Perceptions of Appropriateness of Care Among European and Israeli Intensive Care Unit Nurses and Physicians FREE

Clinicians perceive the care they provide as inappropriate when they feel that it clashes with their personal beliefs and/or professional knowledge.¹ Intensive care unit (ICU) workers who provide care perceived as inappropriate experience acute moral distress and are at risk for burnout.² This situation may jeopardize the quality of care and increase staff turnover.^2- 4

The principal causes of moral distress reported in ICU nurses are delivery of futile care, unsuccessful patient advocacy, and communication of unrealistic prospects to the patients and families.^4- 8 ICU physicians may be troubled by a perceived lack of power to make the clinical decision that most benefits a specific patient.⁵ A survey among 504 European ICU physicians showed that 73% of units frequently admitted patients with no realistic hope of survival, although only 33% of the physicians felt that such patients should be admitted.⁹ More recently, 87% of 114 Canadian ICU physician directors reported that futile care was provided in their ICU over the last year.¹⁰ However, earlier studies of perceived inappropriateness of care in the ICU did not provide data linked to individual cases. Consequently, the extent of perceived inappropriateness of care in the ICU is unknown and the magnitude of situations causing moral distress may be underestimated.

The primary objective of this study was to determine the prevalence of perceived inappropriateness of care among clinicians in European and Israeli ICUs, to describe the patient-related situations associated with perceived inappropriateness of care, and to explore the level of agreement among clinicians concerning perceived inappropriateness of care. The secondary objective was to evaluate the hypothesis that perceived inappropriateness of care is associated not only with situational factors, but also with personal characteristics and work-related factors as well as with intentional job leave. The theoretical framework is given in Figure 1.^{1- 7,11- 21}

We conducted a single-day cross-sectional study among clinicians in European and Israeli adult ICUs including nurses, head nurses, and junior and senior ICU physicians. Ten members of the European Society of Intensive Care Medicine (ESICM) ethics section agreed to serve as national coordinators with 1 representative in each country (Belgium, France, Germany, Israel, Italy, Malta, Poland, Portugal, Switzerland, and The Netherlands). Each national coordinator recruited adult ICUs for the study and obtained approval from the relevant ethics committee for each ICU. In each ICU, a local investigator contacted and enrolled the ICU clinicians scheduled to work in the ICU on the study day and organized an information session during the week before the study.

The study took place from 8 AM, on Tuesday May 11, 2010, to 8 AM, on Wednesday May 12, 2010, in all participating countries except Israel, where the study took place on May 25, 2010, for organizational reasons. The local investigators were asked to establish a coded list of the patients admitted to the ICU on the survey day. This list was destroyed after data collection to preclude identification of the patients. The local investigators were asked to resend the questionnaires within 1 week, making recall bias unlikely.

Three questionnaires were used for data collection: the ICU questionnaire, the clinician questionnaire, and the perceived inappropriateness of care questionnaire (eAppendices 1, 2, 3).

The ICU Questionnaire. In each study ICU, the local investigator completed the ICU questionnaire about ICU characteristics (type of hospital and ICU; mortality rate; number of ICU clinicians; and availability of an ethics consultant, psychologist, or both) and end-of-life practices (symptom control, decision making, and discharge of dying patients to the wards).

The Clinician Questionnaire. Each nurse and physician working in the ICU on the day of the survey completed a questionnaire about personal characteristics (including age, sex, religion, professional role, and work experience), perceived work characteristics (job strain and ethical environment), and intent to leave. The respondents indicated the number of patients in their care on the survey day and the number of patients perceived as receiving inappropriate care.

The clinician questionnaire included the Job Strain Scale, a validated 12-item scale exploring job demand, control, and social support.^20- 21 According to the job strain model developed by Karasek and Theorell,²⁰ job strain occurs when job demands (workload) are high and job control (sum of skill use and decision-making authority) is low. A third factor in this job strain model is social support (from the supervisor and coworkers), which protects against job strain. The total score is obtained by adding the control and social support subscores then subtracting the demand score. Higher scores indicate less job strain.

The ethical environment was defined as “the organizational conditions and practices that affect the way ethically difficult patient care problems are discussed and decided.”²² We assessed 7 aspects of the ethical work environment previously identified in scientific studies: tolerance of different opinions and values; possibility of ethical debate^{5- 6,22- 24}; empathic understanding provided by colleagues; collaboration among colleagues^3,8,23- 24; nurse-physician collaboration^5- 6; presence of nurses during communication of end-of-life information; and active involvement of nurses in decision making.^25- 27 These 7 items showed good internal reliability (Cronbach α, 0.79; P < .001).

The clinicians were asked to report whether they had thoughts of leaving their current job or profession. Past effective job leave due to disagreement about patient care was recorded.^5- 7

The Perceived Inappropriateness of Care Questionnaire. Clinicians who reported perceived inappropriateness of care were requested to complete the perceived inappropriateness of care questionnaire for each patient who was perceived as receiving inappropriate care. The questionnaire evaluated the reasons leading the clinician to consider that care was inappropriate. The patient code allowed us to link the questionnaire responses to data about the relevant patient and therefore to assess the level of agreement among clinicians regarding perceived inappropriateness of care for a given patient.

In this study, we defined perceived inappropriateness of care as a patient-care situation perceived by the respondent to fit 1 or more of the following statements or scenarios: (1) disproportion between the amount of care given and the expected prognosis (too much or too little care); (2) persistent nonadherence of the patient; (3) other patients would benefit more from ICU care; (4) inaccurate information was given to the patient or family; (5) the patient's wishes concerning treatment preferences were known but not respected; (6) one of the parties involved did not participate in decision making related to the patient; and (7) the patient was not getting good-quality care.

To build the study questionnaires, we asked a panel of experts in intensive care, palliative care, and communication to use a Delphi method to develop a consensus about the 7 scenarios and the content of the 3 questionnaires. The original English-language questionnaire was translated into the first language of each participating country then back-translated to English (Brislin method).

The prevalence of perceived inappropriateness of care was defined as the number of clinicians reporting perceived inappropriateness of care for at least 1 of their patients divided by the total number of surveyed clinicians in the same ICU. The perceived inappropriateness of care rate for each clinician was defined as the ratio of the number of patients with perceived inappropriateness of care reported by the clinician over the total number receiving care from the same clinician.

This study has been approved by the appropriate institutional review board in all participating ICUs and countries. Except for Belgium, where written informed consent was obtained from the participating clinicians, completing the questionnaire was taken as evidence of consent to study participation.

Values were described as median or percentage. The χ² test was used to assess differences between nurses and physicians and to assess differences in patient characteristics between patient groups.

Two hierarchical multivariate models were built to identify ICU and clinician characteristics (fixed effects) associated with (1) the perceived inappropriateness of care rate and (2) intentional job leave. We modeled the correlation between clinicians working in the same ICU by including a random ICU effect, nested within a given country, to take into account a possible correlation between ICUs in the same country. The full model included all the variables of the ICU and clinician questionnaires. A stepwise backward selection procedure with a significance level of 5% was used to build the final model. All statistical analyses were performed using SAS statistical software version 9.2 and SPSS version 17.

Of the 99 ICUs invited to join the study, 82 participated and 17 declined (2 because of no institutional review board approval) (Figure 2). In total, 1953 clinicians worked on the survey day and were eligible to receive the questionnaire (median clinicians/ICU, 19.5; IQR, 15-29). The median response rate within participating ICUs was 93% overall (IQR, 82%-100%), 93% among nurses (IQR, 82%-100%), and 100% among physicians (IQR, 80%-100%). The characteristics of the ICUs and clinicians are described in Table 1, Table 2, Table 3, and Table 4.

Of the 1651 clinicians who provided responses for calculating the perceived inappropriateness of care rate (number of patients with perceived inappropriateness of care over the total number receiving care from the same clinician), 439 reported perceived inappropriateness of care in at least 1 patient (27%; 95% CI, 24%-29%; Figure 2) (range across countries, 8%-49%). Of the 1218 nurses who completed the perceived inappropriateness of care questionnaire, each provided care to a median of 2 patients (IQR, 1-3); among them, 300 reported perceived inappropriateness of care (25%; 95% CI, 22%-27%). The 407 ICU physicians provided care to a median of 6 patients (IQR, 4-9) and among them, 132 (32%; 95% CI, 27%-38%) reported perceived inappropriateness of care in at least 1 of their patients. Seven of 26 clinicians failed to indicate their job title (nurse or physician) in the questionnaire.

In all, 397 clinicians completed 445 perceived inappropriateness of care questionnaires (Figure 2). The most common reported reason for perceived inappropriateness of care was perceived disproportionate care (65%) (Figure 3); in 89% of these cases, the amount of care was perceived as excessive and in 11% as insufficient. Disproportionate care was the leading reason for perceived inappropriateness of care among nurses (182/286, 64%) and physicians (99/144, 69%) (15 answers missing on professional role, P = .33). The second most common reason for perceived inappropriateness of care was a feeling that other patients would benefit more from ICU care than the present patient (38%) (Figure 3), This feeling of distributive injustice was significantly more common among physicians (64/144, 44%) than among nurses (98/286, 34%) (P = .05). Observing a lack of participation in decision making, persistent nonadherence of the patient, a lack of accurate information giving, perceptions of poor-quality patient care, and disregarding a patient's wishes were less frequently given as reasons to report inappropriateness of care in this study (Figure 3).

Of the 379 reports of perceived inappropriateness of care for which this information was available, 237 (63%; 95% CI, 55%-70%) stated that similar situations were common in the ICU. The recurrence of situations was more often reported by nurses when compared with physicians (73% vs 43%; P < .001). In 214 of 377 reports (68 missing this response; 57% [95% CI, 49%-64%]), the clinician was not confident that the situation associated with perceived inappropriateness of care would be resolved in the near future (nurses 39% vs physicians 48%; P = .08). More nurses, when compared with physicians, were quite, very, or strongly distressed by the perception of inappropriate care (68% [165/241] in nurses compared with 55% [71/128] in physicians; P = .01).

Patient codes were correctly recorded in 69 ICUs (Figure 4). Perceived inappropriateness of care was reported for 207 patients, corresponding with 23% of 883 ICU beds (95% CI, 20%-27%). For 136 of these patients (66%; 95% CI, 55%-77%), a single clinician, who in most cases was a nurse vs a physician, reported perceived inappropriateness of care (71% vs 29%; Figure 4). For 71 of the 207 patients (34%; 95% CI, 26%-42%), more than 1 clinician reported perceived inappropriateness of care; and in 66% of these patients (45/68 [≥1 professional role unknown in 3 cases]), at least 1 nurse and 1 physician reported the same view (Figure 4). These 71 patients represent 8% (95% CI, 6%-10%) of the 883 ICU beds.

Except for a longer length of stay, no other patient characteristics were associated with agreement on appropriateness of care (eTable 1).

The perceived inappropriateness of care rate is the ratio of the number of patients perceived as receiving inappropriate care, as reported by the clinician, over the total number of patients receiving care from the clinician. The results of univariate analysis are presented in the online supplement (eTable 2).

Multivariate analysis revealed that the following factors were independently associated with lower perceived inappropriateness of care rates (fixed effects): (1) decisions about symptom control shared by nurses and physicians as opposed to being made by the physicians only; (2) involvement of nurses in end-of-life decisions; (3) good collaboration between nurses and physicians; (4) work autonomy; and (5) perceived lower workload, only among nurses (Table 5).

The perceived inappropriateness of care rates were correlated with one another within ICUs and countries (random effect), showing some degree of homogeneity in perceived inappropriateness of care rates in ICUs within a given country.

Nine percent of clinicians (95% CI, 7%-11%) left a previous job because of disagreements related to patient care (147/1593; 58 answers missing). More nurses compared with physicians (10% vs 6%) reported past effective job leave (P = .02). Almost one-third of the respondents (31%; 95% CI, 28%-33%) had thoughts about leaving their current job (500/1630; unreported professional role for 21; 27% physicians vs 32% nurses; P = .08).

Perceived inappropriateness of care was independently associated with higher intentional leave from a job (Table 5). Being a nurse or a physician had no independent effect on job departure (Table 5).

To our knowledge, this is the first large-scale observational study describing perceptions of inappropriate care linked to patient-care situations both in ICU nurses and ICU physicians involved in direct patient care. We found that about 1 in 4 ICU nurses and 1 in 3 ICU physicians believed that they delivered inappropriate care to at least 1 of their patients on the day of the survey. Most of the respondents indicated that similar situations were common in their ICU, and more than half were not confident that these situations would be resolved in the near future.

Repeated perceived inappropriateness of care may strongly influence perceptions of a new patient care situation and as such, affect the quality of patient care.^3- 4,12 Moreover, in our study perceived inappropriateness of care was independently associated with intentional job leave both in nurses and physicians.

The most commonly reported reason for perceived inappropriateness of care was excessive intensity of care. In the ETHICUS study (end-of-life practices in European intensive care units), 89% of ICU physicians reported feeling comfortable with the end-of-life decisions they had made.²⁸ In our study, end-of-life decisions were mostly reported as being made too late or too infrequently. In addition to disproportionate care inducing perceived inappropriateness of care, a perceived failure to observe distributive justice was common, most notably among physicians.^25,29- 30

For two-thirds of patients receiving care from more than 1 respondent, only 1 respondent reported perceived inappropriateness of care. No severity of illness–related characteristics of the ICUs such as average ICU stay length or ICU mortality were significantly related to perceived inappropriateness of care. In addition, the prevalence of perceived inappropriateness of care varied widely across countries and across ICUs and clinicians within a given country. These data underline the subjective nature of perceived inappropriateness of care.^5,13,15,31 The high variability in judgement about appropriateness of care reflects that an individual clinician's judgement is a personal issue related to the clinician's own world view and is therefore colored by his or her own emotions, attitudes, backgrounds, and beliefs.^32- 37

As such, perceived inappropriateness of care will always be part of health care; however, in those workplaces with higher prevalence of perceived inappropriateness of care, there are organizational factors that are intensifying or not helping clinicians to cope with perceived inappropriateness of care.^{3,22- 23,32,37} In our study, the variability in perceived inappropriateness of care was largely associated with differences in the ethical environment across ICUs. For example, perceived inappropriateness of care was less common in ICUs in which physicians and nurses had a certain degree of job autonomy, an acceptable workload, and a high level of interdisciplinary collaboration and decision making. Interventions aimed at improving these factors may decrease the likelihood of perceived inappropriateness of care via both an effect on subjective determinants of perceived inappropriateness of care and improved objective matching of the level of care to the expected outcome.

Another interesting finding from our study is the strong link between perceived excessive workload and perceived inappropriateness of care among the nurses only. Conceivably, nurses may be more likely to suffer from a perceived imbalance between the efforts they expend in caring for the patients and the perceived likelihood that their efforts will be rewarded by better patient outcomes.³⁸ Furthermore, nurses spend considerable time at the bedside and are consequently more acutely aware of the suffering of their patients than are the physicians.^5,15,39- 40 Another possible factor is that the medical decisions lie chiefly in the hands of the physicians, with the nurses being asked to accept and to execute those decisions.^15,39- 40 Perceived powerlessness is a key determinant of moral distress in nurses and is related to a lack of collaboration in patient-care decision making.^5,23,33 Integrating the perspectives of nurses and the physicians may lead not only to greater mutual understanding with fewer conflicts,⁴¹ but also to better end-of-life decision making and care for the patients and their families.^{13,39,42- 46} Teaching individual ICU clinicians to create a symbolic distance from their work experiences and outcomes by becoming aware of their own personal values and beliefs might be another effective intervention.^{32,38,47- 48} Realizing that there are different ways of thinking about moral issues can help the clinicians understand their own process of decision making and tolerate differences both in other clinicians' moral reasoning and decision making and in patients’/families' moral reasoning.^11,35,37 As such, disagreeing on the appropriateness of care and openly discussing these different views may be the starting point of good quality decision making truly adapted to the needs and preferences of the patient (or the family in case of incompetence).^{34,37,47- 49}

The challenge for ICU managers is thus to create ICUs in which self-reflection, mutual trust, open communication, and shared decision making are encouraged in order to improve the well-being of the individual clinicians and, thereby, the quality of patient care.

First, the study was not facilitated in a randomly selected sample of countries and ICUs. We chose to work with motivated national coordinators and local investigators to obtain high response rates and therefore to draw sound conclusions about the participating ICUs.

Second, patient coding was not performed in 13 of the 82 ICUs and our evaluation of agreement among clinicians regarding perceived inappropriateness of care for individual patients was consequently incomplete.

Third, a longitudinal study design would be needed to infer causal relationships between perceived inappropriateness of care and burnout or intent to leave. A longitudinal study might also allow an evaluation of the moral residue left by each instance of perceived inappropriateness of care in a given clinician.^3- 4,12

In conclusion, perceived inappropriateness of care is common among nurses and physicians in ICUs and is significantly associated with an intent to leave the current clinical position, suggesting a major impact on clinician well-being. The main reported reason for perceived inappropriateness of care is a mismatch between the level of care and the expected patient outcome, usually in the direction of perceived excess intensity of care. Perceived inappropriateness of care is a subjective factor that does not necessarily indicate a failure to adhere to recommendations for patient care but that may serve as a marker for inadequate communication, decision sharing, and job autonomy within the ICU.