Advertisement

Humanization of critical care—psychological effects on healthcare professionals and relatives: a systematic review

  • Imelda M. GalvinEmail author
  • Jordan Leitch
  • Rebecca Gill
  • Katherine Poser
  • Sandra McKeown
Review Article/Brief Review

Abstract

Purpose

To systematically review and evaluate the effects of humanized care of the critically ill on empathy among healthcare professionals, anxiety among relatives, and burnout and compassion fatigue in both groups.

Source

MEDLINE, PsycINFO, EMBASE, CINAHL, Cochrane Central Register of Controlled Trials (CENTRAL), and ProQuest Dissertations were searched from inception to 29 June 2017 for studies that investigated the effects of interventions with potential to humanize care of the critically ill on the following outcomes: empathy among critical care professionals, anxiety among relatives, and burnout and compassion fatigue in either group. We defined a humanizing intervention as one with substantial potential to increase physical or emotional proximity to the patient. Two reviewers independently selected studies, extracted data, and assessed risk of bias and data quality.

Principal findings

Twelve studies addressing four discrete interventions (liberal visitation, diaries, family participation in basic care, and witnessed resuscitation) and one mixed intervention were included. Ten studies measured anxiety among 1,055 relatives. Two studies measured burnout in 288 critical care professionals. None addressed empathy or compassion fatigue. Eleven of the included studies had an overall high risk of bias. No pooled estimates of effect were calculated as a priori criteria for data synthesis were not met.

Conclusions

We found insufficient evidence to make any quantitative assessment of the effect of humanizing interventions on any of these psychologic outcomes. We observed a trend towards reduced anxiety among family members who participated in basic patient care, liberal visitation, and diary keeping. We found conflicting effects of liberal visitation on burnout among healthcare professionals.

Humanisation des soins critiques — effets psychologiques sur les professionnels de la santé et les membres de la famille : une revue systématique

Résumé

Objectif

Étudier et évaluer systématiquement les interventions de soins au potentiel « plus humain » chez les patients en état critique sur l’empathie chez les professionnels de la santé, l’anxiété des proches et l’épuisement et la compassion dans les deux groupes.

Source

La recherche a été menée dans les bases de données MEDLINE, PsycINFO, EMBASE, CINAHL, le Registre des essais cliniques Cochrane Central (CENTRAL), et ProQuest Dissertations depuis leur création jusqu’au 29 juin 2017 pour identifier les études portant sur les effets d’interventions susceptibles d’humaniser les soins de patients en état critique sur les critères d’évaluation suivants : empathie chez les professionnels de soins critiques, anxiété chez les membres de la famille, épuisement et usure de la compassion dans l’un ou l’autre groupe. Nous avons établi qu’une intervention d’humanisation était une intervention susceptible d’augmenter la proximité physique et émotionnelle avec le patient. Deux chercheurs ont sélectionné les études de façon indépendante, en ont extrait les données et évalué les risques de biais ainsi que la qualité des données.

Constatations principales

Douze études abordant quatre interventions particulières (ouverture libérale des visites, journaux personnels, participation de la famille aux soins de base et ressuscitation devant témoin) et une intervention mixte ont été incluses. Dix études ont mesuré l’anxiété chez 1 055 membres de la famille. Deux études ont mesuré l’épuisement chez 288 professionnels de soins critiques. Aucune étude n’a abordé l’empathie ou l’usure de la compassion. Onze des études retenues comportaient un risque global élevé de biais. Aucune estimation groupée de l’effet n’a été calculée, car les critères préétablis pour la synthèse des données n’ont pas été satisfaits.

Conclusions

Nous avons trouvé une insuffisance de données probantes pour quantifier une évaluation des interventions d’humanisation sur l’un ou l’autre des critères d’évaluation psychologiques. Nous avons observé une tendance à la baisse de l’anxiété chez les membres des familles ayant participé aux soins de base aux patients, ayant bénéficié d’horaires de visites libéraux et de la tenue de journaux personnels. Nous avons constaté des effets contradictoires de la libéralisation des visites sur l’épuisement chez les professionnels de la santé.

Dehumanization is the process of depriving someone of positive human qualities. It is usually seen as an overt, active, extreme, conflict-driven phenomenon1 and not typically thought of as being associated with healthcare.

There is however a more subtle form of dehumanization called infrahumanization that can occur outside conflict situations. First described by Leyens,2 infrahumanization differs from blatant dehumanization in that the latter involves the denial of primary emotions (fear, sadness, anger, joy, and disgust) while infrahumanization involves the denial of uniquely human secondary emotions (compassion, admiration, fondness, contempt, and disillusionment).1 Individuals tend to attribute more positive secondary emotions (e.g., compassion) to ‘in groups’ and more negative secondary emotions (e.g., contempt) to ‘out groups’, suggesting that infrahumanization is not simply an expression of dislike but rather a general diminution of the humanity of those perceived to be outside one’s group.2 Infrahumanization can occur whenever groups identify others as being outside their cohort and has been shown to occur in medical settings.3,4 Capozza found that physicians and nurses caring for oncology patients perceived their own professional group as possessing more uniquely human traits than the patient group.3 A study of nursing professionals working in different specialties in an urban hospital showed that denial of uniquely human traits (reasoning, rationality, morality, and intellectual ability) to patients was associated with lower levels of stress among nurses who exhibited high levels of organizational and affective commitment, suggesting that infrahumanization is a subconscious coping strategy to reduce the emotional burden of caring for patients.4

Those suffering from critical illness are particularly vulnerable to this unintentional dehumanization. Several factors contribute:

Critical illness often impairs agency and capability. Loss of agency can be seen as loss of competence and so, although viewed with warmth, those with life-threatening illness may be subconsciously regarded by others as being less human.5,6

Evidence shows that people are more likely to dehumanize those who appear different from themselves.6,7 The physical changes of critical illness, reduced mobility, and a standard hospital gown serve to make patients similar to each other and dissimilar to both their usual selves and their carers. This is further exacerbated by dependency on mechanical devices, breathing tubes, intravenous lines, ventilators, and extracorporeal circuits. Unlike most animals, humans rely on shape and facial recognition rather than smell to identify another human.6 The addition of equipment distorts the basic human shape, may obscure facial features, and impedes our ability to perceive the face as a whole. Piece-by-piece as opposed to whole facial processing is the mode our brain uses to distinguish objects and is known to be a dehumanizing mode of facial perception.8 Functional magnetic resonance imaging has shown that medial prefrontal cortex activation is necessary for social cognition and is activated when seeing other humans but is not activated when seeing objects.6

Critically ill patients are often sedated, reducing their ability to make eye contact. Eye contact is a powerful communication tool that conveys a host of social information. Direct gaze triggers increased activity in the amygdala and activation of approach-related responses, preparing us for social interaction. Averted gaze triggers feelings of disengagement and disinterest.9

Critical illnesses themselves are complex, multifactorial, and difficult to treat. Communication between healthcare professionals requires rapid and accurate appreciation of complicated disease processes. Labelling facilitates this. Mr. Jones, a 52-yr-old farmer with septic shock, diabetes, and renal failure, is often more easily understood as a case of septic shock in a diabetic with acute-on-chronic renal dysfunction.

In an effort to refocus care of the critically ill through a more person-centred lens, several initiatives have been advocated, including diaries, liberal visitation policies, involvement of relatives in basic patient care, and opportunities for relatives to be present during resuscitation.10-12 To know if these initiatives allow care providers to see patients in a more human light, we need to determine what effect they have on expressions of empathy among healthcare professionals, empathy being essential to the appreciation of another’s humanity.1,13

In the light of evidence that dehumanization may be protective against burnout and14 that subtle dehumanization may be motivated by the need to reduce the emotional cost of caring,4,15 and recognizing that care of the care giver is a key component of humanized care,12 we also need to find out what effect humanizing interventions have on the risk of burnout and compassion fatigue among both healthcare professionals and relatives.

Finally, as anxiety is common among relatives of the critically ill and can impede critical decision-making,16 and since many of these initiatives focus on increased relative presence and involvement, we also need to determine the effect of humanized care on anxiety among relatives. The objectives of this review were therefore to systematically review the literature to determine the effects of interventions that humanize care of the critically ill on empathy in healthcare professionals, burnout and compassion fatigue among healthcare professionals and relatives, and anxiety among relatives.

Methods

This systematic review was conducted in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses PRISMA guidelines.17,18 A study protocol was designed and was not published separately (available as Electronic Supplementary Material [ESM] appendix 1- Methods section of protocol).

Study type

We included clinical studies (both with and without comparator groups) that measured our pre-specified psychologic outcomes of interest. We excluded non-human studies, studies that did not measure our outcomes of interest, and studies that expressed outcomes qualitatively rather than quantitatively. We applied no language, geographical, or report age restrictions.

Participants

We included studies that measured our outcomes of interest among healthcare personnel who cared for adult (≥ 18 yr of age) critically ill inpatients or among the relatives of these patients. We defined a relative as ‘any person (adult or child), related or otherwise who identified themselves or was identified by the healthcare team as being part of the circle of support for the critically ill person and was described in any eligible study was being a relative or family member’.

Interventions

We defined a humanizing intervention as ‘one with substantial potential to increase physical or emotional proximity between the patient and healthcare professionals or between the patient and their relatives’. We chose this definition on the basis of evidence that increased physical and psychologic distances are key contributors to dehumanization.19,20 Both greater and higher quality inter-group contacts are well supported in the psychologic literature as the most reliable method of overcoming dehumanization.1,21-23

While interventions that increase physical proximity (e.g., rounds in the patient’s room, presence of relatives at the bedside, etc.) are easy to identify, those that increase emotional proximity are more subjective. We therefore used a list of potentially humanizing interventions that were agreed on a priori by all authors (Table 1). The list was recognized from the outset as not exhaustive and therefore any non-listed intervention that any author felt might have humanizing potential was discussed with at least one other author. The agreement of at least two authors was required to include or exclude that intervention.
Table 1

Humanizing interventions

Healthcare professional engagement

Humanized briefing on rounds, encouragement to use patient’s name, involvement in diary keeping

Environmental

Photographs, posters providing background information about who the person is, family photos, drawings, personal items, pet visits, music

Revitalization

Sedation minimized, mobilization optimized, hair dressing, own clothes

Patient and family engagement

Open visitation policy, family presence on rounds or during resuscitation, diaries, family involvement in care

We excluded studies of interventions broadly directed at enhancing the emotional or physical health of healthcare professionals or relatives without substantial potential to increase physical or emotional proximity to the patient. We therefore excluded interventions designed to enhance general coping skills or reduce stress among healthcare professionals or relatives including educational programs, resilience training, relaxation techniques, mindfulness training, general communication improvement strategies, stress management programs, and rearrangement of work patterns. Although some of these interventions may indirectly increase appreciation of the patient’s humanity, we felt that it would be impossible to separate out the effects of any indirect humanization from other effects.

Outcomes

To ensure our research question had sound scientific rationale and to optimize the chance that that any association observed was likely to be real, we focused on psychologic outcomes that had either been shown in studies of other populations to be associated with humanization (empathy, burnout, compassion fatigue)13-15,24 or were outcomes that had a plausible association and were likely to develop while the patient was still in hospital (anxiety).

Primary

Empathy among healthcare professionals measured by any validated method.

Secondary

Burnout and compassion fatigue among healthcare professionals and relatives, anxiety among relatives, measured by any validated method, and adverse events of any type in healthcare professionals, relatives, or patients. We did not pre-specify adverse events so as to capture all adverse events reported by included studies.

Excluded outcomes

We excluded outcomes that had a plausible but unproven association with humanization and could not be expected to fully manifest while the patient was still in hospital.

Therefore, post-traumatic stress disorder, complicated grief, depression, and family post - intensive care unit syndrome were not addressed by this review.

Although plausibly associated with humanization and likely to occur while the patient was still in hospital, moral distress was excluded because of a well-recognized lack of conceptual clarity and substantial global differences in terminology, making identification of any association prone to significant inaccuracy.25,26

Settings

We included studies conducted in hospital settings in any country. We excluded studies conducted in pre- or post-hospital settings. For hospital settings, we included patients cared for in any critical care setting including medical, surgical, or specialist critical care units. We also included those cared for in postoperative recovery units, recognizing that critically ill patients may sometimes be cared for in this setting (beyond the immediate postoperative period) because of bedspace limitations.

Eligibility

For a study to be included at least two review authors had to agree that it satisfied all of the above criteria for study type, participants, interventions, outcomes, and settings. Studies that did not satisfy all of the above criteria were excluded.

Search strategy

The following databases were searched: MEDLINE, PsycINFO, Embase, CINAHL, Cochrane Central Register of Controlled Trials (CENTRAL), and ProQuest Dissertations and Theses. The initial search covered their dates of inception to 29 June 2016, with an updated search by the same librarian (SMcK) using the same search criteria covering 29 June 2016 to 29 June 2017. Reference lists, conference abstracts, and the World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) were searched for additional unpublished studies and ongoing studies (available as ESM; Appendix 2-MEDLINE search strategy).

The full electronic search strategies for all databases can be accessed via QSpace, Queen’s University’s research repository service [http://hdl.handle.net/1974/23768].

Study selection

Search results were uploaded to Covidence systematic review software,27 from where four authors (I.M.G., J.L., R.G., and K.P.) independently screened citations. This was done in two stages. In the first stage, authors independently examined all citations and voted ‘Yes’, ‘No’, or ‘Maybe’ using the Covidence blinded voting system. Citations that received a ‘No’ vote by a minimum of two authors were excluded at this stage. All other citations went on to full-text screening.

In the second stage, two authors (I.M.G. and J.L.) independently examined the full-text version of the studies selected in the first stage to determine eligibility. Full-text articles that received two ‘Yes’ votes were included in the review, those that received two ‘No’ votes were excluded, and those that received any other combination (‘Yes/No’, ‘ Yes/ Maybe’, No/Maybe’, or Maybe/Maybe) were examined by a third author (K.P. or R.G.) to determine eligibility. Any conflicts were resolved by discussion among all four authors. To reduce the risk of missing relevant outcomes that might have been measured but not reported by otherwise eligible studies, one author (J.L.) contacted authors of studies of humanizing interventions that did not report our outcomes of interest.

Data extraction

Two authors (I.M.G. and J.L.) independently extracted data from the included studies using a comprehensive data extraction form. Any conflicts were resolved by discussion among the authors.

Risk of bias assessment

Two authors (I.M.G. and J.L.) independently assessed risk of bias for all included studies using Cochrane’s tool for assessing risk of bias as described in the Cochrane Handbook of Systematic Reviews for Interventions.28 Any discrepancies were resolved by discussion. For each included study, bias was assessed in the following seven domains, random sequence generation, allocation concealment, blinding of participants, blinding of outcome assessors, losses to follow-up, reporting bias, and other bias, and deemed to be ‘low’, ‘unclear’, or ‘high risk’, using a priori criteria based on the effect that bias in each domain may have on the validity of the primary outcome (i.e., empathy).

An overall high risk of bias was defined as either a high-risk rating in four or more of the seven domains of bias or a high-risk rating in at least three domains and an unclear risk rating in one or more domains.

Study quality of individual cohort studies

Additionally, for included cohort studies we assessed quality using the Newcastle Ottawa Scale (NOS), a nine-point scale that assigns scores in three domains: selection, comparability, and outcome.29 A score of 7 or above was considered high quality; 5-7, moderate quality; 4 or less, low quality.

Data synthesis

We planned to calculate pooled estimates of effect for the above outcomes where all of the following conditions are met:
  • Outcomes were reported by at least two studies (both judged to have an overall low or moderate risk of bias) using comparable scales and time points.

  • Absence of substantial clinical or methodologic heterogeneity between included studies.

Clinical heterogeneity was assessed by comparing participants, settings, interventions, outcomes, timing of outcome measures, and ancillary treatments. Methodologic heterogeneity was assessed by comparing risk of bias. Where quantitative analysis was feasible, we planned to assess statistical heterogeneity by visual inspection of forest plots, the Chi2 test, and calculation of the I2 statistic, with a P value < 0.1 in the Chi2 test and an I2 statistic > 50% being indicative of significant statistical heterogeneity. Where significant statistical heterogeneity was present, we planned to present the pooled estimate of effect with subsequent discussion as to the likely impact of heterogeneity on the accuracy and quality of the effect estimate.

Where studies reported outcomes on a continuous scale, we planned to calculate mean differences and standardized mean differences, where studies used the same and different scales of measurement, respectively. Where studies reported the outcomes as dichotomous variables we planned to calculate risk ratios. We planned to present all pooled estimates of effect with their respective P values and 95% confidence intervals. Results were considered statistically significant if a P value < 0.05 was achieved. We planned to perform all meta-analyses using a random effects model in Cochrane statistical software Revman 5.3.30 No subgroup or sensitivity analyses were planned.

Results

Study selection

Search results and study selection are depicted in the study flow diagram (Fig. 1).
Fig. 1

Study flow diagram

The initial search returned 11,243 articles and the updated search returned 992 articles, yielding 12,235 articles. Of these, 3,084 were duplicates, leaving 9,151 articles. Nine thousand forty-one were excluded on examination of their abstracts. Articles excluded at this stage included non-human studies, commentaries, letters, and studies in which patient participants were < 18 yr of age.

The remaining 110 articles were assessed for eligibility by examination of their full-text format. Of these, 98 were excluded. Reasons for exclusion were as follows: 41, no intervention tested; 15, outcomes of interest not addressed; three, systematic reviews (these were examined for additional eligible studies and none were identified); one, duplicate study; one, pre-hospital study; 37, intervention did not have substantial potential to increase physical or emotional proximity to the patient. Details of excluded interventional studies are provided (Table 2). 1-81 The remaining 12 studies were included in the review.82-93
Table 2

Excluded interventional studies

Study ID

Intervention

No substantial potential to increase emotional or physical proximity to the patient

 Alford31

Preoperative teaching package

 Ali32

Different physician staffing schedules

 Fernandez33

Simulation-based teaching for residents

 Barbret34

Family-directed information package

 Barnett35

Family-directed information package

 Barsolaso36

Nurse-led family support program

 Bajoka33

Simulation-based teaching for residents

 Beumer37

Staff-focused moral distress workshop

 Bokinskie38

Family-focused conferences to reduce their anxiety about patient transfer from critical care to ward

 Chaboyer39

Family-focused liaison services to reduce their anxiety about patient transfer from critical care to ward

 Chavez40

Family-focused education and orientation program

 Chien41

Family-focused education program

 Curtis42

Family-focused communication facilitator

 Daly43

Family-focused information pamphlet

 Dodd McCue44

Family-focused communication protocol

 Dracup45

Family-focused communication services

 Deore46

Family-focused information package

 Duchemin47

Staff-directed mindfulness program

 De Lucio48

Nurse-focused communication training program

 Garrouste Orgeas49

Nurse participation in family conferences

 Garland50

Physician staffing models

 Halm51

Family-focused support groups

 Harris52

Family-focused psychologic support

 Johnson53

Family-focused telephone intervention

 Kitchens54

Family-focused pre-transfer brochure

 Kowal55

Nurse-focused peer support program

 Krupa56

Educational video

 Lewis57

Family-focused web-based education program

 Lickiewicz58

Family-focused communication algorithm

 Maillet59

Family-focused transfer support services

 Mitchell60

Family-focused support services

 Moreau61

Family-focused information

 Reider62

Family-focused support services

 Ricou63

Nurse-focused psychology services

 Rudnick64

Family-focused information package

 Schooley65

Family-focused teaching

 Singh66

Family-focused educational video

Outcomes of interest not addressed

 Black67

Family participation in psychologic care of the patient. Outcome – patient cognition

 Van den Bulcke68

Music. Outcomes – patient anxiety, sedation, and physiologic variables

 Combe69

Diaries. Outcomes – patient and family perceptions

 Cook70

Three wishes. Outcomes – quality of end-of-life care

 Egerod71

Diaries. Outcomes – how they helped patients construct the illness narrative and affected relatives’ subjective feeling of support

 Lee72

Music and aromatherapy. Outcomes – effects on patient anxiety and physiologic variables

 Koohi73

Family participation in care. Outcomes – pain and anxiety among burn ICU patients

 Mitchell74

Flexible visiting. Outcomes – family and staff satisfaction

 Tracey75

Music. Outcomes – patient and family experiences

 Twibell76

Family presence. Outcomes – family coping strategies

 Gemunden77

Music. Outcomes – patient stress and anxiety

 Huynh78

Diaries. Outcome – family satisfaction

 Fumagalli79

Liberal visitation. Outcomes – patient anxiety and depression and family stress and change in role function

 Blair80

Diaries. Outcome – family and staff feedback

 Locke81

Diaries. Outcome- patient, family, and staff perceptions of usefulness

ICU = intensive care unit

Summary of included studies

The characteristics of included studies are described (Table 3) with further details provided
Table 3

Characteristics of included studies

Liberal Visitation

Study ID

Design

Setting

Participants

Exclusion Criteria

Number Enrolled

Outcomes

Number Followed up

Relevant Results

Holl88

Before and after cohort study

Single-centre cardiac critical care unit, USA

Family members of patients who had undergone open heart surgery admitted to cardiac critical care

Not fluent in English

Age < 18 yr

90

Unclear outcome priority

Patient and family satisfaction and anxiety

(Spielberger, Gorsuch, and Lushene Revised Scale)

Not reported

Non-significant decrease in anxiety levels among family members in liberal visitation group

Giannini87

Before and after cohort study

8 Critical care units, Italy

Medical professionals

None

235

Unclear outcome priority

Burnout and anxiety pre- and 12 months post-intervention

(Maslach-Jackson Burnout Inventory

Anxiety; State-Trait Anxiety Inventory)

198

Burnout levels – Pre- vs 12 months post-intervention

34.5 vs 42.6% (P = 0.001)

Anxiety scores – No significant difference

Orlen89

Prospective controlled cohort study

(control group had important baseline differences)

Single-centre cardiac surgical critical care and medical coronary care unit, USA

Wives of patients admitted to each unit

Not fluent in English

53

Unclear outcome priority

Satisfaction and anxiety among wives (State-Trait-Anxiety Inventory

53

Anxiety levels among patients’ wives at 48 hr

Unlimited visiting hours vs limited visiting hours

47 vs 37 (P = 0.02)

Diaries

Study ID

Design

Setting

Participants

Exclusion Criteria

Number Enrolled

Outcomes

Number Followed up

Relevant Results

Garrouste-Orgeas84

Before and after cohort study comparing pre-diary, diary, and post-diary periods

Single-centre critical care unit, France

Family members of patients admitted to critical care for 4 or more days

Not fluent in English

Family members of patients who died on day 4

Family unavailable on day of patient discharge

Patient had dementia

216

Unclear outcome priority

Patient and family anxiety, depression, and post-traumatic stress at ICU discharge and at 3 months

(Anxiety prevalence was measured as a score > 8 on the State-Trait Anxiety Inventory)

143

Anxiety prevalence among family members at ICU discharge in the 3 groups

23 vs 23 vs 25, P = 0.51

Anxiety prevalence among family members at 3 months in the 3 groups

30 vs 18 vs 26, P = 0.05

Kloos83

Randomized-controlled trial

Single-centre cardiothoracic critical care unit

Families of these patients

Not fluent in English

Age < 18 yr

Relatives of patients who developed major complications

160

State Anxiety among family members on the 3rd postoperative day (State-Trait Anxiety Inventory)

91

Family maintained progress journal group

Anxiety pre- vs post-intervention

45 vs 40

Control group

Anxiety pre- vs post-intervention

46.5 vs 39

Family Involvement in Basic Care

Study ID

Design

Setting

Participants

Exclusion Criteria

Number Enrolled

Outcomes

Number Followed up

Relevant Results

Rodriguez-Martinez90

Prospective controlled cohort study

Single-centre critical care unit, Spain

Family members of these patients

Family members of patients who were not able to communicate verbally

No family available

Family of patients on the unit for < 3 days

117

Unclear outcome priority

Cognitive state of patient, anxiety among family members

Opinions of nurses

(Anxiety was measured using the State-Trait Anxiety Inventory)

56

Anxiety levels of families Intervention vs Control

23.5 vs 31.2 (P = 0.002)

Skoog85

Before and after cohort study

Single-centre cardiac critical care unit, United States

Family members of patients admitted to critical care who had heart surgery in the preceding 12 hr

Families who were not fluent in English

Age < 18 yr

< 6th grade reading and writing skills

64

Anxiety levels among family members before versus after the intervention

(State-Trait Anxiety Inventory)

56

Anxiety levels before and after the intervention 53.6 vs 40.6 (P = 0.001)

Pritchard91

Prospective controlled cohort study

Single-centre critical care unit

Family members of these patients

Family members not fluent in English, pregnant, or who were children

30

Mean change in anxiety levels in both groups

(Hospital Anxiety and Depression Scale)

30

Mean change in anxiety scores on day 3

Intervention vs Control

3.87 vs 0.4

Witnessed Resuscitation

Study ID

Design

Setting

Participants

Exclusion Criteria

Number Enrolled

Outcomes

Number Followed up

Relevant Results

Robinson82

Randomized-controlled trial

Emergency Department, United Kingdom

Family members of patients requiring resuscitation for major trauma

Family members unavailable

None of 3 specific senior staff available, successful resuscitation (survival to follow-up)

25

Unclear outcome priority

Post-traumatic stress, anxiety, depression, grief

Adverse psychologic effects

(Anxiety was measured by Hospital Anxiety and Depression Scale and Beck’s Anxiety Inventory)

16

Anxiety score

Intervention vs Control

9 vs 7 at 3 months

6.5 vs 3 at 9 months

No significant difference at either time point

Adverse psychologic effects –None

Note – Study was stopped early because of lack of clinical equipoise

Pasquale93

Retrospective cohort study

Single-centre level 1 trauma centre, United States

 

Families who were not fluent in English

Family members of patients who were children, prisoners, had burns or self-inflicted injuries

50

Anxiety

Satisfaction

Adverse psychologic effects

(Anxiety was measured using the State-Trait Anxiety Inventory)

50

Anxiety score

Intervention vs Control

43.8 vs 47.6 (P = 0.368)

Adverse psychologic effects –None

Leske92

Prospective controlled cohort study

Single-centre level 1 trauma centre, United States

Family members of patients who were successfully resuscitated after major trauma

Families not fluent in English

Families who had more than one trauma victim or where the patient was a child, had sustained burns, was a prisoner, or victim of domestic violence

Combative or intoxicated families

250

Unclear outcome priority

Measured anxiety, acute stress, wellbeing, satisfaction

(Anxiety was measured using the State-Trait Anxiety Inventory)

140

Anxiety

Intervention vs Control

43.5 vs 48.1 (P = 0.43)

Mixed Intervention (liberal visitation, involvement of families in basic care, improved family communication, staff support workshops, staff debriefing, etc.)

Study ID

Design

Setting

Participants

Exclusion Criteria

Number Enrolled

Outcomes

Number Followed up

Relevant Results

Quenot86

Before and after cohort study

Single-centre critical care unit, France

Medical professionals

None

62

Prevalence of severe burnout

(Maslach-Jackson Burnout Inventory)

Depression

49

Prevalence of severe burnout

Pre- and post-intervention 15 vs 7 (P < 0.01)

Severe burnout was defined as a score > 30 on the emotional exhaustion subscale, or a score > 12 on the depersonalization subscale, or a high total score)

ICU = intensive care unit

(available as ESM; Appendix 3). Only two were randomized-controlled studies,82,83 five were pre- and post-intervention cohort studies,84-88 four were prospective cohort studies with non-randomized controls,89-92 and one was a retrospective study.93 Four discrete humanizing interventions were assessed: diaries,83,84 liberalization of visitation,87-89 witnessed resuscitation,82,92,93 and family participation in basic care.85,90,91 One study of a mixed intervention with both humanizing and non-humanizing initiatives was included on the basis that some of the interventions tested had potential to increase physical and emotional proximity to the patient (liberal visitation and family involvement in basic care).86 Ten studies measured anxiety among a total of 1,055 family members.82-85,88-93 Two studies measured burnout among a total of 288 critical care professionals86,87 No studies addressed empathy or compassion fatigue. Two studies of witnessed resuscitation sought unspecified adverse psychologic effects among relatives and reported none.82,93

Risk of bias in individual studies

Eleven of the 12 included studies had an overall high risk of bias.83-93 Details of risk of bias assessment for each study are provided (Table 3) (Fig. 2). Commonly occurring themes were lack of randomization, lack of valid control groups, unclear outcome priorities, and failure to provide sample size justification.
Fig. 2

Risk of bias in included studies.

Study quality in individual cohort studies

For the ten included cohort studies, none achieved a high-quality rating; five studies received a score of 5,84,85,87,90,93 putting them at the lower end of a moderate rating, and five had a low quality rating86,88,89,91,92 (Table 4).
Table 4

Risk of bias in individual studies

Domain

Risk

Rationale

Holl88

 Random sequence generation

High

Not randomized and no control group

 Allocation concealment

High

None

 Blinding of participants

High

No mention of blinding of those accessing outcome and, although outcomes were measured by participant-completed questionnaires, investigator knowledge of outcomes being assessed may have influenced participants’ responses

 Blinding of outcome assessors

Unclear

Not clear from the report whether all participants were followed up; also the authors state that no baseline measures of anxiety levels were made for family members as many were not available at the time the investigator intended to make these measurements

 Incomplete outcome data

High

No mention of blinding of those accessing outcome and, although outcomes were measured by participant completed questionnaires, investigator knowledge of outcomes being assessed may have influenced participants’ responses

 Selective reporting

High

Outcome priorities were not stated and multiple analyses of variables were made, raising the possibility of selective reporting bias

 Other bias

High

No sample size calculation so difficult to know if the study was adequately powered

Giannini87

 Random sequence generation

High

Non-randomized study. All staff working in the eight participating units received the intervention (liberalization of visitation). Without a control population who were not exposed to a visitation policy change, it is impossible to accurately quantify the effect of the policy change on burnout levels

 Allocation concealment

High

None

 Blinding of participants

Unclear

The term ‘burnout’ was not mentioned to the participants but it is possible that some of them may have guessed what was being measured when they completed the questionnaire

 Blinding of outcome assessors

Unclear

No mention of blinding of those accessing outcome and although outcomes were measured by participant-completed questionnaires, investigator knowledge of outcomes being assessed may have influenced participants

 Incomplete outcome data

High

37 participants were lost to follow-up; the absence of these participants may have distorted the effect size

 Selective reporting

Unclear

Outcomes priorities were not pre-specified

 Other bias

High

No sample size was provided; also the authors acknowledge that confounding factors including other unmeasured changes occurring in participating units may have influenced staff burnout levels

Orlen89

 Random sequence generation

High

No randomization. Patient populations in the two groups were different, one group being patients admitted to the surgical cardiac critical care and the other being those admitted to the medical coronary care unit

 Allocation concealment

High

None

 Blinding of participants

High

No mention if participants were blinded to the outcomes being measured; if they were aware that anxiety was being measured then this may have influenced their responses

 Blinding of outcome assessors

Unclear

No mention of blinding of those accessing outcome and, although outcomes were measured by participant completed questionnaires, investigator knowledge of outcomes being assessed may have influenced participants’ responses

 Incomplete outcome data

Low

No losses to follow-up

 Selective reporting

Unclear

Outcomes priorities were not pre-specified

 Other bias

High

No sample size was calculated so it is difficult to know if the study was adequately powered

Garrouste-Orgeas84

 Random sequence generation

High

Not randomized. Consecutive patients admitted to the Critical Care unit were enrolled. No baseline measures of anxiety were made.

 Allocation concealment

High

None

 Blinding of participants

High

No, and if families were aware of the outcomes being measured it may lessen or exacerbate their anxiety

 Blinding of outcome assessors

Unclear

Investigators performing the follow-up telephone interviews were unaware of which patients and families had received diaries; however, the authors mention that investigator blinding may have been compromised as some participants may have mentioned the diary during the interviews

 Incomplete outcome data

High

Family members who were not present on the day of ICU discharge were excluded. Their omission may have distorted the anxiety profile of the study population

 Selective reporting

High

Outcome priorities were not predefined and they calculated P values for 21 different outcomes, without statistical adjustment for multiple testing of data, increasing the risk of a false-positive results

 Other bias

High

No primary outcome was specified and no sample size was calculated so it is difficult to know if the study was adequately powered

Kloos83

 Random sequence generation

Unclear

Participants were assigned to either the treatment group or the no-treatment group according to week (e.g., odd weeks = no intervention; even weeks = intervention) so participants were not truly randomly assigned

 Allocation concealment

High

None

 Blinding of participants

High

No mention if participants were blinded to the outcomes being measured; if they were aware that anxiety was being measured then this may have influenced their responses

 Blinding of outcome assessors

Unclear

No mention of blinding of those accessing outcome and, although outcomes were measured by participant completed questionnaires, investigator knowledge of outcomes being assessed may have influenced participants’ response

 Incomplete outcome data

High

High losses to follow-up (69/160) without details of the reasons raises the possibility of more or less anxious family members being less likely to complete the study

 Selective reporting

Low

Outcomes were reported as specified

 Other bias

High

No sample size was calculated so it is difficult to know if the study was adequately powered. It is also likely that either group’s responses may be influenced by knowing that there was another group

Rodriguez-Martinez90

 Random sequence generation

Unclear

Participants were not randomly selected; however they had similar baseline anxiety levels, which partially mitigated the risk of selection bias

 Allocation concealment

High

None

 Blinding of participants

High

No mention if participants were blinded to the outcomes being measured; if they were aware that anxiety was being measured then this may have influenced their responses

 Blinding of outcome assessors

Unclear

No mention of blinding of those accessing outcome and, although outcomes were measured by participant completed questionnaires, investigator knowledge of outcomes being assessed may have influenced participants’ responses

 Incomplete outcome data

High

High losses after enrolment with a lack of data as to which group these losses were from and why

 Selective reporting

Unclear

Outcomes priorities were not pre-specified

 Other bias

High

No sample size was calculated so it is difficult to know if the study was adequately powered

Skoog85

 Random sequence generation

High

No randomization and no control group

 Allocation concealment

High

None

 Blinding of participants

High

Participants were not blinded to the outcomes being measured; if they were aware that anxiety was being measured then this may have influenced their responses

 Blinding of outcome assessors

Unclear

No mention of blinding of those accessing outcome and, although outcomes were measured by participant completed questionnaires, investigator knowledge of outcomes being assessed may have influenced participants’ responses

 Incomplete outcome data

Low

No losses to follow-up

 Selective reporting

Low

Outcomes reported as specified

 Other bias

High

Small study and no sample size calculation provided so difficult to determine if the study was adequately powered

Prichard91

 Random sequence generation

High

No randomization and the groups had significant differences in the baseline prevalence of anxiety with significantly greater anxiety prevalence in the control group 93% vs 67%, P = 0.03

 Allocation concealment

High

None

 Blinding of participants

High

Participants were not blinded to the outcomes being measured; if they were aware that anxiety was being measured then this may have influenced their responses

 Blinding of outcome assessors

Unclear

No mention of blinding of those accessing outcome and, although outcomes were measured by participant completed questionnaires, investigator knowledge of outcomes being assessed may have influenced participants’ responses

 Incomplete outcome data

Low

No losses to follow-up

 Selective reporting

Low

Outcomes reported as specified

 Other bias

High

Small study and no sample size calculation provided so difficult to determine if the study was adequately powered

Robinson82

 Random sequence generation

Low

Sealed envelopes; the unit of randomization was the patient

 Allocation concealment

Low

Allocations were concealed by sealed envelopes

 Blinding of participants

Low

The senior nurse in charge of the resuscitation was aware of the allocations but participants were blinded to the outcomes being measured

 Blinding of outcome assessors

Unclear

No mention of blinding of those accessing outcome and, although outcomes were measured by participant-completed questionnaires, investigator knowledge of outcomes being assessed may have influenced participants’ response

 Incomplete outcome data

Unclear

Even though total losses to follow-up were small (7 participants), they were a high percentage (28%) of the overall study cohort and unevenly distributed between the groups, with 5 being lost from the intervention group and 2 from the control group.

 Selective reporting

Unclear

Outcomes priorities were not pre-specified

 Other bias

Unclear

Small study that was stopped early as staff became convinced of the benefits of giving relatives the option of witnessing resuscitation, risking compromise of the randomization process

Pasquale93

 Random sequence generation

High

Not randomized. Compared family members who were present during resuscitation because they opted to be present and were deemed suitable for presence by the trauma team with family members who were not present because of their absence at the time of enrolment, their choice not to be present, or the trauma team’s recommendation that they not be present. Potentially these populations have important baseline differences in their anxiety profiles, which may have introduced selection bias

 Allocation concealment

High

None

 Blinding of participants

High

No mention if participants were blinded to the outcomes being measured; if they were aware that anxiety was being measured then this may have influenced their responses

 Blinding of outcome assessors

Unclear

No mention of blinding of those accessing outcome and, although outcomes were measured by participant-completed questionnaires, investigator knowledge of outcomes being assessed may have influenced participants’ responses

 Incomplete outcome data

Low

No losses to follow-up

 Selective reporting

Low

Outcomes reported as specified

 Other bias

Low

None identified

Leske92

 Random sequence generation

High

Since there was no randomization and the control group was chosen on the basis of their own choice, their physical absence at the time, or the trauma team’s judgement, they were potentially a very different population from the intervention group, which may have led to selection bias

 Allocation concealment

High

None

 Blinding of participants

High

No mention if participants were blinded to the outcomes being measured; if they were aware that anxiety was being measured then this may have influenced their response

 Blinding of outcome assessors

Unclear

No mention if participants were blinded to the outcomes being measured; if they were aware that anxiety was being measured then this may have influenced their responses

 Incomplete outcome data

High

Very high losses to follow-up in the control group (104/174) may have distorted the true results

 Selective reporting

Unclear

Outcomes priorities were not pre-specified

 Other bias

Unclear

Outcomes were measured any time during the first 72 hr, creating potential for differences in anxiety levels due to differences in the time course of the patients’ illness rather than the intervention being studied

Quenot86

 Random sequence generation

High

Not randomized and as the study was conducted over a two-year period other changes in the conditions that the staff were exposed to may have influenced the results

 Allocation concealment

High

None

 Blinding of participants

High

Participants were not blinded to the outcomes being measured; if they were aware that burnout was being measured then this may have influenced their responses

 Blinding of outcome assessors

Unclear

No mention of blinding of those accessing outcome and, although outcomes were measured by participant-completed questionnaires, investigator knowledge of outcomes being assessed may have influenced participants’ responses

 Incomplete outcome data

High

22 of 62 participants lost to follow-up without explanation

 Selective reporting

Low

Outcomes reported as specified

 Other bias

High

No sample size calculation provided, two-year gap between study periods, and multiple interventions tested at the same time make it difficult to know if the study was adequately powered and if the observed differences in outcome were due to time bias or confounding effects of other interventions

ICU = intensive care unit

Outcomes

No eligible studies were found that measured empathy or compassion fatigue in healthcare professionals or relatives.

Burnout among healthcare professionals and relatives

Two studies addressed burnout among healthcare professionals; none addressed burnout among relatives.86,87 Both studies used the Maslach-Jackson Burnout Inventory (MBI) to measure burnout.94 Gianni studied liberal visitation (a minimum of eight hours a day).87 Locally organized staff training sessions were provided at each centre before the change. Burnout among healthcare professionals was measured before and at six and 12 months after the policy change. Baseline levels of burnout were higher among nurses than physicians. A small but significant increase in burnout levels was seen in the year following the policy change, with the increase being greater for nurses than physicians. Of note, staff perceptions of liberal visitation reflected burnout levels with those with high burnout scores expressing more negative opinions than those with low burnout scores.

Quenot also used a before-and-after cohort study to examine the effects of an ‘intensive communication strategy’ that included unrestricted visiting hours, greater family involvement in basic patient care, educational sessions, more frequent family meetings, staff debriefing, role playing, and working groups.86 The strategy was designed to improve communication among healthcare professionals, patients, and families and was designed in response to information obtained by psychologists, who interviewed staff beforehand to identify specific areas for intervention and who remained available for consultation during the study period. The prevalence of severe burnout and depression among staff decreased significantly after the intervention, with all three components of the MBI (emotional exhaustion, depersonalization, and personal accomplishment) showing a significant change. As both studies had an overall high risk of bias and investigated interventions with important differences in their content, a pooled estimate of effect was not calculated.

Anxiety among relatives of the critically ill

Ten studies addressed this outcome.82-85,88-93

Of these ten studies, two measured the effects of liberal visitation on anxiety among relatives,88,89 using the State-Trait Anxiety Inventory.95 Only one reported the actual figures for this outcome and the timing of outcome measurement.89 One study found a significant decrease89 and the other a non-significant decrease in anxiety.88

Two studies evaluated the effects of family-maintained diaries on anxiety among relatives.83,84 One measured the prevalence of anxiety at three months after patient discharge from critical care, using a score of > 8 on a self- administered questionnaire as an indicator of anxiety.84 The other measured mean anxiety levels on the third postoperative day, using the State-Trait Anxiety Inventory.83 Both found a non-significant decrease in anxiety in the diary groups.

Three studies measured the effects of family involvement in basic patient care on anxiety among relatives.85,90,91 Two used the State-Trait Anxiety Inventory,85,90 and one used the Hospital Anxiety and Depression Scale.91 Timing of outcome measurement was reported by only one study.91 All reported significant results.

Three studies investigated the effects of family presence during resuscitation on anxiety among relatives.82,92,93 Two used the State-Trait Anxiety Inventory,92,93 and one used the Hospital Anxiety and Depression Scale and Beck Anxiety Inventory.82 Timing of outcome measures ranged from 48 hr to nine months. None reported significant results.

No pooled estimates of effect were calculated for this outcome as no two studies (each with less than a high risk of bias) of any of the included interventions measured the same outcome using comparable scales at comparable time points.

Adverse events

No adverse events were reported by the two studies that sought them; hence, no pooled estimate of effect was calculated.82,93

Overall quality of evidence

The overall quality of the available evidence was poor with evidence being either low or very low quality. Details and reasons why the evidence was downgraded are provided (summary of findings in Table 5).
Table 5

Study quality – Newcastle- Ottawa quality assessment scale for cohort studies

 

Selection

Comparability

Outcome

Are participants representative of the wider population?*

Were intervention and control groups drawn from the same population?

Can we be sure that the intervention group actually received the intervention ?

Can we be sure that the outcome of interest was absent at baseline or if not that it was adequately adjusted for in the analysis?

Did the study control for severity of illness of the patient population?

Did the study control for participants’ psychologic or psychiatric history?

Was outcome assessor blinded to group assignment?

Was follow-up long enough for the outcome of interest to occur?§

Was follow-up complete?

Holl88

-

□□□□□□□★

-

-

-

-

-

-

Giannini87

□□□□□□□□-

-

-

-

Orlen89

-

-

□□□□□□□★

-

-

-

-

Garrouste-Orgeas84

-

-

□□□□□□□★

-

-

Rodriquez-Martinez90

□□□□□□□□-

-

-

-

Skoog85

-

□□□□□□□★

-

-

-

Prichard91

-

-

-

-

-

-

Pasquale93

-

□□□□□□□★

-

-

-

Leske92

-

□□□□□□□★

-

-

-

-

-

Quenot86

-

□□□□□□□★

-

-

-

-

*Points were not given for studies performed in specialized units such as cardiac or coronary care units, studies that did not clearly define inclusion and exclusion criteria, and studies that excluded large numbers of eligible participants

Points were only given if baseline measures of the outcome of interest were made and were either similar or any inequality was adequately adjusted for in the analysis

Severity of illness of the patient population and participants’ previous psychologic or psychiatric history were thought to be likely confounding factors for psychologic outcomes among healthcare professionals and relatives

§Points were only given if length of follow-up was reported and if follow-up was at least 48 hr for anxiety and empathy and at least one month for burnout and compassion fatigue

Discussion

Summary of main results

We found insufficient evidence to make any quantitative assessment of the effect of any of the included humanizing interventions on empathy among healthcare professionals, anxiety among relatives, or risk of burnout and compassion fatigue in either group. We observed a trend towards reduced anxiety among family members who participated in basic patient care, liberal visitation, and diary keeping. Nevertheless, the quality of the data was low and although the trend was consistent it was not consistently significant. We found conflicting effects of liberal visitation on burnout among healthcare professionals again with the limitation of low data quantity and quality.

Agreements and disagreements with other studies or reviews

To our knowledge, this is the first systematic review that aimed to measure the effects of a range of interventions with potential to humanize care of the critically ill on the psychologic wellbeing of healthcare professionals and relatives. This limits the scope for comparison with similar studies. There are however studies that have addressed similar but more focused aspects of this question. Considering these in terms of individual interventions:

Diaries

Ullman and colleagues reviewed the effect of diaries on critically ill patients and their relatives.96 They found only three eligible studies, only one of which addressed psychologic outcomes in relatives. This single study showed a reduced risk of post-traumatic stress in relatives of patients in the diary group.97 They found no eligible studies addressing other psychologic outcomes. Their review differs from ours in that they focused on diaries alone and only included randomized-controlled studies. Aiken et al. did not limit inclusion criteria to randomized-controlled studies and found 11 eligible studies examining the effects of diaries on psychologic outcomes.98 The majority of these studies reported open-ended patient perceptions and were not designed to measure specific psychologic outcomes. Our own finding of a non-significant trend towards reduced anxiety among family members of critically ill patients is supported only by the two included studies that addressed this outcome83,84 and should be interpreted in the light of the substantial limitations in both quality and quantity of the contributing data (Table 6).
Table 6

Summary of findings for reported outcomes

Liberal visitation

In their review of the benefits of flexible visitation in intensive care, Errasti-Ibarrondo et al.99 found 15 studies describing positive responses from families including increased satisfaction, reduced stress, and reduced anxiety. Their review aimed to identify themes rather than quantify effects. Flexible visitation appears to be well received by relatives,99-101 but the evidence tends to be observational, exploratory, and limited by substantial bias, making it impossible to be sure of its effects in terms of measurable, reliable outcomes. There is a paucity of research regarding the effects of liberal visitation on healthcare professionals. We found only one study addressing the effects of liberal visitation alone on the risk of burnout among medical professionals; this study reported a small but significant increase in risk of burnout when visiting hours were increased to a minimum of eight hours.87 Interestingly, another study of a multifaceted program that included, among other things, more liberal visitation showed a lower prevalence of severe burnout among critical care professionals.86 The potentially important difference was that liberal visitation here was only one aspect of a comprehensive quality improvement program designed in response to issues raised by healthcare professionals, benefited from psychologic expertise from the outset, and included comprehensive staff training and support sessions with a focus on improved communication between healthcare professionals and relatives. This may explain why this study showed a decrease in burnout levels while the other showed an increase. It is both logical and plausible that certain new practices, no matter how desirable, bring with them ‘change stress’ and that unless introduced thoughtfully and with a change support structure may have unanticipated undesirable consequences. It may well be that it is not liberal visitation per se that increases the risk of burnout among healthcare professionals but rather inadequately supported liberal visitation.

Family participation in care

The three studies included in our review showed a consistent trend towards reduced anxiety among family members who had the opportunity to participate in the basic care of their critically ill relative.85,90,91A narrative review of the literature identified an overall theme of increased satisfaction and reduced anxiety among relatives who participated in care. Nevertheless, this review was exploratory and did not attempt to define outcomes or synthesis effects.102

Witnessed resuscitation

The three studies included in our review showed no significant effect of witnessed resuscitation on anxiety among family members.82,92,93 A Canadian Critical Care Society position paper on family presence during resuscitation based on a literature search of the topic found that in general family members were supportive of the opportunity to be present during the resuscitation of their loved ones and physicians and nurses were largely supportive of the practice.11 The summary recommendation was that the practice was ethically sound and should be considered an important component of patient- and family-centred care. The review upon which this recommendation is based differs from our review in that it included randomized-controlled studies only and included family members of patients resuscitated in a pre-hospital setting and family members of pediatric patients.103 The primary outcomes were patient mortality and quality of resuscitation with family psychologic outcomes being secondary.

What this review adds to the current state of knowledge

Our review highlights the current lack of knowledge regarding the effects of humanized care of the critically ill on empathy and compassion fatigue and provides some potentially useful observations regarding its effects on anxiety and burnout. We found a consistent trend towards reduced anxiety among family members who were allowed more time to visit or were involved in diary keeping or basic patient care.83-85,88-91 Although these studies were small and had a high risk of bias, the consistent direction of their results may well imply a true effect. Interestingly, witnessed resuscitation did not reduce anxiety,82,92,93 which makes sense considering the acute emotional turmoil one is likely to experience when watching vigorous attempts to bring a loved one back to life.

The other interesting observation was the opposite directions of effect we found in the two studies that reported burnout among healthcare professionals,86,87 which may speak to the importance of change support strategies when introducing any new initiative no matter how ‘right’ it feels. Taking these results to the bedside of the critically ill person, it’s probably fair to say that involving families and providing comprehensive support for staff who facilitate greater family presence and participation are likely important contributors to humanized care.

Strengths

Novel and important area of research

First, this is a novel and important area of research. Evidence from psychologic studies is often not widely known to clinicians so their relevance to medical practice can be all too easily missed. When we think of translational research we often think of basic science studies, but translation can and should occur in multiple dimensions. Psychology, the study of human behaviour, has particular value as medicine, no matter how technologically advanced, is ultimately delivered by humans. Infrahumanization is well known to psychologists. This study helps to introduce it to medical clinicians and highlight its relevance to critically ill people and those who care for them.

While many studies have investigated various initiatives aimed at making care more person centred, few have identified these interventions as potentially humanizing, investigated a range of such interventions, or addressed quantitative outcome measures that are both clinically relevant and likely to be influenced by humanization.

Quantitative outcomes addressed

Second, although qualitative studies are ideal for exploring, describing, and understanding perceptions and behaviours, they tend to be less generalizable and are not designed to measure specific defined outcomes, magnitude, or directions of effect or to inform risk-benefit estimation.

Examining humanization through a quantitative lens allows important effects that may otherwise be missed to be captured and measured.

For example, some studies that measured quantitative patient outcomes when families were present during resuscitation found concerning results. A study of simulated codes where the relative present was displaying an overt grief reaction showed delayed time to defibrillation and fewer total shocks delivered.104 A retrospective study of more than 300 patients in a medical intensive care unit showed significantly lower rates of return of spontaneous circulation and survival to hospital discharge for patients whose families were present during cardiopulmonary resuscitation.105 Although these results are not consistent across studies, they do raise an important issue. While humanization is innately right, it does not necessarily follow that all efforts to optimize it should be pursued without considering the risk-benefit balance. As for all interventions, we need to know which work best, what the effects are (good and bad), the size and direction of those effects, and for whom the benefits outweigh the risks. Quantitative data about clinically relevant outcomes can provide us with this information.

Robust methodology

Lastly, this review was conducted using a rigorous methodology in accordance with the PRISMA guidelines for systematic reviews.19 The search was not restricted on the basis of language, study date, or location, ensuring maximum retrieval of studies that met our inclusion criteria.

The use of a blinded voting system to select studies minimized the risk of selection bias and the use evidenced-based criteria to define humanizing interventions allowed precise appraisal of the effects of interventions that were likely to have substantial potential in this regard.

Limitations and potential biases

This review has a number of limitations and potential biases

Outcome-based search strategy

Our search strategy was population and outcome rather than population and intervention focused. We chose this approach for two reasons. First, we felt that it allowed us to identify all interventions with humanizing potential that evaluated our outcomes of interest and reduced the risk of missing studies that evaluated interventions that we had not thought of as humanizing a priori. Second, it allowed us to identify a manageable number of studies to screen. Searching for the concept of ‘humanization’ using terms such as humaniz* or respect* retrieved over one third of OVID Medline’s records, which is approximately 9 million citations (where * allows for any truncation of the word). While our choice of search strategy was both logical and feasible, we cannot exclude the possibility that we missed studies that measured our outcomes of interest but described them without using the specific labels of empathy, burnout, compassion fatigue, and anxiety. To mitigate this risk, for studies examined in full-text format and excluded on the basis that they did not report our outcome of interest, one author contacted study authors to see if they measured our outcomes of interest but expressed them in other ways. None were identified.

Focused definition of humanizing interventions

To get a clear view of the effect humanizing interventions on healthcare professionals and relatives, it was necessary to select interventions that had substantial potential to increase the awareness of others of the patient’s humanity. For this reason, we only included studies that had substantial potential to increase patient-healthcare professional or patient-relative inter-group contact through increased physical or emotional proximity. We excluded studies of interventions that were broadly directed at enhancing the emotional or physical health of healthcare professionals or relatives without substantial potential to increase physical or emotional proximity to the patient. This meant excluding educational programs, resilience training, relaxation techniques, mindfulness training, general communication improvement strategies, stress management programs, and rearrangement of work patterns.

Defining humanizing interventions in this way meant that any humanizing potential of broader, more care provider- and relative-directed interventions is not captured in this review. While this may be a weakness in terms of measuring the effect of all interventions that may have any humanizing potential, it allowed a more precise appraisal of those with substantial potential in this regard.

Limited outcomes addressed

Humanization is a psychological concept and cannot be fully understood by quantitative methods alone, with both quantitative and qualitative research having important roles to play in furthering our understanding of this area. For this review, we chose to focus on selected quantitative outcomes, choosing to sacrifice breath of knowledge to get a more focused and accurate view of the current state of knowledge in terms of definable, measurable, and reliable outcomes. This meant that research describing the feelings of patients, relatives, and healthcare professionals in descriptive or narrative terms was not captured in this review. We also excluded outcomes that had a plausible association with humanization but were not likely to fully manifest while the patient was still in hospital; we cannot say whether humanization influences late-occurring psychologic outcomes including complicated grief, depression, and post-traumatic stress. Identifying any association between humanization and more long-term psychologic outcomes quantitatively is subject to substantial distortion by confounding factors. Even if known confounders are adjusted for, there are likely to be many unknown confounders that may blur any real association. Qualitative methods may be much more appropriate for investigating those outcomes.

Moral distress was not evaluated by this review and, although there is no proven association, when we consider that moral distress arises in situations where one is required to act in a manner that conflicts with their own belief of what is morally correct, it is plausible that humanized patient care may lessen or exacerbate moral distress depending on the individual caregiver’s perception. This would be worth investigating in future studies.

Conclusions

This review identifies a knowledge gap regarding the effects of humanizing interventions on clinically important quantitative psychologic outcomes among healthcare professionals and relatives. It does however suggest that relatives of the critically ill are less anxious when they feel more involved or are allowed more time at the bedside and that liberal visitation may have opposing effects on healthcare professional burnout dependant on the context in which it is introduced. One should be cautious, however, about basing practice change on the findings of a few, small, generally poor-quality studies. As we move towards increasing patient and family-focused care, it is important we consider the emotional impact on healthcare professionals and families. While the natural human assumption may be that only good things come of person-focused and family-inclusive care, we do not know to what extent that is true. To get an accurate picture of the broader effects of humanized care, we need to study its effects on patients, families, and healthcare professionals both qualitatively and quantitatively.

Differences between the protocol and the review

In response to expert peer review, we revised our original outcomes to include burnout and compassion fatigue among relatives of the critically ill. As our original search used all the terms needed to identify these studies, we did not repeat the search. Two authors rescreened the studies to identify any that might have measured these outcomes among relatives of the critically ill. None were identified.

Notes

Acknowledgements

We sincerely acknowledge and thank Dr. John Drover and Dr. John Muscedere who both provided advice on study conduct and Dr. Dean Tripp who provided input on psychologic outcome measures and idea development.

Conflicts of interest

None of the authors have any conflicts of interest to declare.

Editorial responsibility

This submission was handled by Dr. Sangeeta Mehta, Associate Editor, Canadian Journal of Anesthesia.

Authors contributions

Imelda M. Galvin conceived the idea, prepared the protocol, selected studies, extracted and assessed the quality of the data, and drafted the manuscript. Jordan Leitch contributed to the development of the idea, selected the extracted studies, assessed the quality of the data, and provided input for manuscript refinement. Sandra McKeown assisted with refining the search strategy, conducted the searches, and wrote the search section of the manuscript. Rebecca Gill and Katherine Poser both screened and selected studies and assisted with manuscript preparation and refinement.

Supplementary material

12630_2018_1227_MOESM1_ESM.pdf (209 kb)
Supplementary material 1 (PDF 208 kb)

References

  1. 1.
    Haslam N, Loughnan S. Dehumanization and infrahumanization. Annu Rev Psychol 2014; 65: 399-423.PubMedGoogle Scholar
  2. 2.
    Leyens JP, Rodriguez-Perez A, Rodriguez-Torres R, et al. Psychological essentialism and the differential attribution of uniquely human emotions to ingroups and outgroups. Eur J Soc Psychol 2001; 31: 395-411.Google Scholar
  3. 3.
    Capozza D, Visintin EP, Falvo R, Testoni I. Deumanizzazione del paziente oncologico nei contesti medici. Salute e Società 2015; 13: 74-86.Google Scholar
  4. 4.
    Trifiletti E, Di Bernardo GA, Falvo R, Capozza D. Patients are not fully human: a nurse’s coping response to stress. J Appl Soc Psychol 2014; 44: 768-77.Google Scholar
  5. 5.
    Todres L, Galvin KT, Holloway I. The humanization of healthcare: a value framework for qualitative research. Int J Qual Stud Health Well-being 2009; 4: 68-77.Google Scholar
  6. 6.
    Harris LT, Fiske ST. Dehumanizing the lowest of the low—neuroimaging responses to extreme out-groups. Psychol Sci 2006; 17: 847-53.PubMedGoogle Scholar
  7. 7.
    Leyens JP, Cortes B, Demoulin S, et al. Emotional prejudice, essentialism and nationalism. The Tajfel lecture. Eur J Soc Psychol 2003; 33: 703-17.Google Scholar
  8. 8.
    Fincher KM, Telock PE, Morris MW. Interfacing with faces: perceptual humanization and dehumanization. Curr Dir Psychol Sci 2017; 23: 288-93.Google Scholar
  9. 9.
    Khalid S, Deska JC, Hugenberg K. The eyes are the windows to the mind: direct eye gaze triggers the ascription of others’ minds. Pers Soc Psychol Bull 2016; 42: 1666-77.PubMedGoogle Scholar
  10. 10.
    Bäckman CG, Walther SM. Use of a personal diary written on the ICU during critical illness. Intensive Care Med 2001; 27: 426-9.PubMedGoogle Scholar
  11. 11.
    Oczkowski SJ, Mazzetti I, Cupido C. Fox-Robichaud AE; Canadian Critical Care Society. Family presence during resuscitation: a Canadian Critical Care Society position paper. Can Respir J 2015; 22: 201-5.PubMedPubMedCentralGoogle Scholar
  12. 12.
    Le Calle GH, Oviés AA, Gómez Tello VG. A plan for improving the humanisation of intensive care units. Intensive Care Med 2017; 43: 547-9.Google Scholar
  13. 13.
    Preston SD, De Waal FB. Empathy: its ultimate and proximate bases. Behav Brain Sci 2002; 25: 1-20.PubMedGoogle Scholar
  14. 14.
    Vaes J, Muratore M. Defensive dehumanization in the medical practice: a cross-sectional study from a health care worker’s perspective. Br J Soc Psychol 2013; 52: 180-90.PubMedGoogle Scholar
  15. 15.
    Cameron CD, Harris LT, Payne BK. The emotional cost of humanity. Anticipated exhaustion motivates dehumanization of stigmatized targets. Soc Psychol Personal Sci 2016; 7: 105-12.Google Scholar
  16. 16.
    Day A, Haj-Bakri S, Lubchansky S, Mehta S. Sleep, anxiety and fatigue in family members of patients admitted to the intensive care unit: a questionnaire study. Crit Care 2013; 17: R91.PubMedPubMedCentralGoogle Scholar
  17. 17.
    Moher D, Liberati A, Tetzlaff J. Altman DG; Prisma Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 2009; 6: e1000097.PubMedPubMedCentralGoogle Scholar
  18. 18.
    PRISMA Preferred Reporting Items for Systematic Reviews and Meta-Analysis. Available from URL: http://prisma-statement.org/prismastatement/Checklist.aspx (accessed August 2018).
  19. 19.
    Haslam N. Dehumanization: an integrative review. Pers Soc Psychol Rev 2006; 10: 232-64.Google Scholar
  20. 20.
    Lee MK, Fruchter N, Dabbish L. Making decisions from a distance: the impact of technological mediation on riskiness and dehumanization. Proceedings of the 18th ACM Conference on Computer Supported Cooperative Work & Social Computing 2015; 1576-89.Google Scholar
  21. 21.
    Tam T, Hewstone M, Cairns E, Tausch N, Maio G, Kenworthy J. The impact of intergroup emotions on forgiveness in Northern Ireland. Group Process Intergr Relat 2007; 10: 119-36.Google Scholar
  22. 22.
    Viki G, Fullerton I, Raggett H, Tait F, Wilshire S. The role of dehumanization in attitudes towards social exclusion and rehabilitation of sex offenders. J Appl Soc Pyschol 2012; 42: 2349-67.Google Scholar
  23. 23.
    Capozza D, Trifiletti E, Vezzali L, Favara I. Can intergroup contact improve humanity attributions? Int J Psychol 2012; 48: 527-41.PubMedGoogle Scholar
  24. 24.
    Fiske ST. From dehumanization and objectification to rehumanization: neuroimaging studies on the building blocks of empathy. Ann N Y Acad Sci 2009; 1167: 31-4.PubMedPubMedCentralGoogle Scholar
  25. 25.
    Pauly BM, Varcoe C, Storch J. Framing the issues: moral distress in health care. Hec Forum 2012; 24: 1-11.PubMedPubMedCentralGoogle Scholar
  26. 26.
    Hamric AB. Empirical research on moral distress: issues, challenges and opportunities. HEC Forum 2012; 24: 39-49.PubMedGoogle Scholar
  27. 27.
    Covidence. A better systematic review management. Available from URL: https://www.covidence.org/ (accessed August 2018).
  28. 28.
    Higgins JP, Green S. Cochrane Handbook for Systematic Reviews of Interventions. Version 5.1.; 2011. Available from URL: https://training.cochrane.org/handbook (accessed August 2018).
  29. 29.
    Wells GA, Shea B, O’Connell D, et al. The Newcastle Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. The Ottawa Hospital. Available from URL: http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp (accessed August 2018).
  30. 30.
    Cochrane Community. Review Manager (RevMan) [Computer program]. Version [5.3]; 2014. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration. Available from URL: https://community.cochrane.org/help/tools-and-software/revman-5 (accessed August 2018).
  31. 31.
    Alford DL. Differences in levels of anxiety of spouses of critical care surgical patients when the spouse participates in preoperative teaching. ProQuest Dissertations and Theses 1993; 1353980.Google Scholar
  32. 32.
    Ali NA, Wolf K, Hammersley J, Hoffmann S, et al. Comparison of alternative intensivist staffing schedules on physician and patient outcomes in the intensive care unit. Am J Respir Crit Care Med 2011; 183: A1026.Google Scholar
  33. 33.
    Bajoka R, Fernandez DC, Markin A, et al. Simulation based encounters with actors improve resident perceptions of their preparedness to conduct family meetings in the ICU. Am J Respir Crit Care Med 2014; 189: A1148.Google Scholar
  34. 34.
    Barbret LC, Westphal CG, Daly GA. Meeting information needs of families of critical care patients. J Healthc Qual 1997; 19: 5-9.PubMedGoogle Scholar
  35. 35.
    Barnett B, Henderson M, Hochhalter A, Spradley C. Impact of videos in the intensive care unit: a novel approach to communication. Chest 2011; 140: 333A.Google Scholar
  36. 36.
    Barsolaso EG, Mendoza JF, Panlilio VA, et al. Effects of assessment, reorientation and therapy (A.R.T.) program in decreasing anxiety and meeting the needs of families of CVA patients. Cerebrovasc Dis 2013; 36: 28-9 (abstract).Google Scholar
  37. 37.
    Beumer CM. Innovative solutions: the effect of a workshop on reducing the experience of moral distress in an intensive care unit setting. Dimens Crit Care Nurs 2008; 27: 263-7.PubMedGoogle Scholar
  38. 38.
    Bokinskie JC. Family conferences: a method to diminish transfer anxiety. J Neurosci Nurs 1992; 24: 129-33.PubMedGoogle Scholar
  39. 39.
    Chaboyer W, Thalib L, Alcorn K, Foster M. The effect of an ICU liaison nurse on patients and family’s anxiety prior to transfer to the ward: an intervention study. Intensive Crit Care Nurs 2007; 23: 362-9.PubMedGoogle Scholar
  40. 40.
    Chavez CW, Faber L. Effect of an education-orientation program on family members who visit their significant other in the intensive care unit. Heart Lung 1987; 16: 92-9.PubMedGoogle Scholar
  41. 41.
    Chien WT, Chiu YL, Lam LW, Ip WY. Effects of a needs-based education programme for family carers with a relative in an intensive care unit: a quasi-experimental study. Int J Nurs Stud 2006; 43: 39-50.PubMedGoogle Scholar
  42. 42.
    Curtis JR, Ciechanowski PS, Downey L, et al. Development and evaluation of an interprofessional communication intervention to improve family outcomes in the ICU. Contemp Clin Trials 2012; 33: 1245-54.PubMedGoogle Scholar
  43. 43.
    Daly K, Kleinpell RM, Lawinger S, Casey G. The effect of two nursing interventions on families of ICU patients. Clin Nurs Res 1994; 3: 414-22.PubMedGoogle Scholar
  44. 44.
    Dodd-McCue D, Tartaglia A. The impact of the Family Communication Coordinator (FCC) protocol on the role stress of hospital chaplains. J Pastoral Care Counsel 2005; 59: 345-60.PubMedGoogle Scholar
  45. 45.
    Dracup K. Beyond the patient: caring for families. Commun Nurs Res 2002; 35: 53-61.PubMedGoogle Scholar
  46. 46.
    Deore S, Shetty P, Pandit R. To educate patients family about intensive care paraphrenalia. Indian J Crit Care Med 2014; 18(Suppl 1): S46 (abstract).Google Scholar
  47. 47.
    Duchemin AM, Steinberg B, Marks D, Vanover K, Klatt M. Small randomized pilot study of a workplace mindfulness-based intervention for surgical intensive care unit personnel: effects on salivary α-amylase levels. J Occup Environ Med 2015; 57: 393-9.PubMedPubMedCentralGoogle Scholar
  48. 48.
    De Lucio LG, López FJ, López MT, Hesse BM, Vaz MD. Training programme in techniques of self-control and communication skills to improve nurses’ relationships with relatives of seriously ill patients: a randomized controlled study. J Adv Nurs 2000; 32: 425-31.Google Scholar
  49. 49.
    Garrouste-Orgeas M, Max A, Lerin T, et al. Impact of proactive nurse participation in ICU family conferences: a mixed-method study. Crit Care Med 2016; 44: 1116-28.PubMedGoogle Scholar
  50. 50.
    Garland A. Effects on patients, physicians and families of 24 hour, on-site intensivist coverage in academic and community ICU care. Am J Respir Crit Care Med 2011; 183: A1024.Google Scholar
  51. 51.
    Halm MA. Effect of support groups on anxiety of family members during critical illness. Dimens Crit Care Nurs 1990; 9: 311.Google Scholar
  52. 52.
    Harries C, Kaur S, Fleming S. Effects of nurse led psychological support on reducing anxiety and satisfying the needs of families with relatives in intensive care: a randomised controlled trial. RCN International. Nursing Research Conference 2005. Available from URL: https://www.rcn.org.uk/-/media/royal-college-of-nursing/documents/professional-development/research/previous-research-conferences/2004-research-book-of-abstracts-and-conference-programme.pdf?la=en&hash=8A5C0AF0A93055A7041F7177900F242BDEC7079B (accessed September 2018).
  53. 53.
    Johnson MJ, Frank DI. Effectiveness of a telephone intervention in reducing anxiety of families of patients in an intensive care unit. Appl Nurs Res 1995; 8: 42-3.PubMedGoogle Scholar
  54. 54.
    Kitchens JL. Effectiveness of an intensive care unit transfer brochure on family member associated knowledge and anxiety. Clin Nurse Spec 2009; 23: 99.Google Scholar
  55. 55.
    Kowal C. A case study of the effect of organizational peer diffusing among critical care nurses. Crit Care Med 2012; 40: 1-328.Google Scholar
  56. 56.
    Krupa A, Litell J, Wilson M, et al. Video education about resuscitation preferences is well-tolerated by patients and their surrogate. Chest 2011; 140: 260A.Google Scholar
  57. 57.
    Lewis CL. A comparison of the effects of a web-based education program about the ICU environment and a standard education program on anxiety, depression, and acute stress experienced among family members of ICU patients. Saint Louis: University of Missouri; 2015 .Google Scholar
  58. 58.
    Lickiewicz J, Serednicki W, Zasada E. Breaking bad news: how to improve communication. Resuscitation 2013; 84: S64.Google Scholar
  59. 59.
    Maillet RJ, Pata I, Grossman S. A strategy for decreasing anxiety of ICU transfer patients and their families. Nursingconnections 1993; 6: 5-8.PubMedGoogle Scholar
  60. 60.
    Mitchell DL. Effect of a structured nursing intervention program on family anxiety in the intensive care unit. ProQuest Dissertations and Theses 1982; 1318271.Google Scholar
  61. 61.
    Moreau D, Goldgran-Toledano D, Alberti C, et al. Junior versus senior physicians for informing families of intensive care unit patients. Am J Respir Crit Care Med 2004; 169: 512-7.PubMedGoogle Scholar
  62. 62.
    Reider JA. Anxiety during critical illness of a family member. Dimens Crit Care Nurs 1994; 13: 272-9.PubMedGoogle Scholar
  63. 63.
    Ricou B, Delaloye S, Merlani P, et al. A psychologist for nurses and nurse-assistants in an ICU: impact on the burnout and the anxiety of the caregivers. Intensive Care Med 2011; 37: S135.Google Scholar
  64. 64.
    Rudnick WC. Relationship of a structured information transfer process for family members of patients transferred from SICU to an in-patient unit on transfer anxiety. ProQuest Dissertation Theses 2000; MQ51794.Google Scholar
  65. 65.
    Schooley YM. Effects of an informational-teaching protocol on stress levels of families of patients in the intensive care unit. ProQuest Dissertations and Theses 1991; 1346879.Google Scholar
  66. 66.
    Singh B, Krupa A, Hinds R, et al. Comforting patients and their health care surrogates and allaying anxiety of resuscitation strategies using audiovisual aids. Am J Respir Crit Care Med 2012; 185: A5205.Google Scholar
  67. 67.
    Black P, Boore JR, Parahoo K. The effect of nurse-facilitated family participation in the psychological care of the critically ill patient. J Adv Nurs 2011; 67: 1091-101.PubMedGoogle Scholar
  68. 68.
    Van den Bulcke B, Dehaene E, De Vos AS, et al. Effects of music intervention on anxiety, stress response, and the need for sedative medication. Crit Care Med 2016; 44: 282.Google Scholar
  69. 69.
    Combe D. The use of patient diaries in an intensive care unit. Nurs Crit Care 2005; 10: 31-4.PubMedGoogle Scholar
  70. 70.
    Cook D, Swinton M, Toledo F, et al. Personalizing death in the intensive care unit: the 3 Wishes Project a mixed-methods study. Ann Intern Med 2015; 163: 271-9.PubMedGoogle Scholar
  71. 71.
    Egerod I, Christensen D. Analysis of patient diaries in Danish ICUs: a narrative approach. Intensive Crit Care Nurs 2009; 25: 268-77.PubMedGoogle Scholar
  72. 72.
    Lee CH, Lee CY, Hsu MY, et al. Effects of music intervention on state anxiety and physiological indices in patients undergoing mechanical ventilation in the intensive care unit. Biol Res Nur 2017; 19: 137-44.Google Scholar
  73. 73.
    Koohi M, Bagheri-Nesami M, Esmaeili R, Seyyed NM, Seyyed HH. Effect of family participation in primary care provision to reduce pain anxiety among burn ICU patients. J Mazandaran Univ Med Sci 2017; 26: 88-99.Google Scholar
  74. 74.
    Mitchell ML, Aitken LM. Flexible visiting positively impacted on patients, families and staff in an Australian intensive care unit: a before-after mixed method study. Aust Crit Care 2017; 30: 91-7.PubMedGoogle Scholar
  75. 75.
    Tracy MF, Chlan L, Staugaitis A. Perceptions of patients and families who received a music intervention during mechanical ventilation. Music Med 2015; 7: 54-8.PubMedPubMedCentralGoogle Scholar
  76. 76.
    Twibell RS. Family coping during critical illness. Dimens Crit Care Nurs 1998; 17: 100-12.PubMedGoogle Scholar
  77. 77.
    Weinhouse G, Gemunden S, Sekhar P, et al. Medical use of song in critical care: the MUSICC trial. Am J Respir Crit Care Med 2015; 191: A5214.Google Scholar
  78. 78.
    Huynh TG, Covalesky M, Sinclair S, et al. Measuring outcomes of an intensive care unit family diary program. AACN Advanced Critical Care 2017; 28: 179-90.PubMedGoogle Scholar
  79. 79.
    Fumagalli S, Calvani S, Gironi E, et al. An unrestricted visitation policy reduces patients’ and relatives’ stress levels in intensive care units. Eur Heart J 2013; 34: 951.Google Scholar
  80. 80.
    Blair KT, Eccleston SD, Binder HM, McCarthy MS. Improving the patient experience by implementing an ICU diary for those at risk of post-intensive care syndrome. J Patient Exp 2017; 4: 4-9.PubMedPubMedCentralGoogle Scholar
  81. 81.
    Locke M, Eccleston S, Ryan CN, Byrnes TJ, Mount C, McCarthy MS. Developing a diary program to minimize patient and family post-intensive care syndrome. AACN Adv Crit Care 2016; 27: 212-20.PubMedGoogle Scholar
  82. 82.
    Robinson SM, Mackenzie-Ross S, Campbell Hewson GL, Egleston CV, Prevost AT. Psychological effect of witnessed resuscitation on bereaved relatives. Lancet 1998; 352: 614-7.PubMedGoogle Scholar
  83. 83.
    Kloos JA, Daly BJ. Effect of a family-maintained progress journal on anxiety of families of critically ill patients. Crit Care Nurs Q 2008; 31: 96-107.PubMedGoogle Scholar
  84. 84.
    Garrouste Orgeas M, Coquet I, Perier A, et al. Impact of an ICU diary on family and patient’s psychological symptoms after an ICU stay. Intensive Care Med 2010; 36: S154.Google Scholar
  85. 85.
    Skoog M, Milner KA, Gatti-Petito J, Dintyala K. The impact of family engagement on anxiety levels in a cardiothoracic intensive care unit. Crit Care Nurse 2016; 36: 84-9.PubMedGoogle Scholar
  86. 86.
    Quenot JP, Rigaud JP, Prin S, et al. Suffering among carers working in critical care can be reduced by an intensive communication strategy on end-of-life practices. Intensive Care Med 2012; 38: 55-61.PubMedGoogle Scholar
  87. 87.
    Giannini A, Miccinesi G, Prandi E, Buzzoni C. Borreani C; ODIN Study Group. Partial liberalization of visiting policies and ICU staff: a before-and-after study. Intensive Care Med 2013; 39: 2180-7.PubMedGoogle Scholar
  88. 88.
    Holl RM. The effect of role-modeled visiting in comparison to restricted visiting on the well being of clients who had open-heart surgery and their significant family members in the critical care unit. ProQuest Dissertation and Theses 1992; 9225603.Google Scholar
  89. 89.
    Orlen CA. Visitation in the intensive care unit: satisfaction with limited and unlimited visiting hours and anxiety levels of patients’ wives. ProQuest Dissertations and Theses 1993; 1355049.Google Scholar
  90. 90.
    Rodriguez Martínez M, Rodriguez Morilla F, Del Pino AR, et al. Family involvement in the critically ill patient basic care (Spanish). Enferm Intensiva 2003; 14: 96-108.Google Scholar
  91. 91.
    Prichard C, Newcomb P. Benefit to family members of delivering hand massage with essential oils to critically ill patients. Am J Crit Care 2015; 24: 446-9.PubMedGoogle Scholar
  92. 92.
    Leske JS, McAndrew NS, Brasel KJ, Feetham S. Family presence during resuscitation after trauma. J Trauma Nurs 2017; 24: 85-96.PubMedPubMedCentralGoogle Scholar
  93. 93.
    Pasquale MA, Pasquale MD, Baga L, Eid S, Leske J. Family presence during trauma resuscitation: ready for primetime? J Trauma 2010; 69: 1092-9.PubMedGoogle Scholar
  94. 94.
    Maslach C, Jackson SE. The measurement of experienced burnout. J Organ Behav 1981; 2: 99-113.Google Scholar
  95. 95.
    Spielberger CD, Gorsuch R, Lushene RE, Vagg PR, Jacobs GA. Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press; 1983 .Google Scholar
  96. 96.
    Ullman AJ, Aitken LM, Rattray J, et al. Intensive care diaries to promote recovery for patients and families after critical illness: a Cochrane systematic review. Int J Nurs Stud 2015; 52: 1243-53.PubMedGoogle Scholar
  97. 97.
    Jones C, Bäckman C, Griffiths RD. Intensive care diaries and relatives’ symptoms of posttraumatic stress disorder after critical illness: a pilot study. Am J Crit Care 2012; 21: 172-6.PubMedGoogle Scholar
  98. 98.
    Aitken LM, Rattray J, Hull A, Kennedy JA, Le Brocque R, Ullman AJ. The use of diaries in psychological recovery from intensive care. Crit Care 2013; 17: 253.PubMedPubMedCentralGoogle Scholar
  99. 99.
    Errasti-Ibarrondo B, Tricas-Sauras S. Benefits of flexible visitation in the intensive care units for the family of critical patients (Spanish). Enferm Intensiva 2012; 23: 179-88.PubMedGoogle Scholar
  100. 100.
    Fumagalli S, Calvani S, Gironi E, et al. An unrestricted visitation policy reduces patients and relatives’ stress levels in intensive care units. Eur Heart J 2013; 34: P5126.Google Scholar
  101. 101.
    Mitchell ML, Aitken LM. Flexible visiting positively impacted on patients, families and staff in an Australian intensive care unit: a before-after mixed method study. Aust Crit Care 2017; 30: 91-7.PubMedGoogle Scholar
  102. 102.
    Aliberch Raurell AM, Miguel Aymar IM. Need for the role of the patient’s family members at the intensive care unit (Spanish). Enferm Intensiva 2015; 26: 101-11.PubMedGoogle Scholar
  103. 103.
    Oczkowski SJ, Mazzetti I, Cupido C, Fox-Robichaud AE. The offering of family presence during resuscitation: a systematic review and meta-analysis. J Intensive Care 2015; 3: 41.PubMedPubMedCentralGoogle Scholar
  104. 104.
    Fernandez R, Compton S, Jones KA, Velilla MA. The presence of a family witness impacts physician performance during simulated medical codes. Crit Care Med 2009; 37: 1956-60.PubMedGoogle Scholar
  105. 105.
    Krochmal RL, Blenko JW, Afshar M, et al. Family presence at first cardiopulmonary resuscitation and subsequent limitations on care in the medical intensive care unit. Am J Crit Care 2017; 26: 221-8.PubMedGoogle Scholar

Copyright information

© Canadian Anesthesiologists' Society 2018

Authors and Affiliations

  • Imelda M. Galvin
    • 1
    • 2
    • 5
    Email author
  • Jordan Leitch
    • 1
    • 2
    • 5
  • Rebecca Gill
    • 3
    • 5
  • Katherine Poser
    • 4
    • 5
  • Sandra McKeown
    • 2
    • 5
  1. 1.Department of Anesthesiology & Perioperative Medicine and Department of Critical Care MedicineKingston Health Sciences CentreKingstonCanada
  2. 2.Queens UniversityKingstonCanada
  3. 3.Critical Care, Kingston Health Sciences CentreKingstonCanada
  4. 4.St Lawrence CollegeKingstonCanada
  5. 5.Surgical Perianesthesia ProgramKingston Health Sciences CentreKingstonCanada

Personalised recommendations