Moral distress and attitudes toward intramural goal-of-care discussions

Introduction

Advance care planning (ACP) is a broad term used to describe the process and outcomes of planning for one’s future socio-medical care. ACP focuses on two key areas: first, the person’s values, preferences, and goals in general and regarding future care; second, medical conditions, treatment (options), and prognostic factors. Ideally, ACP produces some form of documentation that is accessible to relevant individuals if the person becomes incapacitated. ¹ One such form is an advance directive (AD). In Austria, ADs are legally recognised written declarations in which a person prospectively refuses specific medical treatments (typically life-sustaining measures), and this takes effect if they lose decision-making capacity when the treatment decision arises. ²

In clinical practice, ACP is often implemented through goals of care (GOC) discussions, which focus on the current hospital stay or treatment plan. These discussions may be documented to guide treatment decisions. ³ Both ACP and GOC discussions aim to ensure that medical care aligns with the patient’s values, preferences, and goals, thereby respecting their autonomy. Achieving goal-concordant care, in which treatment matches the patient’s preferences, is widely regarded as the most important outcome of ACP. ⁴

In Austria, the overall prevalence of ADs within the health care sector remains low. At the end of 2022, the country’s two largest AD registries (the Patientenverfügungsregister der österreichischen Rechtsanwälte and the Patientenverfügungsregister des österreichischen Notariats) contained 67,419 entries, about half of which were registered in the preceding 5 years. ⁵ This represents roughly 0.89% of the adult population (≥18 years) as of 31 October 2023 (7,567,393 individuals ⁶ ). It should be noted that some ADs may be recorded in both registries, so the exact number is likely lower. The central registry of legal surrogate decision makers (Österreichisches Zentrales Vertretungsverzeichnis) does not publish up-to-date numbers. According to one report, it had 176,000 entries in June 2021. ⁷

It seems reasonable to assume that AD prevalence is indicative of general ACP activity, which has not been studied in Austria to date. This is particularly relevant given that ACP has been shown to reduce decision conflict and psychological burden in clinical care. ⁸ Taking part in treatment that is seen as overly aggressive or not in the best interest of the patient might lead to moral distress for health care providers (HCPs). ⁹

Moral distress was first described loosely by Jameton in 1984 as a condition that arises when institutional constraints do not allow individuals to do what they believe is the right thing, specifically in nursing practice. ¹⁰ In the scientific literature, it was later coined by Corley et al. in 2001, who defined it as a ‘painful psychological disequilibrium that results from recognizing the ethically appropriate action, yet not taking it, because of such obstacles as lack of time, supervisory reluctance, an inhibiting medical power structure, institution policy, or legal considerations’ (pp. 250f). ¹¹ It has been associated with burnout and job turnover.^12–14

Aims of the current study

In light of the low prevalence of ADs in Austrian hospitals, this study aimed to assess attitudes toward GOC discussions among patients and HCPs in general wards, identify barriers to and facilitators of their implementation, and compare levels of moral distress among HCPs.

Methods

The current study is a cross-sectional survey conducted in six wards of Vienna General Hospital. The study protocol was presented at a hospital-wide kick-off meeting, and wards were recruited for participation on a voluntary basis. These included cardiology, pulmonology, infectious diseases, maxillofacial surgery, palliative care, and chronic haemodialysis. The survey was conducted in a paper-based format over 4 weeks at each ward, from October 2022 to June 2023. All patients and HCPs were asked to take part in a voluntary questionnaire survey. During the 4 weeks, all newly admitted patients (and their caregivers, if present) were approached upon admission to the ward. HCPs were recruited through an information folder situated at the central hub of each ward. Additionally, informational posters were placed in the admission area of each ward.

We wanted to assess the following:

(1) Whether the respective wards had existing standard operating procedures (SOPs) concerning GOC discussions;

A comprehensive list of variables is provided in Table 1.

OPEN IN VIEWER

Table 1.

Variables.

Patients and caregivers
• Satisfaction with the current treatment
• Is a GOC discussion appropriate?
• Has the discussion been helpful, burdening or relieving (or, in case it never happened before, would it be)?
HCPs
• Existence and execution of SOPs for GOC discussions
• Resources to facilitate GOC discussions (i.e. staff, material, time, processes, translators)
• Possibility of creating an AD during the hospital stay
• Moral distress

To assess all of the above, we created three surveys, one for HCPs, one for patients, and one for the patients’ caregivers. The survey was discussed and adapted in a multidisciplinary expert panel consisting of experts in anaesthesiology and intensive care medicine, palliative medicine, psychology, law, and ethics. The patients’ and caregivers’ surveys were both professionally translated into 11 languages and thus available in German, English, French, Polish, Bosnian/Croatian/Serbian, Arabic, Farsi, Turkish, Russian, Ukrainian, and Romanian. The English version of the questionnaire is included in the supplement to this article.

In addition, the HCP survey included the Measure of Moral Distress – Health Care Professionals (MMD-HP), ⁹ which allowed for the assessment of moral distress and the comparison of professions in this regard. The MMD-HP is the most recent revision of the Moral Distress Scale, ¹¹ which has been translated into many languages^15–20 and is used extensively worldwide. The MMD-HP has been translated into Spanish, ²¹ Swedish, ²² and Japanese ²³ but not yet into German. In the validation study of the English version, Cronbach’s alpha was 0.93 for the overall sample and >0.90 for the subsamples (nurses, physicians, other HCPs). ⁹ For this study, we used a German translation, which we created using to-and-back translation, as well as a discussion and cultural adaptation by the expert panel described above. The MMD-HP can be divided into four subscales and yields a score for each subscale in addition to a total score. ⁹ Higher scores refer to higher moral distress. Table 2 explains the subscales in more detail.

OPEN IN VIEWER

Table 2.

The four main factors of the MMD-HP, as identified by Epstein et al. ⁹

Factor	Three example items
System-level root causes	• Be required to care for more patients than I can safely care for.
	• Experience compromised patient care due to a lack of resources/equipment/bed capacity.
	• Have excessive documentation requirements that compromise patient care.
Clinical root causes at patient level	• Continue to provide aggressive treatment for a person who is most likely to die regardless of this treatment when no one will make a decision to withdraw it.
	• Follow the family’s insistence to continue aggressive treatment, even though I believe it is not in the best interests of the patient.
	• Feel pressured to order or carry out orders for what I consider unnecessary or inappropriate tests and treatments.
Team-level root causes (colleagues)	• Fear of retribution if I speak up.
	• Witness a violation of a standard of practice or a code of ethics and not feel sufficiently supported to report the violation.
	• Participate in care that I do not agree with, but do so because of fear of litigation.
Team-level root causes (patient)	• Feel pressured to ignore situations in which patients have not been given adequate information to ensure informed consent.
	• Be part of a team that gives inconsistent messages to a patient/family.
	• Be required to care for patients who have unclear or inconsistent treatment plans or who lack goals of care.

The HCP and patient surveys were analysed using descriptive statistics. Results were calculated using relative frequencies, accounting for missing answers.

In the patient group, subgroups were compared using a Kruskal-Wallis test to analyse whether attitudes toward GOC discussions differed between genders, religions, and level of education.

The MMD-HP total and subscale scores were compared between physicians and nurses using a Mann–Whitney U-test since our sample did not meet the criteria for parametric testing. Results were deemed significant at α = 0.05. A multiple linear regression was modelled to analyse possible predictors for moral distress beyond profession, where the total moral distress score was the dependent variable, while gender, working experience, income and age were included as predictors. The results from the MMD-HP in our study were also compared with the three studies from the USA, ⁹ Japan, ²³ and Spain ²¹ using descriptive statistics.

All analyses and calculations were executed using IBM SPSS 24. ²⁴

Ethical considerations

The study was approved by the ethics commission of the Medical University of Vienna (vote number 1546/2022) as well as the data protection commission of the Medical University of Vienna. Only participants of 18 years of age or older were included and informed consent was obtained from all participants. The questionnaires contained no identifying information and were randomised. After digitalisation, the questionnaires were stored in a locked safe at the research institute. Participation was voluntary and not participating carried no disadvantages. In case of adverse emotional events, we were given full institutional support by the department of clinical psychology and psychotherapy at the Vienna General Hospital.

Results

Forty-seven patients and 75 HCPs participated in the assessment. Tables 3 and 4 show the sample characteristics for the patient and HCP groups, respectively.

OPEN IN VIEWER

Table 3.

Sample characteristics – patients.

​	% (n = )
Total	100 (N = 47)
Sex	100 (N = 46)
Female	41.3 (19)
Male	56.5 (26)
Diverse	2.2 (1)
Age	100 (N = 45)
<25	0 (0)
25–34	11.1 (5)
35–44	4.4 (2)
45–54	11.1 (5)
55–64	35.6 (16)
65–74	24.4 (11)
75–83	13.3 (6)
>83	0 (0)
Religious orientation	100 (N = 45)
Christian	44.4 (20)
Muslim	8.9 (4)
Other	2.2 (1)
Not religious	44.4 (20)
Highest education	100 (N = 46)
No secondary education	4.3 (2)
Mandatory school years	15.2 (7)
High school diploma	19.6 (9)
Training/middle school a	43.5 (20)
Tertiary education	17.4 (8)
Income in €/m	100 (N = 41)
<1300	34.1 (14)
1300–1700	24.4 (10)
1701–2300	24.4 (10)
2301–3000	12.2 (5)
>3000	4.9 (2)

OPEN IN VIEWER

Table 4.

Sample characteristics – HCPs.

​	% (n = )
Total	100 (N = 75)
Profession
Physician	36.0 (27)
Nurse	53.3 (40)
Other a	10.7 (8)

Speciality	Physicians (N = 27)	Nurses (N = 41)
Internal medicine	70.4 (19)	68.3 (28)
Surgical	29.6 (8)	31.7 (13)

Experience in years	Physicians (N = 27)	Nurses (N = 40)
<2	25.9 (7)	10.0 (4)
2–5	25.9 (7)	30.0 (12)
5–10	22.2 (6)	10.0 (4)
10–15	18.5 (5)	15.0 (6)
15–20	0 (0)	2.5 (1)
>20	7.4 (2)	32.5 (13)

HCP = healthcare provider.

^aIncludes psychologists, therapists, and dietologists.

Patient survey

Figure 1 shows the results of the patient survey. A total of 248 surveys were distributed, from which 47 (19%) were returned. Most (n = 44) surveys were answered in German, while the remaining three surveys were answered in Romanian, Arabic and Bosnian/Serbian/Croatian, respectively. A little over half of the patients answered that they had never been asked about their values, wishes, and goals, while around one-third answered that they had. The vast majority had never been asked about their ADs, while only one out of nine had ever been asked about them during a hospital stay. The general attitude toward a GOC discussion during a hospital stay was positive. When asked whether a discussion of values, wishes, and goals was appropriate during a hospital stay, the vast majority answered ‘yes’ and less than 5% answered ‘no’. Three out of four said they would feel little to no burden due to the discussion, while almost the same number said they would feel some relief or be very relieved by the discussion. The vast majority answered that the discussion was somewhat or very helpful, while less than 5% described the discussion as either somewhat or very unhelpful.

OPEN IN VIEWER

Figure 1.

Results from the patient survey (in percent).

No significant differences were found between patients of different gender, religion or level of education.

HCP survey

Figure 2 shows the results of the HCP survey. The majority of HCPs answered that GOC discussions are suitable for 50% of patients or more, and less than one out of seven answered that they are suitable for less than 25% of patients. One-third answered that even healthy patients are suited for GOC discussions. The majority considered GOC discussions suitable for patients with multimorbidity (80.0%) and life-shortening conditions (86.7%) but also with a simple medical condition (54.7%). For patients at the end of life, two-thirds of HCPs regarded the GOC discussions as suitable.

OPEN IN VIEWER

Figure 2.

Results of the HCP survey (in percent). GOC = goals of care; SOP = standard operating procedure.

The resource assessment revealed a general scarcity of resources. Most HCPs reported insufficient access to personnel (54.1%), time (56.2%), translators (75.7%), or material (71.6%) required to assess the patient’s values, wishes, and goals and to formulate corresponding GOC for the current hospital stay. Almost half reported that institutional processes to facilitate this were largely absent.

Half of HCPs practice GOC discussions for less than 10% of patients. One in five does it for more than half of patients and only 10% for over 75% of patients.

Moral distress

In the current study, the mean total MMD-HP score was 105.01. Nurses reported a significantly higher total score than physicians (125.66 vs 86.85, p = 0.020). In the subscales, nurses reported significantly higher scores for systemic (49.24 vs 34.30, p = 0.017) and patient factors (33.95 vs 15.85, p < 0.001). No difference was found regarding the two team factors, working with colleagues (16.53 vs 15.59, p = 0.274), and working with the patient’s caregivers (25.95 vs 12.11, p = 0.263, see also Figure 3).

OPEN IN VIEWER

Figure 3.

Comparison of moral distress scores between physicians and nurses. Comparison between physicians’ and nurses’ MMD-HP total and subscale scores. Data are presented as boxplots with outliers. *statistically significant at α = 0.05.

Our linear regression model met the criteria of linearity, independence of observations, homoscedasticity, independence of residuals, and lack of multicollinearity. The model was significant (F [4, 63] = 3.093, p = .022) and explained 11.1% of the variance (adjusted R² = 0.111). The working experience of HCPs was revealed as the only positive predictor for moral distress (β = .42, p = .046), while gender (β = −.12, p = .328), income (β = .08, p = .508) and age (β = −.11, p = .596) were not significant.

Overall scores from our sample (106.51) were comparable to the samples from the USA ⁹ (108.9) and Japan ²³ (98.2) but a lot higher than those in the Spanish ²¹ sample (68). Looking at nurses, the overall score was highest in our sample (125.66), followed by the samples from the USA ⁹ (112.3) and Japan ²³ (97.2) and more than double that of the Spanish ²¹ sample (61). Physicians scored lower in our sample (86.85) than in those from the USA ⁹ (96.3) and Japan ²³ (104.0), and the scores were comparable to those of the Spanish ²¹ sample (61). The results from the current study are consistent with the data from the USA, ⁹ although the mean difference was larger in our sample (38.79, p = 0.020 vs 16.02, p = 0.023). All result data are provided in table format in the supplementary file.

Discussion

This study demonstrates the broad acceptance of goal-of-care (GOC) discussions among hospitalised patients, who largely perceived them as appropriate, helpful, and relieving. Nevertheless, only a small proportion had previously been asked about ADs during a hospital stay.

Although most healthcare professionals considered GOC discussions appropriate, particularly for patients with complex or life-limiting conditions, they reported conducting them infrequently in routine practice. Key barriers included the absence of structured procedures, limited time and personnel resources, language barriers, and insufficient institutional support.

The findings regarding our assessment of attitudes toward GOC discussions in the general wards of the Vienna General Hospital strongly suggest that GOC discussions are of high relevance for most patients in the hospital setting. We included patients and HCPs from a variety of different medical specialities, ranging from acute to chronic care. Our sample included young, otherwise healthy patients admitted for maxillofacial surgery; multimorbid cardiology and pulmonology patients; patients on chronic haemodialysis; and those in palliative care at the end of life. The results were similar across all specialities. Most patients considered the discussions to be appropriate (68.1%), relieving (71.1%), at least somewhat helpful (80.5%), and to not present a burden (71.8%). Only a small minority of patients considered them not appropriate (6.4%) or not helpful (2.2%). This is consistent with data from Germany on patients scheduled for surgery, where only 7.3% felt that GOC discussions prior to surgery presented a burden. ²⁵ The prevalence of legal ADs in the Austrian general population is less than 1%. This is noteworthy, given the high acceptance of GOC discussions in the current study. The discrepancy may be explained in part by regulatory and financial barriers. In Austria, the establishment of a legally binding AD requires consultations with a physician and a legal representative, both of which constitute private liabilities. ²⁶

The low prevalence is similarly reflected in the clinical setting. In the only cross-sectional study assessing the prevalence of ADs and legally appointed surrogate decision makers in ICUs across Austria, among 475 patients, 0.6% had an AD and 0.8% had a legal surrogate decision maker. ²⁷ These findings are concerning, as a substantial proportion of ICU patients experience at least temporary decisional incapacity while simultaneously facing an increased risk of clinical deterioration and the potential need to escalate life-sustaining therapies. The limited availability of documented ADs or designated surrogate decision makers may therefore impede timely, preference-concordant decision-making in critical care contexts.

Moral distress levels were substantial overall and significantly higher among nurses than among physicians. In our sample, working experience was the only demographic predictor for moral distress, albeit explaining only 11% of the variance. HCPs with more working experience were more likely to have higher total moral distress scores. Despite the relatively small predictive value, the results are concerning and underline the need for measures to mitigate moral distress, which seems to accumulate over time.

Our study is the first to assess moral distress among Austrian HCPs. Overall levels were comparable to findings from the USA and Japan. However, the gap between nurses and physicians was notably higher in our sample (38.8 points in the total score) than in the US study (16.0 points). ⁹ This difference may partly reflect cultural factors but is also likely influenced by study settings. In the US study, participants worked exclusively in acute care environments (ICU, operating room, emergency department, and outpatient clinics), with no representation from internal medicine or general inpatient care. In contrast, the majority of participants in our study were employed in internal medicine (see Table 4), which may have contributed to the discrepancy.

The higher moral distress scores observed among nurses in our sample are not unexpected and likely reflect the demanding work environment and structural pressures inherent in nursing practice. A report from the Austrian Ministry of Health showed that nurses in Austria take approximately 30% more sick leave than the European average and report the second-highest rates of work-related health problems (35%, second only to Finland). Sixty-eight percent experience high levels of daily fatigue. ²⁸

In the context of demographic shifts, workforce projections estimate that more than 40% additional staff will be required in Austria until 2030. ²⁹ Against this backdrop, reducing moral distress should be considered a strategic priority for the Austrian healthcare system, both to safeguard staff well-being and to ensure a sustainable workforce capacity.

Given that respect for patient autonomy is conceptually linked to reductions in moral distress, systematic integration of a patient’s values, preferences, and goals into clinical decision-making appears essential.

In a study by Umgelter et al., 579 patients at the Technical University of Munich scheduled for elective (noncardiac) surgery were surveyed regarding their preferences for medical care in case of an emergency or unplanned ICU admission. Forty-three percent expressed a desire to proactively address future medical care in such scenarios, while only 7.3% felt burdened by the assessment. The desire for autonomy was the sole independent predictor of patients’ willingness to engage in ACP prior to surgery. ²⁵

These findings support the broader implementation of perioperative GOC discussions. Although robust evidence regarding optimal methods and outcome measures for ACP in the perioperative setting remains limited, ³⁰ structured GOC discussions in hospitals represent a feasible approach. It should be noted, that the completion of forms and documentation alone has been found to potentially decrease the overall confidence in ACP and even cause affective harm to patients, caregivers and HCPs. ³¹ Because standardised forms usually focus on easily accessible variables, there is a risk of neglecting core features of the patient’s values, wishes and goals, which can only be assessed through (recurrent) informed discussions. Ideally, there should be institutional SOPs regarding GOC discussions, adapted to the specific setting, including the ideal time to have the discussion, re-evaluation strategies, whom to include and specific contents that are important for the respective setting. While a comprehensive assessment of patients’ values and preferences must remain central, implementation strategies should also be resource efficient to facilitate integration into routine clinical workflows.

Promoting patient autonomy through structured GOC discussions should benefit patients by enhancing goal-concordant care. Conceptually, GOC discussions should also benefit healthcare professionals by potentially mitigating moral distress. When treatment goals have been discussed with the patient, caregivers and the HCP team, some of the main factors that cause moral distress (following the family’s insistence on aggressive treatment; feeling pressured to ignore situations in which a patient has not been given adequate information to ensure informed consent; continue to provide aggressive treatment for a person who is most likely to die regardless of this treatment when no one will make a decision to withdraw it) ⁹ should be mitigated. However, it should be noted that there is currently no evidence regarding a causal relationship between the lack of GOC discussions or ACP in general and moral distress.

Limitations and outlook

The present study has several limitations. First, it was conducted at a single tertiary academic centre in Austria – albeit the largest hospital in the country. Second, the sample size is relatively small, although we included patients from a variety of medical specialities. Because the survey was voluntary, there is also the possibility for non-response as well as selection bias, which is also reflected in the overall response rate of only around 19%. Even though we tried to mitigate the risk by translating the survey into the most spoken languages, it is possible that patients were more willing to participate out of general interest in ACP or more reluctant to participate, for example, due to their cultural background or lack of interest in ACP. These factors limit the generalisability of the findings. Future research should aim to address these issues.

Third, although we aimed to include patients’ caregivers in the survey, COVID-19-related restrictions limited their participation to such a degree that we were not able to extract enough data for a meaningful analysis. Since caregivers may significantly partake in patient care as well as ACP, future research should aim to include them where possible.

Nevertheless, to the best of our knowledge, this is the first study to assess attitudes toward GOC discussions or some form of ACP in the Austrian hospital setting, as well as the first to examine moral distress in an Austrian HCP population.

While our findings indicate a clear need for structured GOC discussions, at least within our institutional context, the optimal implementation strategy remains uncertain. A universal lack of resources was identified as a barrier. However, the exact mechanisms how GOC discussions could be implemented into clinical routine remain unclear. Future studies, including a qualitative element, are needed to gain more insight.

Although conceptually a link between the lack of structured GOC discussions and moral distress seems feasible, the cross-sectional design of the current study cannot answer to this. Future research should therefore focus on developing and evaluating feasible implementation models and assessing their effects on patient autonomy and satisfaction, as well as on the moral distress experienced by HCPs.

Finally, the translation of the MMD-HP used has not yet undergone formal psychometric validation. Therefore, reliability and validity of the results may be compromised. The German translation of the MMD-HP should be validated in future studies. Nationally representative, validated data for Austria, or ideally for the broader German-speaking DACH (D – Germany, A – Austria, CH – Switzerland) region, would be highly valuable. Although moral distress is an inherent aspect of healthcare practice, minimising it should be a central objective of clinical institutions. Achieving this requires not only validated measurement instruments but also robust reference data from representative samples to enable meaningful benchmarking and evaluation.

Supplemental material