A cross-sectional descriptive study was performed to evaluate resuscitation decisions and factors that impact these choices for young people with severe developmental disabilities residing in a skilled nursing facility. Decision-makers were provided with information to clarify resuscitation preferences. Parents/guardians of 30 of the 67 residents also completed a survey. A significant number of decision-makers changed their resuscitation preference to DNR after detailed explanations were provided. Survey results suggest that interpersonal relationships, such as those with family members, religious leader, and physician, were more influential for families who chose full resuscitation compared to those with DNR preferences. Factors such as perception of quality of life and medical condition of the individuals with developmental disabilities were not significantly different between these two groups.
There currently is much discussion regarding end-of-life decisions in the scientific community and lay press. A report on palliative and end-of-life care by the Institute of Medicine indicated many areas in need of further study (Committee on Palliative and End-of-Life Care, 2003). There is much to be learned about how best to provide compassionate and comprehensive patient care while also addressing the needs of families and caregivers. Many important decisions are made in anticipation of death, some of which require a full understanding of options of care.
Life expectancy for people with intellectual disabilities presently has been noted to be comparable to the general population (Botsford, 2004). Individuals with severe to profound intellectual disabilities and complex medical problems, however, continue to live shorter lives albeit improved from previous years (Chaney & Eyman, 2000; Strauss, Shavelle, & Anderson, 1998). Parents and guardians are sometimes faced with difficult decisions regarding aggressive medical treatments, palliative care, and change in resuscitation status to do-not-resuscitate—DNR (Acckermann & Kemle, 1999; Gordon & Sokolowski, 2002; Murphy et al., 1994). Factors that influence resuscitation decisions for children, adolescents, and young adults with these very significant developmental and medical issues have not been clearly defined.
There has been much debate regarding who should make care plan decisions when an individual is not competent to do so (Council on Ethics and Judicial Affairs, 1991; Kunin, 1997; Lawhorne, 1999). End-of-life issues for people with significant neurological impairment have recently gained public attention (Bloche, 2005; deBeaufort, 2005; Roig-Franzia, 2005). However, these decisions usually are made between families and medical staff, with individual consideration for each person. The hope is that the decision-maker acts in the best interest and adheres to the presumed wishes of his or her loved one.
Decision-making may be more difficult when the patient has never been able to express his or her desires and has not lived well into adulthood. As has been seen with adults and families of adult patients who are not competent to make informed decisions, it is unclear if families completely comprehend the issues that must be understood in order to make informed decisions (Nelson, 2003; Nelson & Nelson, 1992; Shmerling, Bedell, Lilienfeld, & Delbanco, 1988). When parents and guardians of medically fragile young people with severe developmental disabilities were provided with explanatory information about resuscitation prior to an acute life-threatening event, there was a significant increase in those who requested a DNR order for their child (Friedman, 2006). Factors such as religion and family members' influence also have been noted to impact some end-of-life decisions. Nelson and Nelson (1992) suggested that the opposition someone may have to a child foregoing aggressive treatment when death is imminent is directly proportional to the geographical, psychological, and personal distance from the child.
We designed this study in order to identify factors that influence parents' and guardians' end-of-life decisions. We felt that a better understanding of contributing factors to this important issue potentially could improve medical professionals' ability to support families and young people in making these important decisions.
We conducted this study in a medium-sized skilled nursing facility (SNF) in a New England state that provided care for young people with severe developmental disabilities and complex medical problems. Although licensed as a pediatric facility, children, adolescents, and young adults also resided there because opportunities to transfer to another facility upon reaching adulthood were usually not available or requested. The residents were all nonambulatory and nonverbal, and the majority of them received feedings through gastrostomy tubes and multiple medical treatments requiring 24-hour skilled nursing care. They participated in educational or adult programming every weekday, except during times of acute illnesses.
In 1997, the home state initiated a Comfort Care program through the Department of Public Health for Emergency Medical Service (EMS) personnel (Department of Public Health, 1999). This program was developed to honor the wishes of patients who wanted resuscitation withheld in the event of cardiopulmonary arrest outside of the hospital or nursing home setting. The patient or guardian would be able to sign a document, along with the treating physician, to indicate these wishes. Emergency medical service workers would, therefore, be obliged to honor DNR status when individuals were in the community setting and had the requisite verification document or wrist bracelet. If this documentation was unavailable, the EMS personnel would initiate medical care that may include cardiac compressions, endotracheal intubation, artificial ventilation, defibrillation, and/or resuscitation medications, as indicated. If an individual had Comfort Care documentation, they would be provided with treatments such as oxygen and pain medications, although not resuscitated in the event that their heart or breathing fails to function.
As a result of this change, and because residents of this SNF often went on field trips and appointments outside the facility, we decided that this program would be implemented for the residents who had DNR orders. In order to accomplish this, families had to be informed about the program and the mandated requirement for documentation with signatures by parents or guardians and physicians. We decided to take this opportunity to review end-of-life decisions with all families and administer a survey regarding potential factors that may affect such decisions.
We developed a 6-page survey with 29 questions and 51 items (available from the first author). There were short answer and multiple-choice questions designed to obtain demographic information. A 4-point Likert scale was used to gather information regarding religious factors, family composition and influence, social support, and physician influence. Likert scale items were also used to measure families' assessment of their child's medical problems, quality of life, and perception of pain. Through the survey we also obtained information regarding resuscitation preference and the stability of this preference over time. There was a cover letter attached to each survey indicating the purpose of the study; the estimated time for completion (15 minutes); and that it was optional, anonymous, and would in no way affect their child's care. Contact telephone numbers were provided to the families if they wished to ask questions or obtain additional clarification. A self-addressed stamped envelope accompanied each survey.
Letters were sent to all families of the young people who resided in the SNF, informing them of a choice of two information meetings that were available for them to attend, either on a weekday evening or Sunday afternoon in April 1998. The purpose of these meetings was to allow us to review end-of-life decisions and resuscitation preferences, clarify the distinctions between resuscitation and different types of treatment, and explain the Comfort Care program for residents who had DNR orders. The letters were followed by phone calls to the families by their child's social worker to confirm attendance to one of the sessions. Parents who attended the informational meetings were given a copy of the survey at the end of the meeting.
Families who did not wish to attend either of the two informational meetings were called by their child's social worker to confirm their child's resuscitation status and asked whether the physician may call them for further discussion or whether they wished to set up an individual meeting with the physician. Families were required to speak to a physician if they wanted to have their child's resuscitation status changed. Decision-makers for individuals with DNR orders were also required to speak to the physician to confirm the order and to review the Comfort Care program. Families were able to indicate to the social worker that they wanted to keep the resuscitation status unchanged (full resuscitation) and did not wish to receive a call or schedule a meeting with the physician. Surveys were sent to families after phone contact by the social worker or the physician.
Similarly, parents who met individually with the physician were provided with a survey at the end of the meeting. The topics covered via individual meetings and per phone contact were the same as those covered at the two informational meetings: (a) definition of resuscitation, (b) the meaning of DNR, (c) clarification of the difference between DNR and providing treatments, (d) situations in which treatments are withheld and care redirected, (e) how this applies to the child, (f) facility policies regarding resuscitation, (g) Comfort Care program, and (h) policy and care implications
Families who were provided with surveys received a follow-up phone call by the social worker to confirm its receipt. If they indicated that they did not plan to complete the survey, they were not contacted again; otherwise, a follow-up call was made 2 weeks later. There were 2 families that the social workers felt should not be sent surveys or contacted by phone by the physician because they had previously indicated that they only wanted to be called by the facility in the event of an emergency. These families, therefore, were not provided with this information. The first author was not aware of which families did not want to complete the survey. There were no identifiers on the surveys that were returned. The change in DNR status was evaluated separately from the survey results because this information required physician involvement and did not affect the anonymity of the surveys.
For all families whose child had a DNR order, its continuation was reviewed and affirmed. The physician explained the Comfort Care program and its requirements, and the parents/guardians and physician completed the needed forms so it could be implemented immediately. If a family contacted via phone wished to change the resuscitation order, the conversation was witnessed via speakerphone by a nurse or social worker and documented in the chart. The families then needed to return signed forms indicating their desire and permission for DNR and the Comfort Care program. These forms were returned separately from the survey.
This protocol was approved by the conducting facility's Institutional Review Board, with the determination that full informed consent was not required because it was a voluntary, anonymous survey.
Means and SDs were calculated for continuous variables. Frequency distributions and proportions were analyzed for discrete variables. Chi-square tests were run for all pre- and posttest analyses. Because of the small sample size, adjustments to the data were necessary. We converted the Likert scales to binomial variables in order to reduce the chance of low expected cell counts for chi-square tests. On the Likert scales, the responses frequently and occasionally were grouped together as were rarely and never. Similarly, highly and moderately were combined as were slightly and not at all. Fisher's exact significance was used when there were low expected cell counts. Because of the exploratory nature and small sample size, we used an alpha of 0.1 to avoid a Type II error.
At the initiation of this study, there were 67 residents at the SNF; 65 were provided with surveys. Most of the residents were non-Hispanic Caucasians; there were 2 African American, 8 Hispanic, and 4 Asian individuals. Twenty residents had DNR orders and 47 did not. Parents or guardians of 16 residents attended one of the two information sessions. Six children of the families in attendance already had DNR orders. The children of 10 of the families at these meetings were in the original full resuscitation group; 4 of them were later changed to DNR. Four families did not attend either of the meetings, but met individually with the physician. Two of these families continued to request full resuscitation efforts in the event of cardiopulmonary arrest, and 2 families changed to DNR.
Ten families contacted by the social workers did not want to change the resuscitation status from full code, did not wish to meet or speak with the physician, and/or had no further questions. Sixteen families spoke to the social worker and expressed the wish to also speak to the physician by phone. Six of those families decided to change from full resuscitation to DNR. The social work staff was able to contact all but 5 families who did not respond to the letter sent in the mail regarding the meeting, were not reachable by phone, and/or did not return phone messages to determine their intent to attend the meeting or wish for individual or phone contact with the physician. They also were sent surveys for completion. All of the families whose child had a DNR order also provided written consent for participation in the Comfort Care program, which was co-signed by the physician. After the forms were completed, the resident wore a bracelet bearing these signatures that would potentially alert the DNR status to paramedics in the community setting.
Of the original group, 12 changed from full resuscitation to DNR when parents were provided with additional information regarding resuscitation and end-of-life issues, resulting in a total of 32 individuals with DNR orders, χ2(1, N = 67) = 31.18, p < .001. Of those individuals whose family changed their resuscitation preference to DNR, 2 died and 2 were transferred to other facilities over the course of the following year. There was no family who changed from DNR to full resuscitation.
Thirty of the 65 surveys provided to families were returned after completion, representing a 46% response rate. The mean age of individuals in both of these groups was 19 years, with a range of 2.5 to 36 years for the entire group, and 2.5 to 34 years for the children of survey respondents. Of the surveys returned, 21 were from families whose child had a DNR order at the time the survey was completed. Nine of the surveys came from families who desired full resuscitation for their child. This difference in distribution of resuscitation preferences between the study group and the entire facility was significant, χ2(1, N = 67) = 10.77, p < .001.
Demographic factors of the survey respondents and their child were not significantly different for the full resuscitation and DNR groups. There were no significant differences between the two groups in frequency of attendance in religious services, religious preference, proximity of family members, frequency of family functions, or frequency of visits to the child. There were no significant differences between full resuscitation and DNR groups in their perceptions of medical condition, disability, quality of life, or perception of pain or discomfort, which are factors that influence resuscitation decisions. (See Table 1 for respondent and resident characteristics.)
Factors identified on the survey that influenced resuscitation decisions are noted in Table 2. There were some respondents who did not answer all questions; therefore, the sample size fluctuated across measures. Families who opted for full resuscitation compared to those who requested DNR were more apt to be influenced in making this decision by family members, χ2(1, N = 22) = 4.71, p = .091, and by their religious leader, χ2(1, N = 28) = 4.54, p = .095. Families whose child was designated as full resuscitation also were more apt to be influenced by discussion with, not opinion of, their child's physician than those in the DNR group, χ2(1, N = 27) = 3.86, p = .073. There also was a trend found when comparing families' discussion with physicians regarding end-of-life, with those in the DNR group being less likely to discuss comfort management in all situations, χ2(1, N = 30) = 3.81, p = .062.
Our findings are consistent with a previous study (Friedman, 2006) performed at another skilled nursing facility caring for young people with severe developmental disabilities, in which provision of information regarding the meaning of resuscitation and treatment options resulted in a significant increase in DNR preference. This finding also is consistent with studies in the adult literature, whereby patients want their physician to discuss resuscitation options as part of routine care (Murphy et al., 1998; Shmerling et al., 1988). A significant increase in DNR preference has been demonstrated when adults are provided with this type of information, especially when prognosis for survival is included in the discussion. Prognostic information was not provided to the families in the present study; however, we note that in the past 15 years, there have been only two survivors of resuscitation efforts, and one of those children died within the following 24 hours. Certainly, ethical issues need to be considered when discussing prognostic information about children compared to elderly adults. In addition to prognosis, discomfort resulting from aggressive medical management when there is little hope of survival has been considered in making end-of-life decisions (Burns & Truog, 1997; Byock, 1998; Gordon, 2001; Gordon, 2003).
In our study, interpersonal relationships seemed to be more influential on decision-makers who chose full resuscitation compared to those who requested DNR. We found that religious leader influence, family pressures, and discussion with their physician affected decisions for the group who chose full resuscitation for their child. The religious denomination, per se, was not found to be a significant factor in the choice of DNR compared to full resuscitation. Nelson and Nelson (1992) reported that religion may impact end-of-life decisions of parents, although quantitative support for that assumption was not provided. Others have also noted the adjustment of parents to their children's disabilities, although not specifically end-of-life decisions, are affected by the response and interaction of family members, friends, and health care providers (Kearney & Griffin, 2001; Smith, Oliver, & Innocenti, 2001). Perhaps those who are ambivalent about DNR orders are more apt to discuss this issue with others and then make a choice that ostensibly appears to provide more options for intervention, whereas those with strong personal feelings in favor of withholding resuscitation efforts in the event of cardiopulmonary arrest are able to make that decision without additional discussions. The influence of interpersonal relationships in making such decisions is an area that requires further exploration.
Decision-makers for the DNR group were also less apt to discuss comfort measures being used for their child in all situations when speaking to their child's physician regarding resuscitation. Families who choose full resuscitation may need assurance that despite the interventions that their child may undergo when he or she is critically ill or dying, they will be made comfortable. It had been anticipated that perceptions of quality of life and/or discomfort would affect the resuscitation decision, however that was not found to be the case based on the questionnaire results. There also was a significant difference in survey completion between families whose child had a DNR order compared to those who did not. Families who choose DNR are mandated to provide written and/or witnessed consent, often after lengthy discussions with the physician. These families may have felt more invested in completing a survey dealing with end-of-life issues.
In the year following this study, there was an adenovirus epidemic, increasing the average yearly mortality rate from 2 to 6. Two of the children who died were in the group that changed to DNR; this change was made prior to the outbreak. There were no families who changed to DNR as a result of their child's illness. Of note is that at the beginning of the study, 2 families had assumed their child would not be resuscitated in the event of cardiopulmonary arrest, even though such an order had never been written. The importance of a regular review of DNR orders is evident. Federal regulation for nursing homes currently mandates that resuscitation status be reviewed at regular intervals (Fogg, 1995). We have found that at the minimum annual reviews are important to provide, continuing opportunities to discuss the young person's medical status and decision-makers' preferences. For children living at home with similar medical and developmental needs, there are no regulations or mandates about when to broach the issue of end-of-life care. The medical home model of care potentially can provide a framework to these types of discussions and decisions (American Academy of Pediatrics, 2002; Moore & Tonniges, 2004). The National Center for the Medical Home Initiative for Children provides resources regarding end-of-life and palliative care, although there are no specific recommendations guiding the introduction and review of these topics (American Academy of Pediatrics, 2006). Myths also need to be dispelled regarding exactly what a DNR order entails, and the distinction between this type of order and the provision of treatment. It is apparent that there are different viewpoints among families, physicians, religious leaders, and the community at large about aggressive treatment options or lack thereof (Department of Mental Retardation, Commonwealth of Massachusetts, 1999; Nelson, 2003).
The surveys sampled did not reflect the population of the facility in terms of resuscitation status, limiting the generalizability of the findings. An alpha of 0.1 was also used in analyzing data because of the small numbers and the exploratory nature of this study. We also did not use standardized measures of quality of life but, rather, the families' reported perception. Other potentially contributing factors, such as parental emotional health and marital relationships, were not evaluated and warrant further study. Although socioeconomic status was not considered, all of the residents received Medicaid funding, so families were not burdened with out-of-pocket expenses for their care. Because this study was done in an SNF in which children, adolescents, and young adults were cared for, our findings cannot be generalized to other groups of individuals. Further, although this study was performed a number of years ago, the same resuscitation policies of this nursing facility continue to be in effect. The same physician spoke to each of these families for this study and continues to provide care at this facility. Despite an increase in public and professional awareness of the need to address issues regarding end-of-life care for individuals with disabilities since the time of this study, we have found little change in parents' attitudes and preferences in our facility.
Although definitive demographic distinctions were not identifiable between parents and their children in the full resuscitation versus the no resuscitation groups, there was a suggestion that interpersonal relationships have a greater influence on families who choose full resuscitation. These relationships occur not only with health care providers, but also with other individuals who play a central role in family life, such as religious leaders and family members. Although the majority of families continue to want resuscitation efforts provided in the event of cardiopulmonary arrest, it was evident that when decision-makers are provided with additional information to clarify resuscitation options, there is a significant increase in those who choose DNR for their loved one.
This research was supported in part by Health Resources and Services Administration; the U.S. Department of Health and Human Services; Leadership Education in Neurodevelopmental Disabilities (LEND), Grant MCJ-299150 from the Maternal and Child Health Bureau; and Grant 90DD0357 from the Administration for Children and Families, Administration on Developmental Disabilities. No official endorsements of this research should be inferred. The authors thank Thomas Ash, Nancy Holtzer, and Karen Lowe for their assistance in this study. We appreciate the input of Joanne Wolfe while developing the survey and are also grateful to Elizabeth R. Woods and Allen C. Crocker for their review of this paper.
Sandra L. Friedman, MD (Sandra.email@example.com), Director, Neurodevelopmental Disabilities Program, Children's Hospital Boston, Division of General Pediatrics, 300 Longwood Ave., Boston, MA 02115. Dana Gilmore, MA, Research Associate, University of Massachusetts Boston. Both authors are affiliated with the Institute for Community Inclusion, and the first author is also affiliated with Harvard Medical School. Correspondence should be sent to the first author