Anda di halaman 1dari 7

Journal of Critical Care (2013) 28, 483489

A narrative-based study on communication by family

members in intensive care unit,,
Stefania Di Gangi PhD , Giuseppe Naretto Dr, Nicole Cravero, Sergio Livigni Dr
Department of Anesthesiology and Resuscitation, San Giovanni Bosco Hospital, Turin, Italy

Narrative-based medicine;
Purpose: This study investigates how informative stories are, as written by patients' families in an
intensive care unit (ICU) guest book, in terms of families' emotional responses, needs, perceptions, and
satisfaction with the quality of care supplied.
Intensive care medicine;
Materials and Methods: Design was retrospective observational. Spontaneously written stories (440),
Cluster analysis
gathered between 2009 and 2011, described experiences of 332 family members and 258 patients.
Multivariate information from stories was analyzed using cluster analysis.
Results: Most frequently, stories were written in the form of letters addressed to patients (38%, 168
stories). Family members wrote mainly to give encouragement and to motivate patients to live (34%,
150 stories), expressing love or affection (56%, 245 stories). Feedback to ICU staff was provided in 65
stories, and competence was the most relevant skill recognized (31%, 20 stories). Cluster analysis
highlighted links between positive feedback and families' positive emotional responses.
Conclusions: The study suggests that ICU guest books can be an effective and simple means of
communication between the family, the patient, and the ICU staff. Families shared thoughts, feelings, or
opinions, which were meant to be supportive for the patients or rewarding for the staff.
2013 Elsevier Inc. All rights reserved.

1. Introduction

Authors' contributions: SDG analyzed the data and drafted the
manuscript. GN supported the study, participated in data collection, and Intensive care unit (ICU) diaries are ongoing records of
contributed to the drafting of the manuscript. NC participated in data collection patients' stay in the ICU, which are kept by a third party and
and data management and reviewed the manuscript. SL contributed to the
often accompanied by pictures [13]. Intensive care unit
drafting and revision of the manuscript for important intellectual content. All
authors read and approved the final version of the manuscript. SDG takes diaries have shown to help patients to remember and under-
responsibility for the integrity of the work as a whole. stand their experience in the ICU [1,2]. They have been used

Institution where work performed: Department of Anesthesiology as a therapeutic tool for psychosocial recovering of patients
and Resuscitation, San Giovanni Bosco Hospital, Turin, Italy. and following up after discharge [4,5], but they have also

Conflicts of interest: All the authors confirm that they have no
been considered as an integral part of the patient's care [6].
conflicts of interest, in any form, in relation to this work.
Corresponding author. S.C. Anestesia e Rianimazione, Ospedale Though diaries are usually written by nurses [15], par-
Giovanni Bosco, Piazza Donatore di Sangue 3, 10154 Torino, Italy. ticipation of patients' families in diary keeping should be
E-mail address: (S. Di Gangi). encouraged [6]. Diaries may improve communication between

0883-9441/$ see front matter 2013 Elsevier Inc. All rights reserved.
484 S. Di Gangi et al.

patients, who are usually unable to speak for themselves, and Form. How were they written?
their families [1,4,6]. It has been shown that writing diaries Feedback to staff. Was feedback to ICU staff provided?
helps families to share their experience, feelings, and presence
with the patients [7]. Moreover, participation of families in When possible, we abstracted from each story the fol-
writing patients' diaries enables patient/family-centered care lowing information:
[1], whose definition requires patients' family members to be
involved in clinical decision making and care [8]. - date and time;
On the other hand, effects of family interactions in the - family member's name or alias, sex, and relationship
ICU are mainly influenced by ICU staff [811]. It has been with the patient;
shown that good communication and collaboration between - patient's name or alias;
patients' family members and ICU staff help to improve the - emotional responses (hope, encouragement, and so on);
quality of care supplied [8,1216]. A collaborative relation- - main purpose of the story (catharsis or emotional
ship, between families and ICU staff, requires ICU staff to release, acknowledgement, and so on);
understand and meet families' needs [8], which have been - expressive medium (drawing, poetry, and so on);
the focus of several studies [10,12,1720]. To the best of our - narrative mode (second-person narrative referred to the
knowledge, family diaries have not yet been used as a means patient);
to learn about family members' needs, perceptions, or satis- - medical history (including patient demographics);
faction with the work of ICU staff. - patient outcome (death, improvement, transferring to
This study focuses on an ICU guest book, whose narratives the general ward);
can be seen as coming from a collection of diaries written by - feedback to ICU staff (professionalism, humanity, and
families or patients' visitors. We aim to investigate how in- so on).
formative stories are, as read from the guest book, in terms of
family members' emotional responses, needs, perceptions, The keywords, used for emotions, purposes, media, and
and satisfaction with the quality of care supplied. feedback, were determined reading the narratives and are
conceptual codes for similar groups rather than specific terms
used by narrators. In this context, we defined poetry as any
2. Materials and methods story that used imaginative language or metaphors. The
keyword coding was done by a single person.
We undertook a retrospective observational study based on
a collection of unstructured stories written by patients' visitors. 2.1. Statistical analysis
The study was carried out in the ICU of Ospedale
Giovanni Bosco, Torino, Italy, as a part of an ongoing project, Qualitative data analysis was combined with cluster
named H.story (, whose aim is to analysis techniques [21], which allowed visualization of
promote narrative-based medicine in health care. multivariate information from stories and hidden pattern
We focused on stories, as collected in the ICU guest book, recognition.
from July 2009 to December 2011. These stories were A tag cloud [22] was used to visually represent the
written spontaneously by individuals (relatives, friends, or keywords describing the contents of stories. More frequent
acquaintances) who visited the patients in the ICU. We tags appeared darker and larger than the less frequent ones.
referred to them as family members. All persons were Hierarchical agglomerative cluster analysis [21] was used to
informed that their private accounts might have been used for find groups (clusters) of homogeneous keywords. We started
research purposes, and their written informed consent was with individual keywords as clusters and successively merged
requested before writing in the guest book. them according to Ward's linkage rule [23]. We defined the
Participation of all family members was voluntary but distance between 2 keywords to be inversely proportional to
encouraged actively throughout the duration of the study. the number of stories containing them. Moreover, we penalized
We allowed narrators the freedom to decide about the pairs in stories with many keywords through a weight that
contents and information provided to encourage openness decreased with the number of keywords. This was done to
and to have a better understanding of what was relevant to highlight pairs of keywords appearing alone. Let a and b be 2
them. They were also allowed to structure the writing any keywords, Ii(a,b) be an indicator that equals 1 if the ith story
way they wished or to express their feelings using art form contains a and b and 0 otherwise, and mi be the number of
(drawings, songs, poetry, and so on). All stories were hand- keywords in the ith story. The distance between a and b, D(a,
written in the Italian language. b) was defined as follows:
Analysis of stories was focused on answering the fol- h X i
lowing questions: Da; b exp i I i a; b=m i :

Contents. What did stories say about family members? This distance was the similarity measure used in the
Purpose. Why were they written? algorithm. A dendrogram, that is, a tree diagram, was used to
Study on communication by family members in ICU 485

plot the sequential merging of clusters based on the degree

of similarity.
Cluster analysis techniques were used for describing the
contents of stories both in terms of feelings and feedback
provided to staff.
All statistical analyses were performed using R. Word
Cloud 2.0 package (
wordcloud/index.html) was used to plot the tag cloud.

3. Results

A total of 440 stories, from July 2009 to December 2011,

were collected and analyzed. In this period, 612 patients
were admitted to the ICU, and we supposed stories referred
to the 398 patients that stayed in ICU more than 48 hours. Fig. 1 Tag cloud of emotional responses. Frequencies are shown
Sometimes the same patient was visited by more than 1 in parentheses as percentages of the total number of stories (440).
On average, each story was described by 4 keywords.
person, or the same person wrote more than once. In fact, we
counted 258 patients and 332 family members, although we
had some missing values in patient and family members'
names. The mean number of entries per patient was 1.71, frequent than thanksgiving (15 against 8%, or 67 against
with standard deviation equal to 2.41. 34 stories).
We observed that 54 patients (28%) were female and 139 Regarding acceptance of patient's illness and stay in ICU,
(72%) were male, with 65 missing values. We found an we found that in 29 stories (7%) family members expressed
opposite pattern in sex composition for family members: 118 resentment and in 13 (3%) disbelief. We used the tag aw-
(68%) were female and 56 (32%) male, with 158 missing areness in 6 stories (1%) when family members stated that
observations. their experience in the ICU made them aware of the essence
Regarding the familial relationship with the patient, we did and the biological significance of life.
not show the complete distribution, but we found that 82 Fig. 2 reports a dendrogram showing how elements
(42%) family members were children, 32 (17%) grand- merged together in clusters. The first pair to form a cluster
children, 17 (9%) nephews, 14 (7%) siblings, 10 (5%) parents, was love/affection and encouragement. In fact, very often,
and 9 (5%) spouses. The missing values were 139. Of the 258 we found messages where expressions of affection followed
patients, we learned from the stories that 14 (5%) died. encouragement, for example, Come on, you can do this. We
In 65 (15%) of the 440 stories, feedback to ICU staff love you. Proceeding with the merging, the tag awaiting
was provided by family members who judged caregivers' joined the cluster. We chose to stop after having identified
skills. In addition, there were 7 (2%) of the 440 stories where the 8 clusters enclosed in dashed rectangles in Fig. 2 and
merely gratitude was expressed. labeled as appropriate. We observed that acknowledgement
or appreciation toward ICU staff was in the same cluster as
3.1. Contents of stories wish, hope, contentment, and solidarity. Nevertheless, family
members expressed gratitude even in difficult situations. In
Contents of stories were described, in terms of emotional fact, we found that, in 8 stories (2%), family members
responses, by a total number of 1749 keywords, of which praised staff after patient's death.
46 were unique tags, that is, an average of 4 keywords
per story. The maximum number of keywords in a story 3.2. Purpose of stories
was 13.
Fig. 1 reports the tag cloud showing frequencies of tags as To understand better family members' needs, we studied
percentages of the total number of stories (440). We observed the reasons or motivations that induced them to write.
that love/affection was the most frequent tag (56%, 245 Results are summarized in Fig. 3, which shows the
stories), followed by encouragement (40%, 176 stories) and distribution of the main purpose of stories. Frequencies are
awaiting (33%, 147 stories). The tag support was used in expressed as percentages of the total number of stories (440).
120 stories (27%). We used the tag acknowledgement or ap- The main reason to write a story for a family member was
preciation when gratitude or appraisal was expressed toward to encourage the patient or sometimes oneself (34%, 150
ICU staff, while the tag thanksgiving was used when the stories). Expressing good wish followed with 13% (56
expression of gratitude was to God or to the patient. We stories) and then catharsis, or emotional release, with 12%
noticed that acknowledgement or appreciation was more (53 stories).
486 S. Di Gangi et al.

Fig. 2 Cluster dendrogram of emotional responses. Final clusters are enclosed in dashed rectangles, and the first and last ones are labeled.

A number of stories were written to express gratitude Fig. 4 shows the distribution of expressive media used for
toward ICU staff or to thank God, respectively, 25 (6%) and stories. We observed that, most frequently, stories were
17 (4%). Another 2% (8 stories) were written to encourage written in the form of letters addressed to the patients (38%,
reflection on ICU experience and on life in general. 168 stories). Short messages were preferred in 158 stories
Moreover, 5 stories (1%) expressed solidarity with family (36%), followed by narrative writings (13%, 58 stories).
members of other patients. Poetry was used in 17 stories (4%), and we would like to
quote some examples: You drive to my heart in a fast lane,
3.3. Expressive form of stories Life is a wheel that goes up, then down, then up again and
makes you appreciate the ride. Drawing was as frequently
Stories were usually written in everyday language and used as poetry (4%, 18 stories).
with simple sentence structure. Very often, we found errors
in spelling, grammar, or sentence structure. Almost all stories 3.4. Feedback to staff
(88%, 387) used a second-person narrative that referred to
the patient. We summarized the feedback provided to the ICU staff
using 34 unique keywords describing the skills attributed to
the staff. On average, 2 keywords per story were used, for a
total of 147 skills. The maximum number of keywords in a
story was 9.
Fig. 5 shows the skills of ICU staff, as judged by family
members, reporting percentages of the 65 stories in which
feedback was provided.
We observed that feedback to ICU staff was mostly
positive. Competence was praised in 20 stories (31%). Pa-
tient and family concern were revealed in 15 stories (23%),
and trustworthiness in 14 (22%).
In 10 stories (2%), family members wrote that caregivers
supplied heavenly care. In fact, physicians and nurses were
called angels on earth or blue and greenish angels,
referring to the color of the uniform they wore. These cases
Fig. 3 Reasons to write a story. Frequencies are shown as per- were classified under the tag excellence.
centages of the total number of stories (440). Only 1 motivation was Fig. 6 reports the cluster dendrogram. The first pair that
abstracted from each story. clustered together was trustworthiness and competence.
Study on communication by family members in ICU 487

normal rooms. A child expressed loss, desolation, and wait

in a picture of a child looking at an airplane in the sky near
his house.
With the expressions needle stabbed in the stomach and
the heart of lead, a family member expressed his suffering
in a metaphoric way that recalled patient's interventions. In
other cases, family members, for their support to the patients,
likened themselves to the light in the dark, stars in the
night, or the sun through clouds that come and go.

3.5.2. Communication with patients

Family members expressed their need to confide their
sorrows, regrets, or unexpressed feelings to the patient: I
fear I have no more chances to speak with you, but I hope
Fig. 4 Expressive media used for stories. Frequencies are shown
that one day you will read my message.
as percentages of the total number of stories (440). Only 1 medium Some family members admitted it was difficult to express
was abstracted from each story. their thoughts in spite of their willingness to write: I'd like
to tell you more, but I think I can't find the right words.

Final clusters are enclosed in dashed rectangles and labeled 3.5.3. Interactions with the ICU staff
as appropriate. In particular, the cluster named Narrative Three family members raised doubts about medicine in
approach pointed out that humanity was close to excellence general or felt impotent, writing There is no result of
and professionalism. clinical examination that can destroy hopes and Being a
mere mortal, I am not able to interfere with the Almighty.
3.5. Qualitative analysis Trustworthiness of staff helped family members to en-
courage themselves and the patients. In fact, very often,
Qualitative analysis gave us more insights into percep- family members reassured patients, writing Don't worry!
tions and needs of family members. You are getting very good care. Meanwhile, we found that
distrust was accompanied by sense of impotence, fear, and
3.5.1. Perceptions of ICU anxiety. A family member wrote: Family can only hope
Family members perceived ICU stay as a nightmare, a that ICU staff does not make mistakes and that treats the
journey without destination, a dark vortex, a prison patient gently.
where the only chance of escape is healing, or stressful ups Family members expressed their need to be supported by
and downs where improvement is accompanied by worsen- the staff. Some of them described their visit as an experience
ing and where joy lives with fear. The ICU appeared more of pain soothed by the help of the staff: The smile and the
frightening and unusual than the general ward. In fact, a kindness of the staff are a balm to the heart.
family member in a story referred to the general ward as Unrestricted visiting policy in the ICU was very wel-
comed: Thank you for your openness, in every sense of
the word.
Some remarkable comments were particularly rewarding
for the staff: Wonderful people who give back the joy,
ICU staff leaves a mark in your heart that will change you
forever, ICU staff: the best I have ever known. They
should teach their skills to the other wards, and finally,
Always work like that.

4. Discussion

To our knowledge, this is the first study that used guest

book stories from patients' family members and combined
qualitative and quantitative research, analyzing information
Fig. 5 Feedback to ICU staff. Frequencies are shown as from stories in a multivariate setting. We chose a guest book
percentages of the number of stories (65) where feedback was because of its potential to encourage sharing thoughts and
provided. On average, these stories were described by 2 skills. leaving comments or greetings without revealing one's
488 S. Di Gangi et al.

Fig. 6 Cluster dendrogram of feedback to staff. Final clusters are enclosed in dashed rectangles and labeled when the number of their
elements was greater than two.

identity. In practice, it was not different from a collection of families and ICU staff could benefit from these diaries.
family diaries. Intensive care unit guest book was revealed as a means to
Similarly to other studies [2,3], we aimed at exploring the help families to communicate with nonspeaking patients.
narrative content of diaries, but unlike them, we did not focus On the other hand, staff could benefit from reading these
on the narrative structure of each story. Moreover, differently stories. For example, they might gain understanding of
from [3], we did not analyze contents of the events leading patients, and, as a consequence, they might follow a more
up to the patient's admission or information about the clinical ethical decision making [24]. Moreover, they could
outside world, such as family news, football scores, and so find reward, acknowledgement, or suggestion for improving
on. This kind of information has shown to be useful for the their work through feedback provided by families. As an
recovery of patients after discharge [5], but this was beyond example, we found evidence that some families were fully
our purposes. satisfied with the unrestricted visiting policy in our ICU and
We focused on families, and we described their in- with the support they received from the staff. These are in fact
teractions with the patients through their emotional responses 2 of the relevant family needs already investigated in the
and feelings and their interactions with the ICU staff through literature [8,11,12,17,25].
their feedback provided to them. We found that the main Cluster analysis on feedback to staff highlighted that
feelings expressed to patients were love and affection (56%, professionalism and excellence were linked to humanity.
245 stories) and that families were satisfied and considered This is in fact the foundation of the narrative-based approach
very positively the work of the staff. Competence of staff [26], which we would like to pursue.
was in fact the main skill praised (31%, 20 stories).
We analyzed the purpose and narrative form of stories to
gain insights into the role of diaries for families and into 4.1. Limitations
families' communication needs. The main reason to write a
story for a family member was to encourage the patient or The analysis of data from stories presented some
sometimes oneself (34%, 150 stories). This result is difficulties. First, collection of data depended on the writing
consistent with the dual roles of the diary, supportive and skills of narrators and their willingness to share thoughts and
supported, highlighted in ref. [1]. We found that stories were opinions. Moreover, negative judgments could have been
written, most frequently, in the form of letters addressed to underestimated because of feelings of vulnerability or privacy.
patients (38%, 168 stories). This is consistent with what was Second, the content of stories was free and unstructured.
found in ref. [1] about the use of the diary as a pretext for Narrators wrote what was important to them regardless of its
communication or as an opening for a conversation. relevance to our research questions.
Although our study was not intended to determine the Furthermore, choice of tags and analysis of con-
effect of the ICU guest book, but to understand the content tents of stories, in addition to statistical skills, required em-
of its stories, our findings suggested that both patients' pathy to identify emotional responses and motivations for
Study on communication by family members in ICU 489

writing. This was meant to bridge the divide between [6] Combe D. The use of patient diaries in an intensive care unit. Nurs Crit
researchers and study participants with a humane approach Care 2005;10:31-4.
[7] Roulin MJ, Hurst S, Spirig R. Diaries written for ICU patients. Qual
to data analysis. Health Res 2007;17:893-901.
However, the effects of subjective interpretation could be [8] Davidson JE, Powers K, Hedayat KM, et al. Clinical practice
reduced using machine learning and artificial intelligence guidelines for support of the family in the patient-centered intensive
algorithms [27] that allow a computer search and an care unit: American College of Critical Care Medicine Task Force
20042005. Crit Care Med 2007;35:605-22.
automatic classification of the rough contents of stories.
[9] Azoulay E, Pochard F, Chevret S, et al. Family participation in care to
Further research could aim both to stratify analysis, by critically ill: options of family and staff. Intensive Care Med 2003;29:
relationship to patients or by sex, and to investigate better the 1498-504.
links between narrators' feelings and feedback to staff. [10] Latour JM, van Goudoever JB, Duivenvoorden HJ, et al. Perceptions
of parents on satisfaction with care in the pediatric intensive care unit:
the EMPATHIC study. Intensive Care Med 2009;35:1082-9.
[11] Pronovost PJ, Rodriguez-Paz J, Mohammad Z. Creating competent
5. Conclusions and caring physicians: ensuring patients are our North Star. Intensive
Care Med 2007;33:1873-5.
[12] Azoulay E, Pochard F, Chevret S, et al. Meeting the needs of intensive
With this study, we proposed methods to explore contents care unit patient families: a multicenter study. Am J Respir Crit Care
from ICU guest books. We would like to encourage research Med 2001;163:135-9.
through family diaries in ICUs because its information may [13] de Vos M, Graafmans W, Keesman E, et al. Quality measurement at
stimulate ICU-level quality improvement efforts. Moreover, intensive care units: which indicators should we use? J Crit Care 2007;
it would be helpful to understand if our findings may be [14] Lautrette A, Darmon M, Megarbane B, et al. A communication
representative of the Italian context as a whole or even of strategy and brochure for relatives of patients dying in the ICU. N Engl
the context of other countries. J Med 2007;356:469-78.
We investigated how written material in ICU guest books [15] McAdam JL, Puntillo K. Symptoms experienced by family members
could inform us about family members' emotional responses of patients in Intensive Care Unit. Am J Crit Care 2009;18:200-9.
[16] Azoulay E, Timsit JF, Sprung CL, et al. Prevalence and factors of
to their loved one's critical illness, as well as their needs, intensive care unit conflicts: the conflicus study. Am J Respir Crit Care
perceptions, and satisfaction with the quality of care. We Med 2009;180:853-60.
suggest that narrative stories can be an effective and simple [17] Bijttebier P, Vanoost S, Delva D, et al. Needs of relatives of critical
means of communication between the family, the patient, care patients: perceptions of relatives, physicians and nurses. Intensive
Care Med 2001;27:160-5.
and the ICU staff. Stories may or may not have positive
[18] Daley L. The perceived immediate needs of families with relatives in
effects on family experiences or patient outcomes. If clini- intensive care setting. Heart Lung 1984;13:231-7.
cians were to read stories, this may enhance their empathy [19] Titler MG, Cohen MZ, Craft MJ. Impact of adult critical care hos-
and understanding of the family perceptions and needs. pitalization: perceptions of patients, spouses, children, and nurses.
Reflecting on such stories could provide insights about ways Heart Lung 1991;20:174-82.
to improve clinical practice. [20] Davidson JE. Family-centered care: meeting the needs of patients'
families and helping families adapt to critical illness. Crit Care Nurse
[21] Hastie T, Tibshirani RF. Hierarchical clustering. In: Hastie T,
Tibshirani RF, editors. The elements of statistical learning. 2 ed.
References New York: Springer; 2009. p. 520-8.
[22] Hammond T, Hannay T, Lund B, Scott J. Social bookmarking tools (I): a
[1] Egerod I, Christensen D, Schwartz-Nielsen KH, grd AS. Con- general review. DLib Mag 2005;11(4),
structing the illness narrative: a grounded theory exploring patients' pril2005-hammond.
and relatives' use of intensive care diaries. Crit Care Med 2011;39: [23] Ward JH. Hierarchical grouping to optimize an objective function.
1922-8. J Am Stat Assoc 1963;58:236-44.
[2] Egerod I, Christensen D. A comparative study of ICU patient diaries [24] Husted JH, Husted GL. Ethical decision making in nursing and health
vs. hospital charts. Qual Health Res 2010;20:1446-56. care: The symphonological approach. New York: Springer; 2008.
[3] Egerod I, Christensen D. Analysis of patient diaries in Danish ICUs: a [25] Giannini A, Miccinesi G, Leoncino S. Visiting policies in Italian
narrative approach. Intensive Crit Care Nurs 2009;25:268-77. intensive care units: a nationwide survey. Intensive Care Med 2008;34:
[4] Jones C, Bckman C, Capuzzo M, et al. Intensive care diaries reduce 1256-62.
new onset post traumatic stress disorder following critical illness: a [26] Charon R. Narrative medicine: a model for empathy, reflection,
randomised, controlled trial. Crit Care 2010;14:R168. profession and trust. JAMA 2001;286:1897-902.
[5] Storli S, Lind R, Viotti IL. Using diaries in intensive care: a method for [27] Bishop CM. Pattern recognition and machine learning. New York:
following up patients. Connect World Crit Care Nurs 2003;2:103-8. Springer; 2006.