Clinicians' Perceptions of Futile or Potentially Inappropriate Care and Associations with Avoidant Behaviors and Burnout

Abstract

Background: Futile or potentially inappropriate care (futile/PIC) for dying inpatients leads to negative outcomes for patients and clinicians. In the setting of rising end-of-life health care costs and increasing physician burnout, it is important to understand the causes of futile/PIC, how it impacts on care and relates to burnout.

Objectives: Examine causes of futile/PIC, determine whether clinicians report compensatory or avoidant behaviors as a result of such care and assess whether these behaviors are associated with burnout.

Design: Online, cross-sectional questionnaire.

Setting/Subjects: Clinicians at two academic hospitals in New York City.

Methods: Respondents were asked the frequency with which they observed or provided futile/PIC and whether they demonstrated compensatory or avoidant behaviors as a result. A validated screen was used to assess burnout.

Measurements: Descriptive statistics, odds ratios, linear regressions.

Results: Surveys were completed by 349 subjects. A majority of clinicians (91.3%) felt they had provided or “possibly” provided futile/PIC in the past six months. The most frequent reason cited for PIC (61.0%) was the insistence of the patient's family. Both witnessing and providing PIC were statistically significantly (p < 0.05) associated with compensatory and avoidant behaviors, but more strongly associated with avoidant behaviors. Provision of PIC increased the likelihood of avoiding the patient's loved ones by a factor of 2.40 (1.82–3.19), avoiding the patient by a factor of 1.83 (1.32–2.55), and avoiding colleagues by a factor of 2.56 (1.57–4.20) (all p < 0.001). Avoiding the patient's loved ones (β = 0.55, SE = 0.12, p < 0.001), avoiding the patient (β = 0.38, SE = 0.17; p = 0.03), and avoiding colleagues (β = 0.78, SE = 0.28; p = 0.01) were significantly associated with burnout.

Conclusions: Futile/PIC, provided or observed, is associated with avoidance of patients, families, and colleagues and those behaviors are associated with burnout.

Introduction

Perceptions of futility, rather than stressful emergencies, may be the cause of clinician burnout.¹ Clinician-perceived futile care or potentially inappropriate care (futile/PIC), which can be defined as care unable to achieve its intended goal² and care that is inconsistent with the provider's professional knowledge or beliefs,³ respectively, is a controversial topic.^2,4–6 Clinician-perceived futile/PIC has been shown to cause moral distress, depression, and other negative outcomes,^3,7–11 and new research describing a direct relationship with burnout, feelings of exhaustion, depersonalization, and reduced personal accomplishment¹² has recently emerged.¹³ Additionally, while it might not be surprising that clinicians who are compelled to provide futile/PIC experience negative downstream effects, little is known about the ways in which providing futile/PIC to a patient might affect a clinician's behavior toward that patient and his/her family.

Studies find that the emotional burden and demoralization of futile/PIC may contribute to physician turnover^7,12 and future intensive care unit (ICU) staffing shortages.¹² Aside from the considerable amount of distress reported by clinicians required to provide futile/PIC,^3,7,8 this care has also been shown to contribute to the high cost of end-of-life care.^14,15 Estimates suggest that between 33% and 38% of patients receive nonbeneficial care, a term closely associated with PIC, at the end of life.¹⁶ In the United States the cost of end-of-life care is remarkable, with 30% of expenditures for Medicare going to the 5% of beneficiaries who die each year.¹⁷ Most of the expense can be attributed to burdensome and potentially inappropriate life-sustaining care such as mechanical ventilation and resuscitation. In fact, acute care in the final 30 days of life contributes to nearly 80% of the costs incurred in the last year of life.¹⁷

As the body of literature on clinician-perceived futile/PIC and its consequences develops^{3,7,8,12,16,18–22} there are a number of questions that have yet to be investigated. Do clinicians change their behavior as an immediate result of providing futile/PIC? If these changes in behavior do occur, are they associated with burnout, and do they reveal a possible target for intervention? In the context of the emotional toll on clinicians, the quality of care for patients and the high cost to taxpayers, this article reports on a research project undertaken in the United States to explore whether clinician-perceived futile/PIC for dying inpatients is associated with avoidant and/or compensatory behaviors, and if these behaviors are associated with burnout.

Methods

Study design and participants

This study involved the completion of a cross-sectional, self-administered online questionnaire sent to registered nurses (RN), physician assistants (PA), nurse practitioners (NP), attending physicians, residents, and fellows. Participants eligible to complete the survey were identified as employees of New York Presbyterian Hospital (NYPH), New York-Presbyterian/Queens (NYPQ), or Weill Cornell Medicine, and worked in the fields of medicine, surgery, neurology, and intensive care. NYPH and NYPQ are large, urban academic medical centers in northeastern United States, each with remarkably diverse patient populations. NYPH is a referral center, and NYPQ is a community-based teaching hospital. To meet inclusion criteria, participants were required to have been employed at least part-time for the past six months, and consulted for at least five inpatients per month. To protect anonymity, data about age, sex, race, and years of practice were not collected. The Institutional Review Boards at both NYPH and NYPQ approved this study.

Instruments

Clinician-perceived futile and/or PIC questionnaire

No validated instruments for measuring clinicians' perceptions of futile/PIC have been developed. Consequently, psychologist from Weill Cornell Medicine, in consultation with clinical staff at NYPH and NYPQ who work in intensive care and palliative medicine, developed a questionnaire for assessing perceptions of the prevalence, causes, and outcomes of clinician-perceived futile/PIC delivered to dying inpatients. For example, one questions asks, “In the last six months, have you provided care that you consider futile (i.e., a treatment or intervention that you do not believe will have a health benefit), or potentially inappropriate interventions, to dying patients?” Responses were recorded on Likert scales. An iterative approach to survey development was used until agreement was reached on wording and content. The survey was tested on a small sample of clinicians for clarity of questions and time for completion before being released to the target populations.

Burnout

To assess burnout, a single item screen developed by Schmoldt et al.²³ and validated separately by both Rohland²⁴ and Hansen and Girgis²⁵ was used. This single item screen was validated against the 22-item Maslach Burnout Inventory-Human Services Survey.²⁶ The screen asks, “Overall, based on your definition of burnout, how would you rate your level of burnout?” The possible responses were as follows: (1) I enjoy my work. I have no symptoms of burnout. (2) Occasionally I am under stress, and I don't always have as much energy as I once did, but I don't feel burned out. (3) I am definitely burning out and have one or more symptoms of burnout, such as physical and emotional exhaustion. (4) The symptoms of burnout that I'm experiencing won't go away. I think about frustration at work a lot. (5) I feel completely burned out and often wonder if I can go on. I am at the point where I may need some changes or may need to seek some sort of help.

Avoidant and compensatory behaviors

No validated instruments have been developed for assessing avoidant and compensatory behaviors in the in-patient, clinical setting. Consequently, in consultation with psychologists and experienced psychometricians at Weill Cornell Medicine who have expertise in research on end-of-life care, and in collaboration with two physicians working as intensivists and two physicians and one NP working in palliative medicine, questions were developed for assessing avoidant and compensatory behaviors of relevance to the delivery of futile/PIC. These behaviors were selected for investigation because compensatory and avoidant behaviors are purported to be evoked by guilt and shame, respectively,^27–29 and guilt and shame have been highlighted in the PIC and moral distress literature.^18,30 In response to the questions “Did you do any of the following due to providing clinician-perceived futile or potentially inappropriate care?” and “Did you do any of the following due to witnessing futile care?” the subjects were asked to choose as many behaviors as were applicable. An example of a compensatory behavior was “treating pain more aggressively” and an example of an avoidant behavior was “avoiding the patient” (see Table 3 for complete list of behaviors). Compensatory and avoidant behaviors were derived from the clinical experience of the researchers and by asking a sample of physicians and nurses, not included in the analysis, how they compensate or avoid patients in their care. The reliability and validity of this tool has not been measured.

Table 3.

Behaviors as a Result of Providing or Witnessing Potentially Inappropriate Care

Behaviors	I activated support servicesa	I responded quicker to complaintsb	I treated pain more aggressivelyc	I made sure the patient wasn't lonelyd	I avoided the patiente	I avoided the patient's loved onesf	I avoided colleaguesg	I transferred the patienth
Did you do any of the following due to PROVIDING futile care? n (%)*	108 (32.4)	88 (26.4)	87 (26.1)	57 (17.1)	29 (8.7)	90 (27.0)	10 (3.0)	12 (3.6)
Association of behaviors with amount of potentially inappropriate care provided, OR (CI)	1.29 (1.04–1.60)	1.34 (1.06–1.68)	1.52 (1.20–1.92)	1.39 (1.08–1.80)	1.83 (1.32–2.55)	2.40 (1.82–3.18)	2.56 (1.57–4.20)	1.59 (1.00–2.52)
	p = 0.02	p = 0.01	p < 0.001	p = 0.01	p < 0.001	p < 0.001	p < 0.001	p = 0.05
Did you do any of the following due to WITNESSING futile care? n (%)*	65 (19.5)	56 (16.8)	37 (11.1)	37 (11.1)	38 (11.4)	73 (21.9)	18 (5.4)	12 (3.6)
Association of behaviors with amount of potentially inappropriate care witnessed, OR (CI)	1.38 (1.12–1.70)	1.47 (1.18–1.83)	1.53 (1.18–1.98)	1.27 (0.98–1.64)	1.66 (1.28–2.16)	1.74 (1.41–2.16)	2.09 (1.44–3.02)	1.34 (0.88–2.03)
	p = 0.02	p = 0.01	p < 0.001	p = 0.01	p < 0.001	p < 0.001	p < 0.001	p = 0.05

The full answers on the survey read as follows:

^aI activated supportive services to make the patients stay more comfortable such as calling the chaplain, getting the patient a room with a window, getting the patient an iPad, etc.

^bI responded more quickly to the patient's complaints of symptoms.

^cI treated the patient's pain more aggressively.

^dI made sure the patient wasn't lonely more than I usually would.

^eI avoided the patient.

^fI avoided the patient's loved ones.

^gI avoided colleagues involved in the patients care.

^hI transferred the patient's care to another physician because I did not want to participate in futile or potentially inappropriate treatment.

^*Percentages sum to greater than 100 because participants could select more than one behavior.

CI, confidence interval; OR, odds ratio.

Statistical analysis

Descriptive statistics (e.g., frequencies) were reported for the study participants. Logistic regression models estimated the likelihood of specific avoidant or compensatory behaviors associated with the observation or provision of clinician-perceived futile/PIC. Linear regression models estimated the effects these behaviors have on the clinician's reported degree of burnout. Both bivariable and multivariable models were estimated, the latter entering all the behaviors simultaneously to determine effect of each behavior over and above the effect of the other behaviors associated with burnout. Statistical analysis was conducted using SAS statistical software, version 9.4 (Cary, NC). Statistical inferences were based on two-sided tests with p < 0.05 considered statistically significant.

Results

We invited 1784 RNs, PAs, NPs, residents, fellows, and attendings via email to complete the survey. Data were collected online from May to August of 2017. We had 349 clinician responses, yielding a response rate of 19.6%. Sixteen of the respondents did not meet eligibility criteria and were excluded from the analysis. Table 1 shows the distribution of responses by both job position and department. RNs represent the largest percent of respondents at 40.5%. Of the departments represented, internal medicine is the largest group at 39.9% with intensive care second at 35.7%. Job position data and department data were missing for 2.1% and 5.1% of respondents, respectively.

Table 1.

Distribution of Responses by Position and Department (n = 333)

Position	Number (%)a
RN	135 (40.5)
Resident/fellow	72 (21.6)
Attending	67 (20.1)
PA/NP	52 (15.6)
Missing	7 (2.1)
Total	333 (100)

Department	Number (%)a
Medical	133 (39.9)
Intensive care	119 (35.7)
Surgical	29 (8.7)
Neurological	19 (5.7)
Missing	17 (5.1)
Medical-surgical	16 (4.8)
Total	333 (100)

^aPercentages do not total 100 due to rounding.

NP, nurse practitioners; PA, physician assistants; RN, registered nurses.

When asked whether they had provided clinician-perceived futile/PIC in the past six months, 75.7% of all clinicians responded “yes,” 15.6% “possibly,” and 8.7% “no” (Table 2). When asked what they consider to be the main reason that futile/PIC is provided, the “patient's family” was cited most frequently at 61.0%, with “physician or consultant request” second at 11.1% (Fig. 1). This question was not answered by 7.2% of respondents. Overall, 43% of respondents screened positive for burnout, which is similar to rates of burnout found in other reports.¹²

Table 2.

Prevalence of Perceived Futile or Potentially Inappropriate Care

“In the past 6 months, have you provided care that you think is futile or potentially inappropriate to dying inpatients?” [Number (%)]
Position	No	Possibly	Yes
RN	9 (6.7)	17 (12.6)	109 (80.7)
PA/NP	6 (11.5)	12 (23.1)	34 (65.4)
Resident/fellow	3 (4.2)	8 (11.1)	61 (84.7)
Attending	10 (14.9)	15 (22.4)	42 (62.7)
Position not provided	1 (14.3)	0 (0)	6 (85.7)
Total	29 (8.7)	52 (15.6)	252 (75.7)

FIG. 1.

“What do you consider the main reason for providing futile care?”

Clinician self-report of providing futile/PIC was significantly (p < 0.05) associated with both compensatory and avoidant behaviors (Table 3). We found that as a result of providing futile/PIC clinicians were significantly more likely to avoid colleagues involved in the patient's care by a factor of 2.56 (1.57–4.20), avoid the patient's loved ones by a factor of 2.40 (1.82–3.19), and avoid the patient by a factor of 1.83 (1.32–2.55) (all p < 0.001).

Also shown in Table 3, witnessing others provide futile/PIC was also associated with both compensatory and avoidant behaviors. The highest odds ratios for clinicians who reported avoidant behaviors as a result of witnessing others provide PIC are as follows: avoid colleagues involved in the patient's care 2.09 (1.44–3.02), avoid the patient's loved ones 1.74 (1.41–2.16), and avoid the patient 1.66 (1.28–2.16) (all p < 0.001).

In bivariable models, all of the compensatory and avoidant behaviors were positively associated with burnout and only two did not achieve statistical significance (Table 4). These two behaviors were as follows: responding quicker to complaints and transferring the patient. Of the avoidant behaviors, it was found that avoiding colleagues (β = 0.78, SE = 0.28; p = 0.006), avoiding the patient's loved ones (β = 0.55, SE = 0.12, p < 0.001), and avoiding the patient (β = 0.38, SE = 0.17; p = 0.03) were all significantly associated with burnout. Of the compensatory behaviors, it was found that making sure the patient was not lonely more than usual (β = 0.32, SE = 0.13; p = 0.01), treating the patient's pain more aggressively (β = 0.26, SE = 0.11; p = 0.02), and activating support services (β = 0.24, SE = 0.11; p = 0.02) were also significantly associated with burnout. When all eight behaviors were estimated in a multivariable model, “I avoided the patient's loved ones” emerged as the behavior most strongly associated with burnout (β = 0.49, SE = 0.13; p < 0.001).

Table 4.

Associations between Clinician Burnout with Avoidant and Compensatory Behaviors

	Bivariate models (unadjusted estimates)			Multivariate models (adjusted estimates)
Behaviorsa	Parameter estimate	Standard error	p	Parameter estimate	Standard error	p
I activated support services	0.24016	0.10573	0.0238	0.07327	0.12107	0.5455
I responded quicker to complaints	0.13829	0.11214	0.2185	−0.18152	0.1468	0.2173
I treated pain more aggressively	0.26145	0.11179	0.02	0.05751	0.15074	0.7031
I made sure the patient wasn't lonely	0.31731	0.12918	0.0146	0.29629	0.16131	0.0673
I avoided the patient	0.38364	0.17179	0.0263	−0.13069	0.20453	0.5233
I avoided the patient's loved ones	0.54762	0.1071	<0.0001	0.48903	0.12876	0.0002
I avoided colleagues	0.77586	0.2816	0.0062	0.48618	0.30799	0.1155
I transferred the patient	0.46875	0.2598	0.0722	0.29157	0.25641	0.2564

^aPlease see Table 3 for full description of behaviors.

Discussion

Our study contributes to the small, but growing, body of literature on the consequences of futile/PIC.^{3,5–8,12,16,18–22,31} Prior research has been largely qualitative³² or studied prevalence,^3,7 cost¹⁴ or negative outcomes, but not how clinicians actually change their behaviors based on exposures to and perceptions of PICs.^10,12 We add to the literature by conducting a quantitative analysis of how clinicians report changing their behavior toward not only the patients, but also the patients' families and professional colleagues, as a direct result of providing (or observing) futile/PIC. We then showed that these behaviors are associated with burnout.

Our results are consistent with other findings as they reveal that in this cohort the majority of clinicians (91.3%) reported providing or “possibly” providing care that they perceive to be futile or potentially inappropriate,⁷ and the majority (61.0%) of clinicians cite family members as the main cause.^14,31 We find further that futile/PIC is associated with avoiding family members, and avoiding family members is the behavior most closely associated with feelings of burnout. Clinicians often feel family members are insisting on futile/PIC. Thus, clinicians blame the family members for compelling them to provide care that they do not want to provide, which causes clinicians to experience existential dilemmas,¹ and as a result they avoid the patient and the family.²⁹ Consequently, not only does the patient suffer from diminished care, but the soon-to-be bereft family members are also left isolated in a time of heightened vulnerability.³³

Improving communication between the health care team and the family may be one means of decreasing futile/PIC,^6,34 and consequently minimizing avoidant behaviors and reducing burnout. One approach shown to be effective is to designate a single, longitudinal contact person from both the health care team and the family.³⁵ Additionally, family meetings with multidisciplinary health care teams within the first 24–48 hours of admission to the ICU have been recommended by the American College of Critical Care Medicine.³⁶

Greater use of palliative care and chaplaincy services may be another means of reducing PIC.³⁷ Palliative care consults have been shown to help families clarify the plan of care to focus on quality of life.³⁸ Palliative care is expected to be inversely related to futile/PIC¹⁶ and can be used as a bridge to hospice.³⁸ Chaplaincy services provide emotional and spiritual support for families, which may help them come to terms with the severity of the patient's prognosis and forego aggressive treatment.^39,40

The limitations of this study include our response rate of 19.6%, which introduces the possibility of nonresponse bias in our results. Physicians and other clinicians are known to have low response rates to online surveys, citing lack of time and survey burden as common reasons.⁴¹ However, we have reason to believe that our sample reflects the general population as our reported frequencies of futile/PIC are similar to those reported elsewhere.^14,16 Additionally, it has been shown that responding and nonresponding health care providers share similar characteristics.⁴² It is also possible that the clinicians who did respond to our survey may have underreported futile/PIC and avoidant behaviors due to social desirability bias. As a result this is likely a conservative estimate of the frequency futile/PIC and its negative sequelae. This survey is also cross-sectional and limits our ability to draw conclusions regarding causal relationships.

In conclusion, this study found that perceived futile or PIC is associated with clinicians avoiding patients, family members, and other health care providers. These avoidant behaviors were associated with burnout, and burnout has been shown to contribute to the looming shortage of qualified clinicians working in settings such as the ICU.¹²

Author Disclosure Statement

No competing financial interests exist.