Feasibility of an Intervention Study to Support Families When Their Loved One Has Life Sustaining Therapy Withdrawn

Abstract

This investigation addressed family member perceptions of preparation for withdrawal of life-sustaining treatment in the ICU. These families are at high risk for psychosocial and physical sequelae. The quantitative results of this mixed methods study are reported. A control group received usual care and in an educational booklet component of the intervention. The experimental group received the above plus exposure to comfort cart items and additional psychological support. Twenty-eight family members enrolled over a 13-month period. Sixty-one percent (10 intervention, 7 control) completed the follow up. Fourteen (82%) family members recalled the booklet. Some family members reported moderate to severe depression (12.5%), anxiety (12.5%), and stress (12.6%). Satisfaction with care (83.7–85.2%) and family member well-being (44.1) were within the norm. SF-36 physical component score was above the norm and the mental component score was below the norm. This study demonstrated feasibility and acceptability of the interventions and follow-up questionnaires when families make the difficult decision to withdraw treatment. Strategies are suggested to strengthen statistical power.

Background

In the United States there are approximately 6.5 million adult ICU admissions per year.¹ Mortality following an ICU admission is 20%, and 60% of these deaths are preceded by decisions to withdraw life-sustaining treatment.² Quality of ICU care at end of life varies broadly in the United States.³ Family members experience significant burden, especially after participating in decisions to withdraw life-sustaining therapy (referred to as withdrawal). ^4,5 In a systematic review of the literature, communication strategies associated with family satisfaction with end-of-life care in the ICU were identified in 14 studies. These strategies included expressions of sympathy, non-abandonment, assurances of comfort and provision of written materials. Additionally, at the time before death increased satisfaction was associated with support for shared decision making, family presence at the time of death and specific care measures such as extubation.⁶

This investigation was designed to address gaps in current critical care practice in which family members perceive a lack of information needed to prepare for withdrawal and a lack of support during the active phase of dying.^7,8,9 It is important to address this gap because these family members are at high risk for negative psychosocial and physical sequelae ^10, For example, family members have reported extremely high levels of stress approaching those experienced in natural disasters and continuing for six months or more.⁹ Supportive interventions can improve individual and family functioning.¹¹

Methods

The objectives of this study were to: 1) develop and refine an educational and psychological support intervention to prepare families for the dying process after withdrawal, and 2) pilot test the acceptability of this intervention in families with a critically ill family member in a medical intensive care unit (ICU) after a deciding to withdraw. We hypothesized that families receiving the intervention would have decreased stress, anxiety, depression, and post-traumatic stress symptoms; better health; improved perception of family well-being; and better satisfaction with ICU care.

Design

This pilot study employed an experimental, randomized two group design. The control group received usual care provided by the ICU team and a copy of an education intervention booklet explaining what to expect during withdrawal, the dying process, and suggestions for family coping. The experimental group received usual care provided by the ICU team, the education intervention booklet, items from a comfort cart, and psychological support from a research nurse specializing in end of life family support. The quantitative results of this mixed methods study are reported in this paper.

Intervention

Johnson’s self-regulation theory guided the foundation for development of the intervention. This theory proposes that individuals respond in ways that are consistent with their understanding of the experience.¹² The theory focuses on interventions that have both educational and psychological components. The representation of an event is key, and individuals can be influenced by informational interventions. In the context of this study, it was recognized that clinicians can influence families’ interpretations of threatening events (the dying process) and their response (stress, anxiety, depression, post-traumatic stress symptoms, etc.) through provision of information and support. Interventions based on self-regulation theory are designed to assist individuals to direct their attention to objective features of the threatening event.

Educational Component of the Intervention

When a decision is made to withdraw, patients and their families are supported by a multi-professional ICU team (usually including nurses, physicians, pastoral care representative, case manager, social worker, and palliative care clinician). Since families often feel unprepared for what to expect during withdrawal, the intervention included an information session with all family members participating in the study.

To guide the session a booklet, “Preparing You for What to Expect” (see addendum for booklet), was developed for this study. The content included recommendations from other studies as reviewed by the research team ^5,13 and from the Hospice and Palliative Nurses Association. Once the research team agreed on the content of the booklet and it was accepted by our multidisciplinary research advisory team, the booklet was piloted. Acceptability and verification of content were established by professional staff and family members who had participated in decisions to withdraw.

In addition, the booklet was evaluated for readability. Almost a third of Americans have limited health literacy.^14,15 Readability is a strong determinant in health literacy. To ensure comprehension of printed material, the recommended readability score of the reading material should be at or below the sixth-grade level.¹⁶ The most commonly used measure is the Flesch-Kincaid formula, but the Simple Measure of Gobbledygook (SMOG) readability formula performs most consistently. In a study of the readability of hospice materials to prepare families for caregiving at the time of death the mean Flesch-Kincaid grade level was 8.95 and the mean SMOG grade level was 11.06. Medical terms accounted for the complexity of the materials.^17,18 Our study team prepared our booklet to avoid complex words and used descriptive, lay terminology instead of medical terms. The study booklet Flesch-Kincade grade level was 3.4 and the Gobbledygook index was 4.1.^17,18 Therefore, the readability of our booklet was between third and fourth grade level, well below the recommended sixth grade level.

Psychological Support Component of the Intervention

After consent was obtained and the booklet was reviewed, families were randomized to control or intervention groups. There was no further contact with the control group until the follow-up interview. Psychological support was provided to the intervention group before, during and after withdrawal. This support from the research nurse included placing a butterfly decal on the window of the patient’s room to alert staff that the patient was transitioning into death and to remind the family of sensitivity to their impending loss; attending to the family; and offering items from the comfort cart.

The comfort cart was arranged in a two-drawer locked and rolling file cabinet labeled “Family Support Cart”. The comfort cart contained items for physical and emotional support for the family. Families were offered tissues and water bottles. With sensitivity to family preferences, materials were available to make keepsake hand molds and secure a lock of hair. In addition, the comfort cart contained blank business-size cards with inspirational quotations. Examples of the quotations are: 1. Those we love don’t go away, they walk beside us every day. 2. Death leaves a heartache no one can heal but love leaves a memory no one can steal. 3. Little by little we let go of loss, but never of love. These comforting words were inserted into a laminated card holder with room to add a photograph of their loved one or a lock of hair. Some families found comfort in music. Therefore, a CD player was available with a diverse collection of music. The comfort cart also contained nourishment and coffee shop gift cards ($5.00) to support family breaks. To support children, the ICU staff recommended Child Life Services. In addition, the Pastoral Care Representative recommended distribution of The Fall of Freddy the Leaf: A Story of Life for All Ages by Leo Buscaglia to provide a basis for discussion of their feelings and grief. This book is appropriate and appeals to children of all ages.

Methods

The study was conducted in an ICU within a large tertiary care urban teaching institution. The study was approved by the Institutional Review Board. The ICU is a 29-bed medical intensive care unit in which approximately 8–12 patients monthly have withdrawal of treatment. Family members were eligible if they self-identified as a family member, were ≥ 18 years old, and were able to speak and understand English. Family members were randomized according to block randomization design with a block size of 10. Within each block, half of the patients were assigned to control and the other half to intervention. The purpose of block design was to randomize subjects into equal groups to balance group assignment over time, thus reducing the opportunities for bias and confounding in experimental design.¹⁹ We anticipated recruiting 40 family members based upon an expected attrition rate of 25% yielding a final sample size 30.

Prior to data collection, the study was explained to the ICU staff and note cards were distributed with a summary of the study details and research nurse contact information. The research nurse attended daily collaborative care meetings to screen for potential study subjects. After a decision by the healthcare team and family to withdraw, a member of the ICU staff asked the family for permission for the research nurse to contact them and explain the study. If a family agreed, the research nurse met with the family to explain the study and obtain signed consent per Investigational Review Board approved protocol. After randomization, each family member completed a demographic questionnaire and received the booklet with an explanation. Each session was audio-recorded. The session then ended for the control group. Family members randomized to the experimental group received additional psychological support and comfort cart items at the time of withdrawal and during the dying process. All family members received usual care provided by the unit staff.

Outcome Measures

Approximately one month after the patient’s death, a second research nurse who had not been involved in the consent process, explanation of the booklet or the extended intervention interviewed each family member and completed the study measures. Using a semi-structured interview guide, family members were asked if they found the booklet helpful and if they felt prepared and supported before and during the dying process of their loved one. After the interview was completed each family member completed the study measures. Measures included: Depression, Anxiety, Stress Scales,^20,21 Family Satisfaction with Care in the ICU-24,^22–25 Family Member Well- Being questionnaire,²⁶ Impact of Events Scale-Revised,^27–30 and Health Survey Short Form-36,^31–33, and recall of the booklet. Each measure has established reliability and validity and all were administered within a 5 to 10-minute time-period. ^20–33

Data Analysis

Two-hundred and ninety ICU patients were screened over a 13-month period. Fifty-four patient families were asked to participate and 22 patients (40.7%) with 29 family members signed consent to participate in the study. Of the 29 family members, 12 were randomized to control (41.4%) and 17 were randomized to the intervention (58.6%). Seven (7) family members from the control group and ten (10) family members from the intervention group completed the follow-up interview. Of all the family member participants, 63% completed the follow up and there was no difference between groups.

The 29 family members mean age was 56 years, 72% were female, 55% were White and 45% were African American, 58% were married and 42% were single, divorced or other. Of the 22 patients were 59% were male, 50% were White and 50% were African American. The primary ICU diagnoses were hemoptysis, post-cardiac arrest, septic shock, Acute Respiratory Distress Syndrome, pneumonia, chronic obstructive pulmonary disease and liver failure. When families decided to withdraw, about two thirds of the patients were removed from the ventilator, one quarter of the patients were extubated and one patient was withdrawn from continuous renal replacement therapy. After withdrawal, most patients died in the ICU, about one quarter were transferred to hospice, and one patient was transferred to a hospital inpatient bed.

Of the family members, 17 of the 29 accepted the follow up visit and completed the survey: seven (7) family members in the control group (41.2%) and 10 family members from the intervention group (58.8%). The survey included Depression, Anxiety, Stress Scales; ^22–24 Family Satisfaction with Care in the Intensive Care Unit-24; ^25,26 Family Member Well- Being questionnaire;²⁷ Impact of Events Scale-Revised; ^28–30 and Health Survey Short Form-36; ^31–33 educational booklet recall information on the what to expect during withdrawal for all family members; and recall information of the comfort cart intervention for those randomized to the intervention arm of the study. Of the family members who completed the follow-up surveys, 72% were female and 55% were white. The attrition rate was 41% which was higher than the 25% expected. Because the number of family members at follow-up was lower than expected statistical analysis could not be performed to determine significant differences between groups, but tables with preliminary results are presented to demonstrate study feasibility. At the follow-up, some family members reported moderate to severe depression (12.5%), anxiety (12.5%), and stress (12.6%). Satisfaction with care scores were at the norm in both groups (83.7–85.2%). Family member well-being was within the normal range (44.1). Family members reported more intrusion (mean=1.95) then avoidance (mean =1.0) or hyperarousal (mean=.81) as the components of post-traumatic stress syndrome. Physical health quality of life was above the norm and mental health quality of life was below the norm. (See Tables 1 and 2).

Table 1.

Descriptive Measures of Family Members Completing Follow Up Survey

			Intervention (n=10)	Control (n=7)
	Mean (SD)	Range	Mean (SD)	Mean (SD)	p-value
Depression	10.06(10.38)	(0.00–33.00)	12.22(12.73)	7.29(6.13)	0.363
Anxiety	5.13(7.5)	(0.00–28.00)	4.13(5.14)	6.28(9.87)	0.597
Stress	10.69(9.19)	(0.00–35.00)	11.11(10.30)	10.14(8.32)	0.843
Overall Satisfaction	84.81(12.92)	(48.91–100)	84.37(15.13)	85.36(10.55)	0.886
Satisfaction with Care	85.24(12.26)	(55.36–100)	83.16(13.69)	87.90(10.53)	0.462
Satisfaction with Decision Making	83.72(16.13)	(38.89–100)	84.39(18.9)	82.85(13.10)	0.182
FMWB	44.12(13.46)	(27.00–74.00)	43.0(13.03)	45.71(14.9)	0.696
Avoidance IES-R	1.10(0.71)	(0.13–3.13)	1.16(0.79)	1.02(0.62)	0.693
Intrusion IES-R	1.95(1.05)	(0.25–3.88)	1.97(1.10)	1.91(1.06)	0.906
Hyperarousal ISE-R	0.81(0.76)	(0.00–2.67)	0.83(0.71)	0.78(0.88)	0.903
SF-36 PCS	51.78(8.87 )	(37.68–65.16)	54.13(7.43)	48.77(10.20)	0.242
SF-36 MCS	43.50(11.13)	(25.0–58.31)	41.90(11.10)	45.55(11.68)	0.533

p value from independent t-test

Table 2.

Depression Anxiety and Stress Scale of Family Members Completing Follow Up Survey

	Categorized measures, n(%)
Measures	Minimum	Minor	Mild	Moderate	Severe
Depression	10(62.5)	1(6.3)	3(18.8)	0(0)	2(12.5)
Anxiety	13(86.7)	0(0)	0(0)	1(6.7)	1(6.7)
Stress	13(81.3)	1(6.3)	0(0)	1(6.3)	1(6.3)

Fourteen (82%) of the 17 family members recalled review of the booklet explaining what to expect during withdrawal and the dying process. Those who found the booklet unhelpful said: “It was not particularly helpful because (patient) was in a vegetative state,” and “I couldn’t absorb the information at the time.” Those who found the booklet helpful said: “…we used part of it for the prayer book” at the funeral, and “There were good reminders.” “I was so discombobulated. It helped to have things explained.”

The 10 family members who were randomized to the intervention group and completed the follow up survey were asked if they remembered the comfort cart interventions. Not every family member had chosen each of the items offered from the comfort cart. Two of 9 family members (22%) appreciated the butterfly decal placement on the ICU patient’s door signifying the patient’s transition from life to death. Six of 10 (60%) family members remembered receiving bottles of water. Six (6) of 9 (67%) family members remembered receiving tissues. Three families (n=3, 30%) had accepted crackers and remembered that. Opportunities to distribute food coupons for the coffee shop were scarce, but the 2 (20%) family members who received them were grateful. One of these family members was diabetic and had not eaten before the withdrawal. She was very relieved to receive the coupon and encouragement to take care of herself while supporting her dying relative. Mementos were offered to families. Four of 5 family members (80%) remembered receiving a lock of the patient’s hair and 6 of 8 family members (75%) remembered receiving hand molds with some requesting more than one. One family member said: “I love it” and another family member reported placing it in a covered box in her dining room and embracing her husband’s “hand” every morning. Four of 7 family members (57%) remembered receiving a card with a comforting verse to carry with them. Two of 4 (50%) families who received the children’s book discussing dying reported sharing this book with children in their family. One family member (10%) requested music during the dying process and recalled this positively. (See Table 3).

Table 3.

Comfort Cart Items Offered and Recalled

Descriptive of Intervention	Accepted	Recalled	% Recall
Booklet describing dying process reviewed with Family	17	14	82%
Family offered bottled water	10	6	60%
Family offered cracker packet	10	3	30%
Family offered tissue packets	9	6	67%
Butterfly was placed on patient door	9	2	22%
Family offered a hand mold	8	6	75%
Family offered card with comforting verse	7	4	57%
Family offered a lock of hair	4	5	80%
Family offered a food coupon	2	2	100%
Family offered Freddie the Leaf book for a child	2	4	50%
Family offered music	1	1	100%

Discussion

The results of this study support the feasibility of communication and comfort cart interventions to support family members who have made the difficult decision to withdraw life sustaining treatment and are in the process of attending their loved one’s death. Studying families at the time of death of a loved one presents unique challenges related to informed consent, data collection and ethical considerations. ³⁴ Our team, in addition to other researchers, concluded that prospective research on a critically ill dying patient and involvement of family members could be conducted in an ethical manner. Dotolo, Nielsen, et. al. ³⁵ explored barriers and facilitators influencing family members decision to participate in end of life research in the ICU. They found altruism, the research experience and the prospect of enhanced resources were the major factors in the decision to participate in the research. Incorporating these factors as family members are approached for consent is important.

Recruitment and attrition were major limitations in this study. In a systematic review of post-traumatic stress symptoms in post-ICU family members, Petrinec and Daly³⁶ reported consent refusal rates are not always published but those reported ranges from 5% to 38%. Refusal rate in this study was almost 41% which is in high. A multicenter research design over a longer time period would support recruitment of a larger cohort to adequately power the study. For example, Curtis, Treece, et al.¹⁰ screened patients over a 5-year period to randomize 168 patients with 268 family members for an intervention study utilizing communication facilitators to reduce family distress in ICU end of life care. Cooperative research groups, such as the Palliative Care Research Cooperative, are also an important resource for increasing sample size and study sites.³⁷

Communication remains an important strategy for supporting family members in end of life care. The booklet explaining the dying process was well received. The communication strategy in this study was supported by a French study. Lautrette, Darmon, Megarbane, et.al.³⁸ conducted a randomized controlled study of 126 family members (n=63 intervention and n=63 control) in 22 ICUs the efficacy of a proactive end of life conference and a bereavement brochure was evaluated. Fifty-six (89%) of the intervention group and fifty-two (83%) of the control group completed the follow up interview. The intervention group had significantly lower Impact of Event scores and Hospital Anxiety and Depression Scores. Families who completed follow-up generally reported low levels of anxiety, depression, and post-traumatic stress symptoms. About 10% of the sample, however, described severe symptoms.

In other studies, family members who decided to withdraw treatment and witnessed the death of their loved on in an ICU experienced depression, anxiety, and stress as well as some elements of post-traumatic stress syndrome ^{4, 11} Glick, Motta, Weigand, et.al.³⁹ studied family members making decisions for their loved ones’ end-of-life care needs in the ICU experienced anxiety and depression along with anticipatory grief. White, Angus, Shields, et.al. ⁴¹ conducted a randomized clinical trial in which family support interventions were offered by ICU staff to critically ill ICU patients and their surrogate decision-makers. The intervention did not significantly affect the patients or the psychological symptom burden of the surrogates, but the surrogates’ ratings of quality of communication and family-centered care were better in the intervention group. These studies suggest family perceptions of caring, information and support at end-of-life are most important yet challenging to enhance¹³

This study was limited in the complex coordination needed with both ICU staff and family members at the time of withdrawal. To elicit support from the ICU clinicians, the research nurses recruited collaboratively with social work, pastoral care, and nursing staff. This at times delayed approaches to available family members. The large nursing staff was informed about the study, reminders kept them engaged in recruiting opportunities, and ICU nurses often contacted research nurses regarding family decisions on impending withdrawal. Physicians changed frequently but were receptive and supportive of the research team. We encountered challenges at the clinician and family level. Timing of recruitment was an important challenge. Although the research nurses made efforts to provide coverage every weekday, there were lapses in availability especially when decisions to withdraw were made in the evening or with no prior indication that the decision was pending.

This study supported the use of printed material and comforting items to assist family members when they have made the difficult decision to withdraw treatment and need preparation to prepare for the impending death of a loved one. It also supported the willingness of family members to recall their experience and complete questionnaires to inform health care workers about the effectiveness of these interventions. Despite a close connection to the study nurses at the time of withdrawal though, there was a high attrition rate.

The booklet explaining the dying process was useful. It is available for clinical use, but outcome measurement is highly recommended to support evidence-based practice. The comfort cart intervention was supported by professional experience and anecdotal accounts in informal nursing publications.⁴⁰ As in our study, comfort carts include a mechanism to communicate to staff that a patient is dying, food, music, memento making materials, and child support items. More recent carts include rubbing stones etched with words like family, peace, or love. The comfort carts have been supported by donations but could be hospital supported in the future. Study of family member outcome measures is warranted to determine the effects of the comfort cart.

This study design or elements of this study may be replicated for nursing research or a Continuous Quality Improvement project. For example, family satisfaction with care and decision making could be measured using the Family Satisfaction with Care-24 instrument before and after introducing the booklet to determine if family satisfaction improved. The same could be done with the introduction of the comfort cart. Funding projects is often challenging. Accessing hospital resources supporting clinical research, partnering with schools of nursing and teaching staff the grant application process may enhance staff career development.

Figure 1.

Education intervention booklet: “Let us help you Prepare”