Impact of an Integrated Psychosocial Intervention for Health Care Professionals in an Intensive Care Unit: Controlled Quasi-Experimental Trial

Abstract

Intensive care units (ICUs) are challenging environments for healthcare professionals (HCPs), patients, and families. Up to 75% of HCPs experience psychosomatic symptoms and 30% intend to leave their job. To address this, a German ICU implemented an integrated psychosocial care model. Its implementation and evaluation are described. In a quasi-experimental design, a psychologist was integrated part-time into an ICU team for 7 months (June 2022-January 2023), tasked to deliver low-threshold psychosocial support to HCPs. Additional components were one-on-one interactions, clinical supervisions, a feedback survey and health circles. Pre- and post-measures from a voluntary sample of the intervention ward (N = 65) and a control ward receiving usual psychosocial care (N = 33) were taken via online surveys assessing the psychosocial safety climate (PSC), emotional and cognitive irritation and intention to leave. Groups were compared using linear mixed-effect models. The intervention was implemented as planned, and HCPs expressed satisfaction with the changes experienced on the ward. Statistical testing revealed no significant interaction effects between point of measurement and group allocation on irritation and intention to leave (P > .01). The PSC interaction effect did not reach the significance (α = .01, P = .02) but indicated a trend towards improvement in the intervention group. The intervention successfully provided accessible psychosocial support and fostered a more positive PSC within the ICU team. Statistical effects were limited, likely due to the quasi-experimental study design, small sample size and group imbalances, highlighting the need for larger, more controlled studies.

Keywords

integrated psychosocial intervention emergency care psychologist intensive care unit controlled quasi-experimental trial psychosocial safety climate irritation scale intention to leave

Introduction

There is broad evidence supporting the observation of rough and stressful working conditions in intensive care units (ICUs) since decades. About 28% of ICU nurses from 23 countries worldwide intend to leave their job.¹ Chronic work stress among ICU healthcare professionals (HCPs) are linked to burnout,² second victim traumatisation,³ depression, addiction and suicidality,⁴ all of them being connected to the workplace which some even perceive as a “hostile environment.”⁵ In addition to health related problems, psychosocial phenomena such as compassion fatigue or moral distress result from it; the latter substantially influencing job turnover intention.⁶ These consequences are not only an individual problem but also a public health matter: burnout is associated with reduced productivity,⁷ early retirement⁸ and compromised patient safety.^9-11

There are currently no nationwide structures in Germany to support ICU teams. Only about 20% of surveyed HCPs reported access to psychosocial support measures such as counselling, case discussions and supervision.^12,13 Although more and more hospitals are offering clinical supervision for HCPs, these programmes are limited in spontaneity and frequency and may be hindered by time, space and trust barriers.¹⁴

Support strategies targeting individuals may only be moderately effective in reducing psychological distress,¹⁵ as HCP stressors arise at a team or organisational level as well. Team-level stressors include for example, challenging relationships with team members,¹⁶ whereas organisational stressors encompass issues such as poor policy guidance on end-of-life decision making.¹⁷ Consequently, experts increasingly call for low-threshold acute psychosocial care^18,19 implemented through organisational interventions.²⁰ Such ratio-preventative measures may help strengthen team cohesion and foster social support¹⁵ to combat the aforementioned stressors.

The Psychosocial Safety Climate (PSC)²¹ theory provides a framework which links organisational culture with employee well-being. PSC describes how employees feel about whether their organisation considers their mental health in decision-making.

High PSC can act as a preventive stress measure for employees through the commitment of senior managers,²² enhancing employees’ physical health.²³ Organisational factors such as positive leadership can also reduce nurse turnover.²⁴ PSC can predict the persistence of depressive symptoms²⁵ and has a positive influence on job burnout²⁶ and work safety.²⁷ PSC is therefore increasingly used to measure the effects of workplace interventions.²⁸

Despite growing awareness, organisational approaches to psychosocial support remain scarce in German ICUs. Our experiences at the University Hospital in Ulm and at the Charité in Berlin during the Corona pandemic^29,30 showed that integrating psychologists into ICU teams on a permanent basis can be beneficial.³¹

In 2022, reacting to the need for psychosocial support identified by the chief physician of a neurosurgical ICU, we developed a rapid-response intervention to address these issues. Building on previous experiences and guided by PSC theory, we developed a complex intervention centred on the prolonged integration of a psychologist into the ICU team, complemented by system-level measures.

Over the following weeks, this model was adapted to the new ICU’s needs through ongoing feedback. To evaluate the intervention’s effects, a pre-post quasi-experimental design was applied. Our research question was: (1) Will a 7-month integrated psychosocial intervention on an ICU improve the team’s wellbeing, measured by PSC, workplace-related stress (irritation) and turnover intention, compared to a control group?

Methods

Study Design

We used a longitudinal quasi-experimental design, as the control and intervention group could not be assigned randomly due to prior agreements. Consequently, potential baseline differences between groups cannot be ruled out. For drafting and editing this manuscript, the SQUIRE reporting checklist³² was used, see Supplemental Appendix 1.

The intervention unit was a neurosurgical ICU of a district hospital in the south of Germany (NC: 12 beds). It consisted of 2 subgroups: staff of the NC unit (Subgroup A) and anaesthesiology department (Subgroup B). Subgroup B works closely with Subgroup A and experiences very similar conditions, thus they were analysed together. The control unit was a neighbouring neurological ICU (NL;12 beds + 2 emergency care beds), selected for similar structure and size. The control group was granted the opportunity to receive the intervention after the end of the study. They were aware that the intervention group received an intervention; blinding was not possible. Occupational health services and human resources were available as usual for both groups. Employment on the intervention or control ward was the only inclusion criterion administered.

The intervention was implemented from June 2022 until January 2023. Initially planned for 5 months, it was extended to 7 months upon request. For an overview of the study timeline and components, see Figure 1.

Figure 1.

Study design and timeline.

We enrolled a survey to evaluate intervention effects at 2 measurement points. Active participation in the intervention (eg, talks with the study therapist) was not required for survey participation, and vice versa.

Intervention

The intervention was installed quickly due to the acute circumstances and refined iteratively.

The development team comprised SN (study psychologist (SP) and advanced trainee psychotherapist), HG (senior physician and head of the group) and KH (senior psychotherapist and supervisor).

The intervention aimed to evaluate, clarify and address the psychosocial needs of HCPs. Five intervention components were used to achieve this goal (see Figure 2): The central element was the presence of a SP in the unit, working as a psychological professional and contact person for HCPs, accompanied by clinical supervision groups, an initial exploratory survey, and team meetings (health circles; German “Gesundheitszirkel”) with all team members.

Figure 2.

Components of the intervention.

The intervention started with the introduction of the SP to the team, followed by weekly visits over the course of 7 months.

Presence of a Psychologist

Presence of the SP was administered on the intervention ward between 08/06/22 and 11/01/23. During this time, the SP visited the ICU once a week for 3 to 5 h, occasionally more often upon request, and was available by phone. For the means of hygiene and as a symbol of integration, she wore the same scrubs as the HCPs. The psychologist’s role was defined as providing psychosocial care on the ward, emphasing organisational and systemic team support. Any issues that fell outside this remit, such as evident mental health conditions, were referred to other professionals. Regular supervision of the SP with a senior clinical psychologist (KH) ensured governance. The SP had 2 main roles: (a) participatory observation and (b) in-situ coaching. A: Becoming a team member, the SP gained insight into unit dynamics and identified team issues through immersion, facilitating understanding of the atmosphere and its challenges. These observations were later shared within the health circle in a processed form, adapting Bion’s^33,34 method of containing. Containing is a therapist’s ability to receive, process and return the client’s overwhelming emotions in a more manageable form, for example by summarising and non-judgmentally reframing staff concerns to support shared understanding.

B: The second role involved informal activities that we referred to as “in-situ coaching” or, in more colloquial terms, “clinical supervision over a cup of coffee,” highlighting the low-threshold, trust-based nature of these supportive and reflective conversations with HCPs. These encounters can be considered as “therapeutic priming”³⁵ while introducing a therapeutic attitude and evidence-based techniques such as resource activation, problem actualisation, and motivational clarification.³⁶ Unlike traditional psychotherapy, this approach was proactive, by reaching out to staff the SP perceived as stressed. Confidentiality was a key factor in combining both roles in 1 person. HCPs were informed that participation in all intervention components was voluntary and that they could decline pro-active offers or withdraw from them at any time without consequences. The SP was bound by professional confidentiality, so all disclosed information was only ever shared in an aggregated and anonymous way, unless otherwise agreed on. The SP’s close integration into the team enabled the mirroring of perceptions and emotions with authenticity and credibility that facilitated acceptance and created a feeling of cohesion. We also assumed that, as Colville,³⁷ (p. 107) put it, “even where the psychologist is not providing formal staff input, their very presence in a team serves as a reminder that there is a psychological dimension to care.”

One-On-One Interactions

Sometimes, interactions between single team members and the SP exceeded the form of brief catch-ups on the ward as they required more time or space. They could then be transferred to a quiet place on the ward or be scheduled to a longer session outside the ward.

Clinical Supervision

N = 3 scheduled clinical group supervision sessions were administered separately for the team (26/07/22), the ward management team (18/10/22) and for the senior physicians’ team (01/12/2022), as requested. The sessions were led by KH, assisted by SN and consisted of 3 to 15 participants. They lasted 90 min. Participation was voluntary.

Feedback Survey

The two-point evaluation survey served 2 purposes: (1) Examining the unit’s needs to objectively substantiate and communicate them. Further, to shape the intervention accordingly by reaching a consensus on the changes the unit wants to work on. And (2) Measuring cues for effectiveness of the intervention (see section 2.4.).

Health Circle (“Gesundheitszirkel”)

A health circle is a group consisting of members from different professions and hierarchies on the ward, brought together with the goal of implementing “workplace-related organisational changes [..] with the participation of employees”³⁸ (p. 259). The health circle met 4 times in total, every 2 to 3 months for 90 to120 min. It consisted of 4 to 20 members from the unit as well as the psychosomatic team. During these meetings, information collected to date was presented and discussed. The psychosomatic team’s role was to support, moderate, and help the ICU team develop actionable measures. The circles also promoted communication among all members of the unit. Decisions resulting from these discussions were then communicated uniformly. The first health circle took place on 01/06/2022 as a “kick-off” meeting for brainstorming issues. The following health circles took place on 13/07/2022 to discuss feedback survey results. The subsequent circle took place on 20/09/2022. In October, a new senior physician working model was implemented, based on decisions made by the health circle. The final health circle took place on 15/12/2022. The meeting protocols were sent to all participants by email afterwards.

Measures

We used the online survey tool Unipark³⁹ to conduct a 40-item questionnaire (see Supplemental Appendix 2) among the members of the intervention and control unit. It included 7 demographic items and 5 questionnaires totalling 30 items, as well as 3 open questions. These invited participants to comment on areas of concern, ideas for improvement, and anything else important to them. Two questionnaires – the Workability Index⁴⁰ and Healthy Work⁴¹ – were applied solely for situational analysis and to provide feedback during the health circles, they are therefore not included in the analyses in this paper.

Approval was obtained from the ethics committee of the University of Ulm (No. 70/22-FSt/Sta) and the staff council. Informed consent for survey participation was obtained from all study participants, digitally via the survey tool. All data was handled as per the EU General Data Protection Regulation (GDPR) and the German Federal Data Protection Act (BDSG), with the goal of best possible data protection and preservation of anonymity.

The T0 and T1 surveys were distributed via E-Mail and QR-code on the ward. The same survey was administered at T0 and T1. At T1, 1 filter item was added to identify prior T0 participation, after which the demographics section was skipped. Missing data were not imputed. The baseline measures for T0 were collected prior to the intervention over a period of 4 months, from February-June 2022. T1 measure took place post-intervention from end of January-May 2023.

Psychosocial Safety Climate (PSC)

PSC was measured using the German 4-item scale.⁴² This reliable, validated scale⁴³ correlates with psychological stress, emotional exhaustion, and engagement. Items are for example, “Psychological well-being of staff is a priority for this organisation.” They are rated on a five-point Likert scale, (1 = strongly disagree, 5 = strongly agree). For calculation and subsequent classification, a sum variable was formed (range 4;20). Based on a Swedish benchmark study,⁴⁴ supported by cross-cultural validation in Germany^43,45, three categories can be differentiated: A PSC score of 8 or lower is an indicator of urgent need for action (red risk), a score above 8 and up to 12 indicates that more consideration should be given to regulations (yellow risk), and a score above 12 indicates good existing health and safety practices (green risk).

Irritation Scale

Irritation is a construct to describe a “state between mental fatigue and mental disorder,”^{46 p. 6}. It can be measured using the irritation scale,⁴⁷ which is a reliable self-report scale “to economically indicate the potentially damaging effect of critical working conditions.” There are positive correlations with anxiety and emotional exhaustion, as well as physiological stress indicators like systolic blood pressure and cortisol levels.⁴⁷ There are negative correlations with life and job satisfaction.⁴⁶

The scale consists of 8 items on a seven-point Likert scale (1 = totally disagree − 7 = totally agree); higher scores indicate higher levels of irritation. It contains 2 subscales: the “cognitive irritation” subscale provides an indication of selective, work-related stress (eg, “I find it difficult to unwind after work”); the “emotional irritation” subscale indicates a generalised stress that goes beyond the work context (eg, “I get irritated easily, even if I don’t show it”). We will report the total and subscale values in percentile ranks, according to recent norms from the German population.⁴⁸ For statistical modelling in the LMMs, the interval-scaled raw scores will be used to better meet model assumptions.

Job and Career Turnover Intention

Turnover intention (intention to change or leave the current employment) was measured as an indicator for actual job and career turnover.⁴⁹ It correlates positively with job demands and negatively with rewards, control, and social support.⁵⁰ It was measured using 3 non-validated items from the Nurse’s Early Exit Study.⁵¹ Each item (eg, “How often in the last year have you thought about leaving your job?”) could be answered on a five-point Likert scale, ranging from “never/ a few times a year/ a few times a month/ a few times a week/ everyday.” Higher scores indicate higher Turnover Intention. We used a sum score for analysis (range 5;15).

Statistical Analysis

Due to the mixed comparisons (within- and between-subject) of participants who completed either T0, T1 or both surveys (see Table 1), we will evaluate the effects of group allocation (intervention vs control) and point of measurement (T0 vs T1) as well as their interaction on the dependent variables (PSC, Irritation, Emotional Irritation, Cognitive Irritation and Intention to leave) using linear mixed effects modelling.⁵² For the analysis, R⁵³ and the package lme4⁵⁴ will be used. Individual participants will be treated as random effects, group allocation, time of data collection and the interaction of those factors will be included as fixed effects. LMM were estimated using Restricted Maximum Likelihood (REML), which allows inclusion of all available cases, including participants with data from only one measurement point. Missing data were assumed to be missing at random, and no imputation procedures were applied. The results will be interpreted using lmerTest which provides P-value estimates via the Satterthwaite approximation (comparable to a Type III ANOVA). Resulting P-values will be interpreted using a Bonferroni-corrected level of significance of α*<.01 to control for Type I error inflation due to multiple testing across the 5 dependent variables.

Table 1.

Size and Demographic Data of Analysed Sample.

Variable	Total sample	CG	IG	Subgroup A	Subgroup B
Total
Sample size	98	33	65	55	13
T0	68	21	47	37	10
T1	52	17	35	29	6
Gender
Female	61	24	37	30	7
Male	26	7	19	16	3
Profession
Nurse	60	30	30	33	0
Physician	31	1	30	17	13
Other (e.g. auxiliary services)	7	2	5	5	0
Age in years
M (SD)	37.28 (11.55)	32.61 (10.17)	39.78 (11.55)	38.08 (11.73)	47.90 (6.17)
Range	20-67	20-59	20-67	20-67	40-62
Time in clinic
M (SD)	10.49 (9.50)	10.33 (10.97)	10.58 (8.71)	10.76 (9.28)	9.80 (5.71)
Range	1-47	1-47	1-34	1-34	3-22
Time in job
M (SD)	14.12 (10.49)	11.54 (9.71)	15.56 (10.72)	14.13 (10.82)	22.00 (7.75)
Range	1-45	3-34	1-45	1-45	14-42
Working hours
M (SD)	35.28 (10.73)	30.82 (11.16)	37.54 (9.83)	36.76 (9.84)	41.80 (9.10)
Range	4-60	4-42	6-60	6-60	26-56
Number of sick leave days
M (SD)	11.26 (13.28)	9.45 (7.71)	12.13 (15.24)	12.47 (14.58)	10.22 (19.49)
Range	1-95	2-30	1-95	1-95	2-62

Note. For participants completing both surveys, only demographic data at T0 was examined.

Results

Participants

N = 120 voluntary survey responses from N = 98 participants across all groups were collected (see Table 1). Based on the pseudonym codes assigned at T0, only N = 22 participants could be reliably matched across both survey time points. However, N = 32 participants self-reported that they had participated in T0 and T1. The reason for this discrepancy remains unclear. We analysed the data of all participants, but only treated confirmed matches as repeated participations. N = 76 participants responded solely to the T0 or T1 survey. N = 3 participants relocated between the subgroups of the intervention group during the intervention.

The ratio of female to male participants was higher in the control group than in the intervention group (3.43:1 vs 1.95:1). The ratio of gender-distribution also varied between times of data collection (T0: 2.63:1 vs T1: 1.58:1). The ratio of nurses to physicians was far greater in the intervention group than the control group (1:1 vs 30:1).

Survey Results

Due to the small sample size, reliability was calculated across both time points. All reliability indices showed good internal consistency, except for the non-validated Intention to Leave items (see Table 2).

Table 2.

Reliability Indices for T0 and T1 Combined.

Scale	Cronbach’s alpha (α)
PSC-4	.87
Irritation scale	.86
Emotional irritation	.82
Cognitive irritation	.85
Intention to leave	.68

Table 3 shows the means and standard deviations of the dependent variables examined in this study, differentiated by group allocation and point of measurement, as well as the estimated interaction resulting from linear mixed-effects modelling. We analysed the full sample of N = 98 participants from both groups using LMM to account for both within- and between-subject effects. Contrary to our hypothesis, none of the interactions were significant on the basis of the Bonferroni-corrected significance level of α*<.01.

Table 3.

Mean Sum Scores (SS), Percentile Ranks (PR), Standard Deviations (SD, in Brackets) and Sample Size (N) per Subgroup of the Dependent Variables by Group Allocation and Point of Measurement along with the Resulting Test Statistic, P-Values and 99% Confidence Intervals of the Interaction Effects.

	T0 (Baseline)		T1 (Follow-up)		LMM of interaction of group allocation and point of measurement	99% CI of interaction effect
Measure	Control	Intervention	Control	Intervention		99% CI of interaction effect
PSC-4 (SS)	9.52 (3.53), N = 21	7.36 (3.16), N = 47	8.71 (4.47), N = 17	9.83 (3.88), N = 35	F(1, 66.13) = 5.68, P = .02	(−0.75, 5.35)
Irritation scale (PR)	59.36 (20.51), N = 21	74.66 (17.60), N = 47	58.05 (22.60), N = 15	70.84 (18.95), N = 34	F(1, 36.24) = 0.3, P = .59	(−6.12, 4.02)
Emotional irritation (PR)	57.38 (20.13), N = 21	70.41 (19.09), N = 47	57.45 (23.67), N = 15	67.32 (19.69), N = 34	F(1, 51.71) = 0.27, P = .60	(−4.73, 3.10)
Cognitive irritation (PR)	64.52 (21.95), N = 21	78.71 (16.64), N = 47	61.48 (22.16), N = 15	75.39 (19.55), N = 34	F(1, 31.31) = 0.47, P = .50	(−2.76, 1.75)
Intention to leave (SS)	6.38 (1.77), N = 21	6.57 (2.81), N = 47	5.35 (1.58), N = 17	5.43 (2.28), N = 35	F(1, 28.61) = 0.41, P = .53	(−1.19, 2.71)

Regarding Irritation Scale, there were no significant main effects of group allocation or time of data collection. We categorised the irritation scores using Stanine⁵⁵ and percentile ranks⁴⁸ into the 3 categories unobtrusive (Stanine <7, PR <77), threshold (Stanine = 7, PR = 77-89) and noticeable (Stanine ≥8, PR >89). All mean percentile ranks fall within the normal range, with the exception of the subscale cognitive irritation (work-related stress) of the intervention group, which is in the threshold category at T0, but in the normal category at T1. Irritation was overall lower within the control group.

For the non-validated Intention to Leave scale, the LMM suggested a trend towards a main effect of time with overall lower scores at T1 (F(1, 74.14) = 6.01, P = .02). This effect, however, did not reach the Bonferroni-adjusted significance level (α* <.01).

The interaction of group allocation and time of data collection on PSC did not reach the significance level (P = .02) but indicated a trend towards improved PSC scores in the intervention group over time in accordance with our hypothesis. Figure 3 presents the distribution of PSC sum scores across the risk-categories proposed by Berthelsen et al.⁴⁴ Descriptively, responses in the intervention group shifted towards the low risk category, while the control group showed an opposite trend with more high risk responses over time.

Figure 3.

Distribution of PSC sum scores among the low, medium and risk categories by group allocation and point of measurement.

Open Question Answers and Observations Throughout the Study Course

Open responses and observations were summarised thematically to highlight common patterns and support interpretation of the quantitative data. All intervention components (presence of a psychologist, one-on-one interactions, clinical supervisions, health circle, survey) came to use.

In line with quantitative results, the free-text field answers indicated an initially high need for support at T0. For example, 1 respondent wrote; “I have always been able to switch off [of work]. The current situation no longer allows this and is placing an enormous strain on my everyday life.”

Based on the T0 survey and health circle discussions, the ward identified and prioritised its most pressing work issues. The 5 most important issues were: Cooperation with other departments, appreciation of employees’ performance, working without interruptions and disruptions, cooperation between professional groups and having to make difficult decisions. Within the health circles, they were summarised into 3 categories: appreciation, cooperation and workflow. Regarding workflow, a working group of senior physicians as well as a working group comprising nurses and assistant physicians was set up to tackle it. During the intervention, the senior physician structure changed, improving daily availability and clarifying responsibilities – a change welcomed by all HCPs on the ward. Although this was planned earlier, the process was refined and encouraged by the health circle and implemented during the intervention period by the newly formed working group. Appreciation and cooperation were issues that were discussed a lot and brought to the attention of management (physician and nurse).

Concerns about nursing ward management emerged within open surveys responses and one-on-one talks, (eg, experienced lack of cooperation and care towards the nursing staff). With consent, the SP shared these with management, who were motivated to make changes. Despite the concerns, HCPs expressed hope for improved interdisciplinary cooperation. They showed appreciation for the survey, as it made them feel heard and valued.

In supervision and one-to-one meetings, HCPs showed satisfaction with the effects of the intervention, as they perceived an improved team atmosphere and management interest. One team member told the SP that, since the new work model for senior physicians was introduced, time pressure had been “eliminated” and working on the ward had become much more enjoyable. They appreciated being able to focus on important tasks. Another team member said that the nursing staff felt more valued because they were more involved in treatment, their opinions were sought, and they had better access to information. This is also reflected in the positive trends within cognitive irritation and PSC.

To our knowledge, both wards did not use any regular psychosocial services during the time of the intervention. When the control ward was later on offered to receive the same intervention, they declined.

Discussion

In this study, we implemented and evaluated a 7-month ad-hoc integrated psychosocial intervention in an ICU and compared it with a control unit, measuring both groups pre- and post-intervention. All planned components were used and, after an acclimatisation period, accepted and appreciated by the unit. While the quasi-experimental design (with no randomisation) and the small, imbalanced sample size entail serious methodological limitations, discussed later, the findings provide important first insight into the feasibility and acceptability of such an approach. The health circles revealed issues related to appreciation, cooperation and workflow, leading to team-level changes during the intervention. The observations throughout the study period indicated that the intervention fostered a more open and supportive work environment.

Statistical analysis indicated that our hypothesis was rejected, as no significant group-and-time interaction effects were found.

Descriptively, cognitive irritation improved slightly within the intervention group, in line with our hypothesis. We also observed a baseline difference suggesting that the intervention group was more irritated initially – consistent with their expressed need for support. Nevertheless, this imbalance restricts validity of statistical analysis.

A non-significant trend towards reduced Intention to leave was observed in both wards. As the scale was non-validated with moderate internal consistency, reliability of this outcome is limited. The results are not in line with the changes within the other measures, yet it may be attributable to a general post-pandemic easing of stress amongst HCPs.⁵⁶

Although not statistically significant, the observed trend towards improved PSC in the intervention group may nevertheless indicate a clinically relevant change, especially as PSC slightly worsened in the control group. These outcomes must be interpreted very cautiously due to the non-randomised design, baseline differences, unequal group sizes and sample imbalances regarding gender and occupation. While some literature indicates effects of gender or occupation on PSC,⁴⁴ no robust effects are known.⁵⁷ As covariate analysis wasn’t included in our hypotheses, we refrained from conducting it a posteriori.

In the literature, PSC correlates negatively with psychological distress⁵⁸ and shows similar patterns to work-related strain.⁵⁹ However, findings are inconsistent, as some studies reported no link between PSC improvement and well-being or work-related outcomes.⁶⁰ Overall, interventions designed to enhance PSC generally appear to succeed in this goal.^59,60 Against this background, the direction of the observed change in PSC in the present study may reflect a meaningful intervention signal. However, this must be interpreted in light of the methodological limitations.

From a theoretical perspective, the observed PSC improvement may relate to psychosocial care being implemented at management’s request, potentially perceived as a sign of management care. Additionally, the intervention may have improved social connections, leadership, trust, workplace values and healthy coping strategies,⁶¹ all of which are embedded within PSC. This interpretation is supported by HCPs reports of workplace improvement. Given the non-blinded nature of the study, a confounding Hawthorne effect cannot be ruled out. The opposite development in the control group could reflect contrast effects, as they may have felt disadvantaged. Alternatively, working conditions may have worsened organisation-wide, but were mitigated by the intervention.

Limitations

As a quasi-experimental field study, this research was subject to significant methodological constraints. Participants were neither randomised nor blinded, and the sample was small and imbalanced. Expectations, hope or floor effects may have influenced the outcomes. Owing to the limited sample size, longitudinal matching (only N = 22 matches) or advanced bias-reduction approaches (eg, propensity score matching) were not feasible, leaving possible residual confounding.

Responses at T1 may have been influenced by the intervention group being informed of the T0 results at the second health circle. Sample size and group imbalances (by gender and profession) did not cause convergence issues within the formulated LMMs,⁶² yet they still influence statistical power and increase the risk of type-II errors.⁶³ In particular, these differences may have influenced the baseline levels, responsiveness to the intervention, and outcomes. Some mismatch of participant codes between T0 and T1, likely due to spelling mistakes, complicated longitudinal analysis and indicate caution. While unplanned, the modest extension of the study duration by 2 months is unlikely to have affected its statistical power or internal and external validity, given the quasi-experimental design and the fact that no formal power calculation was conducted a priori.

The intervention supported a major structural change increasing the availability of senior physicians. As this change overlaps with key intervention components, particularly management engagement and improvements at the ward level, intervention-specific effects cannot be clearly disentangled from it. Integrating a psychological professional into the team posed some challenges. The unfamiliarity came with barriers of incomprehension and occasional rejection. It required sensitivity to group dynamics and awareness of individual and collective motives as well as hierarchies. Uncertainty in acute medical teams about the role of the SP has been well documented,⁶⁴ highlighting the need for clear role definitions.

It remains unclear whether a dose-response effect exists. With an intervention lasting only 3 to 5 h per week, the level of intensity was limited, but the approach was cost-effective.

Conclusions

Although beneficial in practice, the statistical results of our study were weak. As this was a quasi-experimental study design, these effects were to be expected. Future studies require larger samples and a more controlled experimental design which allows for example, matched longitudinal data. Further methodological refinements like stronger intervention definitions, blinding or variation of intervention dose could increase statistical power, yet they limit complexity and freedom of this adaptive approach. To advance psychosocial care in ICUs, the field needs a precise evaluation of the needs and resources of ICU staff and other affected groups. There seem to be shared needs between the different affected groups on ICUs,⁶⁵ which emphasises the need for a broader and holistic cross-group perspective. Ideally, health intervention development should follow stages from a model or theory of the problem to a model of change, programme design, production, and finally implementation and evaluation.⁶⁶ Our work continues in the ongoing “IPS-Pilot” project; refining the intervention and evaluating its feasibility in a pilot RCT in 8 German ICUs.⁶⁷

Supplemental Material

sj-pdf-1-inq-10.1177_00469580261428687 – Supplemental material for Impact of an Integrated Psychosocial Intervention for Health Care Professionals in an Intensive Care Unit: Controlled Quasi-Experimental Trial

Supplemental material, sj-pdf-1-inq-10.1177_00469580261428687 for Impact of an Integrated Psychosocial Intervention for Health Care Professionals in an Intensive Care Unit: Controlled Quasi-Experimental Trial by Sophie Felicitas Nickel, Simone Korger, Karl Christian Hirning, Klaus Hönig, Christian Rainer Wirtz, Ferdinand Keller and Harald Otto Gündel in INQUIRY: The Journal of Health Care Organization, Provision, and Financing

Supplemental Material

sj-pdf-2-inq-10.1177_00469580261428687 – Supplemental material for Impact of an Integrated Psychosocial Intervention for Health Care Professionals in an Intensive Care Unit: Controlled Quasi-Experimental Trial

Supplemental material, sj-pdf-2-inq-10.1177_00469580261428687 for Impact of an Integrated Psychosocial Intervention for Health Care Professionals in an Intensive Care Unit: Controlled Quasi-Experimental Trial by Sophie Felicitas Nickel, Simone Korger, Karl Christian Hirning, Klaus Hönig, Christian Rainer Wirtz, Ferdinand Keller and Harald Otto Gündel in INQUIRY: The Journal of Health Care Organization, Provision, and Financing

Footnotes

Acknowledgements

I acknowledge the use of DeepL Translate (DeepL.com) and ChatGPT-5 (chatgpt.com) as writing assistance to generate better translations from my first language in the preparation of this manuscript.

ORCID iDs

Sophie Felicitas Nickel

Simone Korger

Karl Christian Hirning

Klaus Hönig

Christian Rainer Wirtz

Ferdinand Keller

Harald Otto Gündel

Ethical Considerations

Ethics committee approval of the university of Ulm was obtained on 16/05/2022 (No. 70/22-FSt/Sta), approval of the staff council on 31/03/2022.

Consent to Participate

Informed consent for survey participation was obtained from all participants of the study. The consent form was integrated digitally into the survey tool.

Author Contributions

Sophie Nickel: Conceptualisation, Methodology, Investigation, Writing Original Draft. Simone Korger: Formal analysis, Data Curation, Writing Review & Editing, Visualisation. Christian Hirning: Methodology, Data Curation. Klaus Hönig: Conceptualisation, Methodology, Supervision. Christian Rainer Wirtz: Conceptualisation, Resources, Funding acquisition. Ferdinand Keller: Formal analysis, Writing – Review & Editing. Harald Gündel: Conceptualisation, Methodology, Resources, Writing – Review & Editing, Project administration.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The intervention was funded by the Clinic for Neurosurgery BKH Günzburg, which operates the intensive care unit that received the intervention. No funding was received for conducting this study and preparing the manuscript.

Declaration of Conflicting Interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: The study allowed for a result in both directions. The study therapist is also the author of this paper. However, the analysis was predominantly carried out by SK and FK, who were not involved in the intervention.

Data Availability Statement

Data is available on request.

Supplemental Material

Supplemental material for this article is available online.

References

Guoilian

Zeng

Xiaoying

Xiaoqian

Xiaohua

Global prevalence of turnover intention among intensive care nurses: a meta-analysis. Nurs Crit Care. 2023;28:159-166. https://doi.org/10.1111/nicc.12679

Kumar

Sinha

Varma

Prabhakaran

Phatak

Nimbalkar

SM.

Burnout and its correlates among nursing staff of intensive care units at a tertiary care center. J Fam Med Prim Care. 2021;10(1):443-448. https://doi.org/10.4103/jfmpc.jfmpc_1651_20

Ganahl

Knaus

Wiesenhuetter

Klemm

Jabinger

Strametz

Second victims in intensive care—emotional stress and traumatization of intensive care nurses in western Austria after adverse events during the treatment of patients. Int J Environ Res Public Health. 2022;19(6):3611. https://doi.org/10.3390/ijerph19063611

Arndt

Beerlage

DIVI-Jahrbuch 2019/2020. MWV Medizinisch Wissenschaftliche Verlagsgesellschaft MBH & Co KG. 2019.

Donchin

Seagull

FJ.

The hostile environment of the intensive care unit. Curr Opin Crit Care. 2002;8(4):316-320. https://doi.org/10.1097/00075198-200208000-00008

Dodek

Cheung

Burns

KEA

, et al. Moral distress and other wellness measures in Canadian critical care physicians. Ann Am Thorac Soc. 2021;18(8):1343-1351. https://doi.org/10.1513/AnnalsATS.202009-1118OC

Dewa

Loong

Bonato

Thanh

Jacobs

How does burnout affect physician productivity? A systematic literature review. BMC Health Serv Res. 2014;14(1):325. https://doi.org/10.1186/1472-6963-14-325

Dewa

Jacobs

Thanh

Loong

An estimate of the cost of burnout on early retirement and reduction in clinical hours of practicing physicians in Canada. BMC Health Serv Res. 2014;14(1):254. https://doi.org/10.1186/1472-6963-14-254

Halbesleben

Wakefield

Cooper

LB.

Nurse burnout and patient safety outcomes: nurse safety perception versus reporting behavior. West J Nurs Res. 2008;30(5):560-577. https://doi.org/10.1177/0193945907311322

10.

Tawfik

Scheid

Profit

, et al. Evidence relating health care provider burnout and quality of care: a systematic review and meta-analysis. Ann Intern Med. 2019;171(8):555-567. https://doi.org/10.7326/M19-1152

11.

Teng

Shyu

Chiou

Fan

Lam

SM.

Interactive effects of nurse-experienced time pressure and burnout on patient safety: a cross-sectional survey. Int J Nurs Stud. 2010;47(11):1442-1450. https://doi.org/10.1016/j.ijnurstu.2010.04.005

12.

Niecke

Hartog

Deffner

Janssens

Michels

Need for psychological support in intensive care. Med Klin. 2020;115(2):135-139. https://doi.org/10.1007/s00063-018-0523-x

13.

Deffner

Schwarzkopf

Waydhas

Sektion „Qualität und Ökonomie“ der Deutschen Interdisziplinären Vereinigung für Intensiv- und Notfallmedizin (Section “Quality and Economics” of the German Interdisciplinary Association for Intensive Care and Emergency Medicine. Psychological care in German intensive care units). Med Klin. 2021;116(2):146-153. https://doi.org/10.1007/s00063-019-00638-2

14.

Rothwell

Kehoe

Farook

Illing

Enablers and barriers to effective clinical supervision in the workplace: a rapid evidence review. BMJ Open. 2021;11(9):e052929. https://doi.org/10.1136/bmjopen-2021-052929

15.

Beadman

Carraretto

Key elements of an evidence-based clinical psychology service within adult critical care. J Intensive Care Soc. 2023;24(2):215-221. https://doi.org/10.1177/17511437211047178

16.

Rink

Oyesanya

Adair

Humphreys

Silva

Sexton

JB.

Stressors among healthcare workers: a summative content analysis. Glob Qual Nurs Res. 2023;10:23333936231161127. https://doi.org/10.1177/23333936231161127

17.

Prompahakul

Keim-Malpass

LeBaron

G. Epstein

Thai translation and psychometric testing of the measure of moral distress for healthcare professionals. J Heal Sci -med Res. 2020;41(6). https://doi.org/10.31584/jhsmr.2023976

18.

Waydhas

Riessen

Markewitz

, et al. Empfehlung zur Struktur und Ausstattung von Intensivstationen 2022 (Erwachsene). Anästh Intensivmed. 2023;64:129–137.

19.

Richter-Kuhlmann

Arztgesundheit: Künftig nicht nur eine Floskel. Dtsch Arztebl Int. 2019;116(23-24):11-56.

20.

Azoulay

Timsit

Sprung

, et al. Prevalence and factors of intensive care unit conflicts: the conflicus study. Am J Respir Crit Care Med. 2009;180(9):853-860. https://doi.org/10.1164/rccm.200810-1614OC

21.

Dollard

Dormann

Tuckey

Escartín

Psychosocial safety climate (PSC) and enacted PSC for workplace bullying and psychological health problem reduction. Eur J Work Organ Psychol. 2017;26(6):844-857. https://doi.org/10.1080/1359432X.2017.1380626

22.

Liu

Zhao

Zhan

Can the psychosocial safety climate reduce ill-health presenteeism? Evidence from Chinese healthcare staff under a dual information processing path lens. Int J Environ Res Public Health. 2020;17(8): 2969. https://doi.org/10.3390/ijerph17082969

23.

Melnyk

Tan

Hsieh

, et al. Critical care nurses’ physical and mental health, worksite wellness support, and medical errors. Am J Crit Care. 2021;30(3):176-184. https://doi.org/10.4037/ajcc2021301

24.

Laschinger

HKS

Fida

A Time-lagged analysis of the effect of authentic leadership on workplace bullying, burnout and occupational turnover intentions. Eur J Work Organ Psychol. 2013;23:739-753. https://doi.org/10.1080/1359432X.2013.804646

25.

Zadow

Dollard

Parker

Storey

Psychosocial safety climate: a review of the evidence. In: Dollard

Dormann

Awang

, eds. Psychosocial Safety Climate: A New Work Stress Theory. Springer International Publishing; 2019;31-75.

26.

Yuan

Ren

Yin

, et al. How does psychosocial safety climate cross-level influence work engagement and job burnout: the roles of organization-based self-esteem and psychological detachment. BMC Nurs. 2024;23(1):389. https://doi.org/10.1186/s12912-024-01935-8

27.

Hemingway

Smith

CS.

Organizational climate and occupational stressors as predictors of withdrawal behaviours and injuries in nurses. J Occup Organ Psych. 1999;72(3):285-299. https://doi.org/10.1348/096317999166680

28.

Tripney Berglund

Kombeiz

Dollard

. Manager-driven intervention for improved psychosocial safety climate and psychosocial work environment. Saf Sci. 2024;176:106552. https://doi.org/10.1016/j.ssci.2024.106552

29.

PSNV. Gründung des Psychosozialen Versorgungsnetzwerkes der Charité (Foundation of the Charité psychosocial care network). n.d.. Accessed March 28, 2025. https://psychosoziales-versorgungsnetzwerk.charite.de/historie

30.

Mayer

. Mittendrin in der Pandemie-so geht es Intensivpflegekräften und Ärzten der Uniklinik (In the midst of the pandemic—how intensive care nurses and doctors at the university hospital are getting on). Südwest Presse; 2021.

31.

Rose

Schürmann

Ferentzi

Schmitt

KRL

. Integrated psychosocial support in the ICU. Curr Opin Anaesthesiol. 2025;38(1):37-42. https://doi.org/10.1097/ACO.0000000000001455

32.

Ogrinc

Davies

Goodman

Batalden

Davidoff

Stevens

SQUIRE 2.0 (standards for quality improvement reporting excellence): revised publication guidelines from a detailed consensus process. BMJ Qual Saf. 2016;25(12):986-992. https://doi.org/10.1136/bmjqs-2015-004411

33.

Bion

WR.

Learning From Experience. Heinemann; 1962.

34.

Wolfram

. Containing. In: Stumm

Pritz

, eds. Wörterbuch der Psychotherapie. Springer; 2000;114.

35.

Landbeck

Witt

Marty Petit

, et al. What clinical practices for intensive care psychologists in France? A national survey. Crit Care. 2024;28(1):204. https://doi.org/10.1186/s13054-024-04987-z

36.

Grawe

Psychologische Therapie. Hogrefe; 1998.

37.

Colville

The role of a psychologist on the paediatric intensive care unit. Child Psychol Psychiatry Rev. 2001;6(3):102-109. https://doi.org/10.1017/s1360641701002611

38.

Aust

Ducki

Comprehensive health promotion interventions at the workplace: experiences with health circles in Germany. J Occup Health Psychol. 2004;9(3):258-270. https://doi.org/10.1037/1076-8998.9.3.258

39.

Unipark. 2021. https://www.unipark.com/

40.

Tuomi

Ilmarinen

Jahkola

Katajarinne

Tulkki

, eds. Work Ability Index. 2nd ed. Finnish Institute of Occupational Health; 1998.

41.

Mulfinger

Sander

Stuber

, et al. Cluster-randomised trial evaluating a complex intervention to improve mental health and well-being of employees working in hospital - a protocol for the SEEGEN trial. BMC Public Health. 2019;19(1):1694. https://doi.org/10.1186/s12889-019-7909-4

42.

Dollard

MF.

The PSC-4: a short PSC tool. In: Dollard

Dormann

Awang

, eds. Psychosocial Safety Climate: A New Work Stress Theory. Springer International Publishing; 2019;385-409.

43.

Formazin

Ertel

Kersten

Nübling

Psychosocial safety climate [Validierung der deutschen Fragebogenversion]. Zeitschrift für Arbeits- und Organisationspsychologie A&O. 2022;66:129-142. https://doi.org/10.1026/0932-4089/a000379

44.

Berthelsen

Muhonen

Bergström

Westerlund

Dollard

MF.

Benchmarks for evidence-based risk assessment with the Swedish version of the 4-item psychosocial safety climate scale. Int J Environ Res Public Health. 2020;17:8675. https://doi.org/10.3390/ijerph17228675

45.

Berthelsen

Ertel

Geisler

Muhonen

Validating the psychosocial safety climate questionnaire – integration of findings from cognitive interviews in Germany and Sweden. Scand J Work Organ Psychol. 2019;4:9. https://doi.org/10.16993/sjwop.85

46.

Klingenberg

Süß

Profiling resilience: a latent profile analysis of German nurses’ coping and resilience. Front Health Serv. 2022;2:960100. https://doi.org/10.3389/frhs.2022.960100

47.

Mohr

Müller

Rigotti

Normwerte der Skala Irritation: ZweiDimensionen psychischer Beanspruchung (Norm values of the irritation scale: Two dimensions of psychological strain). Diagnostica. 2005;51:12-20. https://doi.org/10.1026/0012-1924.51.1.12

48.

Gralla

Guendel

Mueller

, et al. Validation of the irritation scale on a representative German sample: new normative data. Sci Rep. 2023;13(1):15374. https://doi.org/10.1038/s41598-023-41829-4

49.

Camerino

Conway

Van der Heijden

BIJM

, et al. Low-perceived work ability, ageing and intention to leave nursing: a comparison among 10 European countries. J Adv Nurs. 2006;56(5):542-552. https://doi.org/10.1111/j.1365-2648.2006.04046.x

50.

Alipour

Shabani

Asady Fakhr

Mahdiyoun

SA.

Study of occupational stress as predictors of nursing intention to leave. J Ergon. 2018;6(1):58-64. https://doi.org/10.30699/jergon.6.1.58

51.

Simon

Tackenberg

Hasselhorn

Kümmerling

Büscher

Müller

Auswertung der ersten Befragung der NEXT-Studie in Deutschland. Universität Wuppertal; 2005.

52.

Field

, ed. Discovering Statistics Using IBM SPSS Statistics. 4th ed. Sage Publications; 2013.

53.

R: a language and environment for statistical computing. R Foundation for Statistical Computing; 2021. https://www.R-project.org/

54.

Bates

Mächler

Bolker

Walker

Fitting linear mixed-effects models using LME4. J Stat Softw. 2015;67(1):1-48. https://doi.org/10.18637/jss.v067.i01

55.

Mohr

Rigotti

Müller

Irritation - ein Instrument zur Erfassung psychischer Befindensbeeinträchtigungen im Arbeitskontext. Skalen- und Itemparameter aus 15 Studien. Zeitschrift für Arbeits- und Organisationspsychologie. 2005;49:44-48.

56.

Niecke

Henning

Hellmich

, et al. Psychische Belastung des intensivmedizinischen Personals in Deutschland im Verlauf der COVID-19-Pandemie. Evidenz aus der VOICE-Studie. Med Klin. 2025;120(4):300-306. https://doi.org/10.1007/s00063-024-01164-6

57.

Fattori

Comotti

Bordini

Dollard

Bonzini

Psychosocial safety climate (PSC) at middle management level in the healthcare sector: a contribution to the Italian validation of psychosocial safety climate-4. Front Psychol. 2022;13:13-2022. https://doi.org/10.3389/fpsyg.2022.1046286

58.

Inoue

Hatakeyama

Kondo

, et al. Post-intensive care syndrome: its pathophysiology, prevention, and future directions. Acute Med Surg. 2019;6(3):233-246. https://doi.org/10.1002/ams2.415

59.

Boß

Ross

Reis

, et al. Effectiveness of an integrated platform-based intervention for promoting psychosocial safety climate and mental health in nursing staff: a pragmatic cluster randomised controlled trial. Int J Nurs Stud. 2025;167:105076. https://doi.org/10.1016/j.ijnurstu.2025.105076

60.

de Wijn

van der Doef

MP.

Reducing psychosocial risk factors and improving employee well-being in emergency departments: a realist evaluation. Front Psychol. 2021;12:728390. https://doi.org/10.3389/fpsyg.2021.728390

61.

Laurent

Fournier

Lheureux

, et al. Mental health and stress among ICU healthcare professionals in France according to intensity of the COVID-19 epidemic. Ann Intensive Care. 2021;11(1):90. https://doi.org/10.1186/s13613-021-00880-y

62.

Meteyard

Davies

RAI

. Best practice guidance for linear mixed-effects models in psychological science. J Mem Lang. 2020;112:104092. https://doi.org/10.1016/j.jml.2020.104092

63.

Knudson

Lindsey

Type I and type II errors in correlations of various sample sizes. Compr Psychol. 2014;3:03.CP.3.1. CP.3.1 https://doi.org/10.2466/03.Cp.3.1

64.

Beadman

Highfield

Wade

Ford

Scott

Wood

The health of intensive care unit teams: moving beyond a trauma focus. Occup Med. 2021;71(8):386-387. https://doi.org/10.1093/occmed/kqab135

65.

Wintermann

Petrowski

Weidner

Strauß

Rosendahl

Impact of post-traumatic stress symptoms on the health-related quality of life in a cohort study with chronically critically ill patients and their partners: age matters. Crit Care. 2019;23(1):39. https://doi.org/10.1186/s13054-019-2321-0

66.

Bartholomew-Eldredge

Markham

Ruiter

Fernandez

Kok

Parcel

, eds. Planning Health Promotion Programs: An Intervention Mapping Approach. 4th ed. Jossey Bass; 2016.

67.

Nickel

Korger

Schindler

, et al. Integrated psychosocial care in intensive care (IPS-Pilot): protocol for the systematic, multimethod development of a complex intervention (phase A). JMIR Res Protoc. 2025;14:e65682. https://doi.org/10.2196/65682

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.13 MB

0.12 MB