hospital leaders need to keep in mind that some patients are deficient in health literacy, and this may make it challenging for them to become educated about themselves or to feel confident in asking questions. Even if patients believe that their role is to know about their illnesses and treatment plans, additional research needs to explore the best ways to facilitate patient learning. There may be the need for different modalities depending on patient health literacy and acuity variations.
Limitations
Although this study obtained some insightful perspectives from individuals with recent hospital stays, as with all research, this study had some limitations that should be noted. Importantly, over half of our respondents were 65 years or older. Although it is typical for greater proportions of older patients to respond to patient surveys, this needs to be considered when evaluating the data. Age could be important because it appears that younger patients are more inclined to be proactive and engaged in their care. Thus, our finding that patients think that their role is to "follow instructions" may be a generational finding. However, this finding still may be important in the coming years given that the general patient population is aging. In addition, our results could have suffered from nonresponse bias effects. The overall response rate in the larger survey was about 33%. Although we would have preferred a higher rate, this appears consistent with current large-scale patient surveys (Goldstein, Elliott, Lehrman, Hambarsoomian, & Giordano, 2010). Only about half of the respondents answered the question regarding the patient role. Those who answered the question may be systematically different from other patients. The qualitative nature of our study makes this less of a concern, however, because our objective was to inductively explore how patients perceive their roles. Future studies should examine relationships with representative samples. Finally, although our study was qualitative, because it was part of a written questionnaire that was mailed to patients, we were not able to probe more deeply into respondent answers. On the other hand, because it was a mailed survey, we obtained a larger set of responses from a wide geographical region, so our findings may be more representative of typical inpatients. It is noteworthy that theme categories were relatively consistent across the three hospitals.
In conclusion, the results of this study lend support to researchers who question the extent to which patients should be expected to assist in facilitating patient safety. Some of the literature has suggested that patients may not be capable of intervening in their own safety, and some has indicated that doing so may shift responsibility from care providers and organizations to patients. This study found evidence to suggest that many patients may not want or be psychologically able to play an active role in patient safety. Thus, at least in terms of hospitalized acute care patients, health care leaders may want to focus on involving patients by having them participate in small segments of their care rather than the full inpatient experience. And perhaps, it will be wise to question whether some types of patient involvement initiatives will result in decreasing the trust that patients have in their care providers and health care organization.
Acknowledgments
The authors thank Vikki Entwistle for initial suggestions about the analytic framework. They also thank three anonymous reviewers from the Academy of Management for their insightful suggestions, as well as the patients who provided thoughtful comments regarding the patient's role in patient safety.
References