Creating Opportunities for Parent Empowerment: Program Effects on the Mental Health/Coping Outcomes of Critically Ill Young Children and Their Mothers
Objective. Increasing numbers of children in the United States (ie, ∼200 children per 100 000 population) require intensive care annually, because of advances in pediatric therapeutic techniques and a changing spectrum of pediatric disease. These children are especially vulnerable to a multitude of short- and long-term negative emotional, behavioral, and academic outcomes, including a higher risk of posttraumatic stress disorder (PTSD) and a greater need for psychiatric treatment, compared with matched hospitalized children who do not require intensive care. In addition, the parents of these children are at risk for the development of PTSD, as well as other negative emotional outcomes (eg, depression and anxiety disorders). There has been little research conducted to systematically determine the effects of interventions aimed at improving psychosocial outcomes for critically ill children and their parents, despite recognition of the adverse effects of critical care hospitalization on the nonphysiologic well-being of patients and their families. The purpose of this study was to evaluate the effects of a preventive educational-behavioral intervention program, the Creating Opportunities for Parent Empowerment (COPE) program, initiated early in the intensive care unit hospitalization on the mental health/psychosocial outcomes of critically ill young children and their mothers.
Design. A randomized, controlled trial with follow-up assessments 1, 3, 6, and 12 months after hospitalization was conducted with 174 mothers and their 2- to 7-year-old children who were unexpectedly hospitalized in the pediatric intensive care units (PICUs) of 2 children’s hospitals. The final sample of 163 mothers ranged in age from 18 to 52 years, with a mean of 31.2 years. Among the mothers reporting race/ethnicity, the sample included 116 white (71.2%), 33 African American (20.3%), 3 Hispanic (1.8%), and 2 Indian (1.2%) mothers. The mean age of the hospitalized children was 50.3 months. Ninety-nine children (60.7%) were male and 64 (39.3%) were female. The major reasons for hospitalization were respiratory problems, accidental trauma, neurologic problems, and infections. Fifty-seven percent (n = 93) of the children had never been hospitalized overnight, and none had experienced a previous PICU hospitalization.
Interventions. Mothers in the experimental (COPE) group received a 3-phase educational-behavioral intervention program 1) 6 to 16 hours after PICU admission, 2) 2 to 16 hours after transfer to the general pediatric unit, and 3) 2 to 3 days after their children were discharged from the hospital. Control mothers received a structurally equivalent control program. The COPE intervention was based on self-regulation theory, control theory, and the emotional contagion hypothesis. The COPE program, which was delivered with audiotapes and matching written information, as well as a parent-child activity workbook that facilitated implementing the audiotaped information, focused on increasing 1) parents’ knowledge and understanding of the range of behaviors and emotions that young children typically display during and after hospitalization and 2) direct parent participation in their children’s emotional and physical care. The COPE workbook, which was provided to parents and children after transfer from the PICU to the general pediatric unit, contained 3 activities to be completed before discharge from the hospital, ie, 1) puppet play to encourage expression of emotions in a nonthreatening manner, 2) therapeutic medical play to assist children in obtaining some sense of mastery and control over the hospital experience, and 3) reading and discussing Jenny’s Wish, a story about a young child who successfully copes with a stressful hospitalization.
Outcome Measures. Primary outcomes included maternal anxiety, negative mood state, depression, maternal beliefs, parental stress, and parent participation in their children’s care, as well as child adjustment, which was assessed with the Behavioral Assessment System for Children (parent form).
Results. COPE mothers reported significantly less parental stress and participated more in their children’s physical and emotional care on the pediatric unit, compared with control mothers, as rated by nurses who were blinded with respect to study group. In comparison with control mothers, COPE mothers reported less negative mood state, less depression, and fewer PTSD symptoms at certain follow-up assessments after hospitalization. In addition, COPE mothers reported stronger beliefs regarding their children’s likely responses to hospitalization and how they could enhance their children’s adjustment, compared with control mothers. COPE children, in comparison with control children, exhibited significantly fewer withdrawal symptoms 6 months after discharge, as well as fewer negative behavioral symptoms and externalizing behaviors at 12 months. COPE mothers also reported less hyperactivity and greater adaptability among their children at 12 months, compared with control mothers. One year after discharge, a significantly higher percentage of control group children (25.9%) exhibited clinically significant behavioral symptoms, compared with COPE children (2.3%). In addition, 6 and 12 months after discharge, significantly higher percentages of control group children exhibited clinically significant externalizing symptoms (6 months, 14.3%; 12 months, 22.2%), compared with COPE children (6 months, 1.8%; 12 months, 4.5%).
Conclusions. The findings of this study indicated that mothers who received the COPE program experienced improved maternal functional and emotional coping outcomes, which resulted in significantly fewer child adjustment problems, in comparison with the control group. With the increasing prevalence of attention-deficit/hyperactivity disorder and externalizing problems among children and the documented lack of mental health screening and early intervention services for children in this country, the COPE intervention could help protect this high-risk population of children from developing these troublesome problems. As a result, the mental health status of children after critical care hospitalization could be improved. With routine provision of the COPE program in PICUs throughout the country, family burdens and costs associated with the mental health treatment of these problems might be substantially reduced.
Increasing numbers of children in the United States (ie, ∼200 children per 100 000 population) require intensive care annually, because of advances in pediatric therapeutic methods and a changing spectrum of pediatric disease.1 It has been projected that there will be continued growth in the number of pediatric intensive care unit (PICU) beds well into the new millennium, with a higher level of illness acuity for children occupying them.2
Critically ill children hospitalized in intensive care units (ICUs) are especially vulnerable to a multitude of short- and long-term, negative emotional, behavioral, and academic outcomes, including a higher risk of posttraumatic stress disorder (PTSD) and a greater need for psychiatric treatment, compared with matched hospitalized children who do not require intensive care.3 In addition, the parents of these children are at risk for the development of PTSD, as well as other negative emotional outcomes (eg, depression and anxiety disorders).4–6
Little research has been conducted to determine systematically the effects of interventions aimed at improving psychosocial outcomes for critically ill children and their parents, despite recognition of the adverse effects of critical care hospitalization on the nonphysiologic well-being of these patients and their families. Although numerous studies have demonstrated the positive effects of including parents in the psychologic preparation of children for planned minor operations and procedures,7–9 it is empirically unknown whether similar approaches initiated early in the ICU hospitalization would benefit critically ill children and their parents. Therefore, the urgent need to test interventions aimed at improving the mental health and coping outcomes for this high-risk population is immediately evident.
Our pilot study demonstrated the positive effects of an educational-behavioral intervention program titled Creating Opportunities for Parent Empowerment (COPE) on short-term coping outcomes for 30 mothers and their 1- to 6-year-old, critically ill children, up to 1 month after hospitalization.6 Because short-term findings from our pilot project were promising, this study was undertaken with the primary aims of determining whether the outcomes could be replicated within the context of a full-scale, randomized, controlled trial and evaluating whether the COPE program could have sustained positive effects on the process and outcomes of maternal and child coping with critical illness for up to 1 year after hospital discharge.
Design of the COPE intervention was driven by self-regulation theory,10,11 control theory,12,13 and the emotional contagion hypothesis.14–16 On the basis of self-regulation and control theories, we hypothesized that the COPE intervention would strengthen mothers’ beliefs and knowledge of the typical behaviors and emotions to expect from their young children as they recovered from a critical illness and would remove barriers to enhance their ability to facilitate their children’s adjustment. COPE mothers were expected to experience less negative mood and to support their children more effectively during and after hospitalization, compared with mothers who received the control program. On the basis of the emotional contagion hypothesis, we thought that effective parental coping among the COPE mothers would lead to better adjustment outcomes for their children.
A randomized controlled trial was designed in which subjects at each of 2 study sites were randomly assigned, after providing informed consent, to receive either the COPE program or a control program, in 1-week blocks of time to decrease the probability of staff-to-parent and parent-to-parent contamination (ie, the likelihood that parents in the PICU who were in different study groups would exchange content that was provided to them in their experimental program). Both groups received an educational-behavioral intervention program (ie, COPE or control), which was delivered in 3 phases, as follows: phase I, within 6 to 16 hours after the child was admitted to the PICU; phase II, within 2 to 16 hours after the child was transferred from the PICU to the general pediatric unit; phase III, within 2 to 3 days after the child was discharged from the hospital. Follow-up assessments were completed 1, 3, 6, and 12 months after discharge. Procedures were followed in accordance with the ethical standards of the institutional research subject review boards and the Helsinki Declaration of 1975, as revised in 1983. Enrollment and follow-up monitoring of participants occurred from January 1998 through March 2002.
All mothers who could read and speak English with children admitted to either of the 2 PICU study sites (ie, a 12-bed PICU at a 124-bed children’s hospital within a 720-bed academic medical center in upstate New York or a 31-bed PICU at a 235-bed children’s hospital in southwestern Pennsylvania) were eligible to participate if their children 1) had an unplanned medical or surgical admission to the PICU, 2) were between 2 and 7 years of age, 3) were expected to survive, 4) had no prior ICU admissions, 5) had no cancer, and 6) had no suspected or diagnosed physical or sexual abuse. Mothers were excluded from data analysis if 1) their children were readmitted to the PICU after transfer from the PICU to the general pediatric unit, 2) their children were hospitalized in the PICU for >21 days, or 3) they made a personal decision to withdraw from the study. We chose to focus solely on mothers to eliminate the possibility of substantial variability in the data, because evidence has indicated that there are marked differences in the responses of fathers and mothers to childhood critical illness.17 Of 212 eligible mothers, 38 (18%) refused, with the majority stating that they were too stressed or too tired to participate (n = 17) or that they wanted to focus only on their children (n = 7).
The convenience sample was composed of 174 mothers (90 in the COPE group and 84 in the control group). After enrollment and randomization to a study group, data for 3 COPE mothers were eliminated, because of a prior admission to the PICU (n = 1), readmission to the PICU before discharge (n = 1), or death after multiple readmissions to the PICU (n = 1). Data for 8 control mothers were eliminated from analyses because of an extended PICU stay of >21 days (30 days) (n = 1), readmission to the PICU before hospital discharge (n = 5), a planned admission (n = 1), or death (n = 1). Therefore, the final sample for data analyses included 163 mothers and their children.
Mothers in the final sample ranged in age from 18 to 52 years, with a mean of 31.2 years (SD: 6.3 years). They had a mean number of 1.7 other children at home (SD: 1.4 children; range: 0–6 children). The sample included 116 white mothers (71.2%), 33 African American mothers (20.3%), 3 Hispanic mothers (1.8%), and 2 Indian mothers (1.2%). Two mothers (1.2%) reported that they were of other races, and 7 (4.3%) did not report their ethnic groups. One hundred thirty-three mothers (81.6%) had completed high school. Seventy-eight (47.8%) were in their first marriages, 36 (22.1%) had never married, 5 (3.1%) were separated from their husbands, 12 (7.4%) were divorced, 21 (12.9%) were in their second marriages, and 11 (6.7%) did not report their marital status. The household income for the majority of families (n = 102, 62.6%) was at least $15 000/year (see Table 1 for baseline characteristics of the mothers in both study groups).
The mean age of the hospitalized children was 50.3 months (SD: 18.9 months; range: 22–88 months). Ninety-nine children were male (60.7%) and 64 (39.3%) were female. The major reasons for hospitalization were respiratory problems (eg, asthma or pneumonia) (n = 71, 43.6%), accidental trauma (n = 26, 16%), neurologic problems (eg, seizures or accidents) (n = 22, 13.5%), infections (eg, meningitis or sepsis) (n = 18, 11%), hematologic problems (eg, bleeding after procedures) (n = 8, 4.9%), cardiac problems (n = 3, 1.9%), ingestions (n = 7, 4.3%), or other causes (eg, acidosis) (n = 8, 4.9%). Fifty-seven percent (n = 93) of the children had never been hospitalized overnight. Pediatric Risk of Mortality scores18 averaged 4.4 (SD: 4.9; range: 0–31) on a scale of 0 to 76, with 0 representing no risk of death. The length of stay in the PICU averaged 64.3 hours (SD: 64.3 hours; range: 10.0–440.0 hours), and the total length of hospital stay averaged 6.9 days (SD: 6.3 days; range: 1–32 days) (see Table 2 for baseline characteristics of children in both study groups).
The COPE Program
The experimental program focused on increasing 1) parents’ knowledge and understanding of the range of behaviors and emotions that young children typically display during and after hospitalization and 2) direct parent participation in their children’s emotional and physical care. Phase I of the intervention in the PICU consisted of audiotaped information and matching written information.
Phase II of the COPE intervention, a “booster” intervention that occurred shortly after transfer from the PICU to the general pediatric unit, consisted of 1) audiotaped and written information that reinforced critical content of the initial audiotape and provided additional information on children’s responses during and after hospitalization, as well as providing mothers with additional suggestions to enhance coping outcomes for their children, and 2) a parent-child activity workbook. The workbook contained 3 activities to be completed before discharge from the hospital, ie, 1) puppet play to encourage expression of emotions in a nonthreatening manner, 2) therapeutic medical play to assist children in obtaining some sense of mastery and control over the hospital experience, and 3) reading and discussing Jenny’s Wish, a story about a young child who successfully copes with a stressful hospitalization.
Phase III of the COPE intervention program, the second booster intervention, occurred 2 to 3 days after hospital discharge and consisted of a telephone call, during which a 5-minute script that reinforced the following was read: 1) young children’s typical postdischarge emotions and behaviors and 2) parenting behaviors that would continue to facilitate positive coping outcomes for the children. Mothers were encouraged to continue to perform the activities from the workbook they had received during hospitalization. Immediately after the telephone intervention, the telephone script was mailed to the mothers.
The Control Program
The structurally equivalent control program also consisted of 3 phases, which occurred at the same times as the COPE program phases. Phase I provided audiotaped and written information on the services and policies of the PICU. Phase II consisted of 1) audiotaped and written information about the general pediatric unit and its policies and 2) a parent-child activity workbook containing “control” activities (eg, reading a non-hospital-related book and working with modeling compound). Phase III of the control program consisted of a telephone call made 2 to 3 days after discharge, during which mothers were informed that they should contact their primary health care providers if their children were having any problems or unusual symptoms. During the telephone call, the mothers were asked to comment on their children’s hospital stays.
Instruments with established construct validity and reliability were used to measure the primary outcomes of the study. These instruments, with their Cronbach α values for this sample, are summarized in Table 3. Higher scores on the State Anxiety Inventory,19 the Profile of Mood States,20 and the Parental Stressor Scale: PICU21 are indicative of greater state anxiety, negative mood state, and parental stress.
The Post-Hospital Stress Index for Parents and the Post-Hospital Stress Index for Children were based on criteria for the diagnosis of PTSD, as published in the Diagnostic and Statistical Manual of Mental Disorders (4th ed).4 Parents completed these two 20-item questionnaires regarding their own symptoms and their children’s symptoms by checking either the “yes” or “no” column in response to each of the questions (eg, “Do you experience frequent distressing memories of the hospitalization?”). Content validity was established by 3 psychiatrists and 3 psychologists. Scoring consisted of totaling the number of positive responses; therefore, total scores could range from 0 to 20. Subjects reporting ≥7 PTSD symptoms were referred to their primary care physicians for follow-up care.
The mothers’ involvement in their children’s care was assessed with the Index of Parent Participation,22 a 36-item instrument describing parenting behaviors in which mothers could engage during their children’s hospitalizations (eg, bathing the child or explaining a test or procedure). Scores were computed by summing the parenting behaviors, with a higher score indicating more parent participation. The nurses completed this measure while the children were in the PICU, and the mothers completed it while their children were in the general pediatric unit. In addition, the children’s primary nurses, who were blinded with respect to the study groups, were asked to make hatch marks on 2 visual analog scales (10 cm), indicating the extent to which the children’s mothers participated in the physical and emotional care of their children in the PICU and in the general pediatric unit. Higher scores on each of these scales indicated greater involvement by mothers in their children’s care.
Children’s adjustment was measured by using the Behavioral Assessment System for Children (BASC) (parent form).23 The BASC is a comprehensive measure of both adaptive and problem behaviors. Mothers rated their children, during the past 6 months, for each of the 131 items on a 4-point scale, ranging from 1 (never) to 4 (always). Composite subscales included externalizing problems, internalizing problems, behavioral symptoms, and adaptive skills.
The Parental Belief Scale for Parents of Hospitalized Children24 was used to measure parental beliefs about their hospitalized children’s responses and their role during hospitalization. Twenty items assessed parental beliefs about the hospitalized child’s responses (eg, “I know what changes in behavior to expect in my hospitalized child”) and the parental role during hospitalization (eg, “I am clear about the things that I can do to best help my child”). Parents indicated agreement with each item on a 5-point scale, ranging from 1 (strongly disagree) to 5 (strongly agree). Item scores were summed, with a possible range of scores of 20 to 100; higher scores indicated stronger beliefs.
To assess whether subjects processed the information they received at the 3 intervention points, mothers answered 12 manipulation check questions (6 of the 12 questions were specific to the COPE program and 6 were specific to the control program). Higher scores for the mothers on the questions relevant to their experimental program provided evidence of processing of the information for the study condition. The timing of the interventions and the collection of outcome measures is summarized in Table 3.
To confirm that randomization produced equivalent study groups within and across study sites, frequency distributions and descriptive statistics (ie, means, medians, modes, and SDs) were calculated for all demographic and clinical variables, as well as preintervention measures. No distribution anomalies were found. Student’s t tests, χ2 tests, and analyses of variance were conducted to examine the data for baseline group or site differences. Subjects from the 2 study sites were similar with respect to most variables, except 1) the gender of the child, 2) the type of admission (surgical or medical), 3) the use of restraints, and 4) the father’s level of education. To help control for these site differences, a decision was made to enter site as a covariate in subsequent analyses. Because maternal trait anxiety was highly correlated with a multitude of the study’s dependent variables (eg, mother’s depression and child outcomes), it also was used as a covariate in subsequent data analyses. To help control for preintervention group differences, maternal education and income were used as covariates, because they were correlated significantly with each other and because education was correlated with experimental condition. Analyses of covariance were used to test the study hypotheses. A repeated-measures approach was not used, because such an approach demands listwise deletion of data, which would severely limit the use of available data for this sample.
Despite multiple attempts to gather all posthospitalization mailed survey questionnaires from the subjects, there was a 58.2% rate of attrition in the study by 1 year after hospitalization (n = 44, 50.5%, in the COPE group and n = 51, 67%, in the control group). Because of the smaller numbers of subjects at postdischarge time points, which resulted in low observed power of our statistical tests to detect significant findings, a decision was made to calculate and report effect sizes for the study’s outcome variables, as well as the traditional results from statistical analyses. Effect sizes are not dependent on sample size and help avoid type II errors in the interpretation of the findings. Pooled within-group SDs were used to calculate effect sizes, and values of .2, .5, and .8 were used to define small, medium, and large effects, respectively.25 Table 4 presents the study findings.
Maternal Emotional Outcomes During Hospitalization
There were no differences with respect to state anxiety and total negative mood state during hospitalization. The 2 groups demonstrated similar patterns of decline across the 4 time points at which these variables were measured (Table 4).
Although there was only a small positive effect of the COPE program on maternal reports of overall parental stress during the PICU stay, as assessed with the Parental Stressor Scale: PICU (Table 4), mothers in the COPE group reported significantly less stress regarding staff communication than did control group mothers (COPE: mean score = 4.3, SD = 3.9; control: mean score = 6.0, SD = 5.9; P < .05). Mothers in the COPE group reported significantly less total stress on the Parental Stressor Scale: PICU after transfer to the general pediatric unit, compared with control mothers (Table 4). Specifically, COPE mothers reported significantly less stress, compared with control mothers, on the following subscales: 1) their children’s appearance (COPE: mean = 2.9, SD = 2.9; control: mean = 4.7, SD = 4.0; P ≤ .01), 2) their children’s procedures (COPE: mean = 10.8, SD = 6.9; control: mean = 13.6, SD = 7.7; P ≤ .05), and 3) their children’s behaviors and emotions (COPE: mean = 19.0, SD = 9.8; control: mean = 23.4, SD = 11.9; P ≤ .01).
Mothers’ Involvement in Their Children’s Care
Nurses who were blinded with respect to study groups rated the COPE mothers as significantly more involved in their children’s physical care on the pediatric unit, compared with control mothers (Table 4). In addition, they rated COPE mothers as being more involved in their children’s emotional care on the pediatric unit, compared with control mothers. There were no significant differences related to maternal involvement in care rated by the PICU nurses.
Maternal Anxiety and Negative Mood After Hospitalization
Although there were no significant differences between groups, positive effect sizes (range: .25–.40) were found for the COPE program with respect to maternal state anxiety at 3 of the 4 postdischarge time points. With respect to total negative mood state, COPE mothers reported significantly less negative mood 1 month after hospitalization, compared with control mothers (Table 4). In addition, there was a trend for COPE mothers to report less total negative mood 1 year after hospitalization, compared with control mothers (P < .10, medium effect size). There were also small positive effects of the COPE program on total negative mood state at all of the postdischarge time points (range: .29–.42). Specifically, COPE mothers reported significantly less depression 1 month after hospitalization and 6 months after hospitalization, compared with control mothers (Fig 1), with effect sizes ranging from .20 to .45.
Because the number of stressful events in the year before hospitalization was correlated significantly with PTSD symptoms after hospitalization (r = .24–.25, P < .05, across the 4 postdischarge time points), this variable was used as a covariate in this analysis, in addition to site and trait anxiety. There were differences in mothers’ work status after hospitalization, and this variable was also controlled for in the analysis.
There were no significant differences in reports of PTSD symptoms in the COPE and control groups 1 and 3 months after hospitalization, although there was a small positive effect (effect size: .22) for the COPE program at 3 months. There was a trend for the COPE mothers to report fewer PTSD symptoms at 6 months, compared with the control mothers, even after controlling for the number of stressful events before hospitalization. At 12 months, the COPE mothers reported statistically fewer PTSD symptoms, compared with the control mothers (Fig 2).
COPE mothers reported significantly fewer behavioral symptoms for their children on the BASC behavioral symptoms index composite 12 months after discharge, compared with control mothers (Fig 3). One year after discharge, there was a significantly higher percentage of control group children (25.9%) with clinically significant behavioral symptoms, compared with COPE group children (2.3%) (χ2 [1 df] = 9.36, P < .01). In addition, although it was not significant, there was a small positive effect (effect size: .27) for the COPE program with respect to the behavioral symptoms index composite 3 months after discharge.
With respect to the other BASC composite subscales, COPE mothers reported significantly fewer externalizing problems for their children 12 months after discharge, compared with control mothers (Fig 4; Table 5). At 6 and 12 months after discharge, there was a significantly higher percentage of control group children with clinically significant externalizing symptoms (6 months, 14.3%; 12 months, 22.2%), compared with the COPE children (6 months, 1.8%; 12 months, 4.5%) (χ2 [1 df] at 6 months = 5.59, P < .05; χ2 [1 df] at 12 months = 5.23, P < .05). Although there was no statistically significant difference between the 2 study groups with respect to the adaptive skills composite and the internalizing composite, there were positive effects of the COPE program at 12 months (effect size: .45 and .38, respectively).
With respect to specific BASC subscales, COPE mothers reported significantly fewer withdrawal symptoms at the 6-month follow-up assessment (COPE: mean = 42.5, SD = 28.5; control: mean = 30.4, SD = 23.1; P = .05, effect size: .45). In addition, although values were not significantly different, there was a small positive effect (effect size: .42) of the COPE program on attention problems 6 months after hospitalization.
Twelve months after discharge, COPE mothers rated their children significantly lower on the hyperactivity BASC subscale and higher on the adaptability BASC subscale, in comparison with control mothers. There was also a trend for the COPE children to be rated as less depressed (P < .10, effect size: .49). Although they were not statistically significant, there were small positive effects of the COPE program on aggression (effect size: .41), anxiety (effect size: .29), and attention problems (effect size: .20). In addition, 22.2% of the control group children met the criteria for clinically significant maladaptive behaviors, compared with 6.8% of the COPE group children (χ2 [1 df] = 3.59, P < .10).
COPE mothers reported statistically stronger beliefs regarding their children’s likely responses to hospitalization and how they could enhance their children’s adjustment, compared with control mothers (COPE: mean = 72.7, SD = 10; control: mean = 69.0, SD = 13.5; P = .05).
The data were analyzed to determine whether families who left the study were different with respect to major demographic factors (eg, age, race, or gender) and/or baseline clinical variables (eg, Pediatric Risk of Mortality scores or mothers’ trait anxiety), as well as the BASC measures before the 12-month follow-up assessment. Analyses indicated that there were no significant differences in any of these measures between families who remained in the study at the 12-month follow-up assessment and those who left the study. Because of the attrition rate from 6 months to 12 months, a systematic analysis was conducted to assess whether the study findings were influenced by patterns of missing data. There were no predictors of missing data except for group assignment (the control group had more missing data with time) and race (the percentage of nonwhite respondents decreased with time). However, the effects of the treatment did not differ according to race, and race did not predict outcomes at any time point.
To examine the robustness of the study findings more thoroughly, multiple-imputation strategies were used. In general, multiple imputation imputes missing scale values with regression-based maximal likelihood procedures but also incorporates random error into the estimates.26 To improve estimates, key demographic factors and the important covariates were used in the imputation strategy, and the procedures described by Allison27 for intervention studies were used. Ten imputations were performed, and each of the 10 datasets was analyzed as described above. The results obtained for each of the 10 datasets were combined by using the techniques described by Rubin.28
Overall, the results obtained with multiple imputation confirmed the findings derived by using listwise techniques. Specifically, COPE mothers 1) were significantly more involved in their children’s physical and emotional care in the pediatric unit, 2) were less stressed regarding their children’s overall appearance, 3) reported less total negative mood 1 year after hospitalization, 4) reported fewer externalizing problems for their children 12 months after discharge, and 5) reported fewer aggressive problems for their children 1 year after discharge (all P < .05). Moreover, there were trends for differences between the 2 groups, with COPE mothers reporting 1) less total stress after transfer to the general pediatric unit, 2) less stress regarding their children’s medical procedures and their children’s behaviors and emotions, 3) less negative mood and depression 1 month after hospitalization, 4) fewer PTSD symptoms 6 months after hospitalization, and 5) less depression among their children 12 months after discharge, compared with control mothers. The level of agreement in the findings between the listwise and multiple-imputation approaches to the handling of missing data increases the robustness of the findings.
The principal finding of this study is that the mental health/psychosocial outcomes of critically ill children can be improved by enhancing their mothers’ coping outcomes, during and after hospitalization, with the COPE program. Despite low statistical power at the 12-month follow-up assessments, there were significant differences between the COPE and control groups with respect to major maternal and child outcome variables, with positive effects of the COPE intervention. Therefore, we think that the findings of this study represent true improvements among critically ill children and their mothers who received the experimental program, with the results of this study supporting the value of the COPE intervention in improving the mental health and coping outcomes of mothers and young children who experience critical care hospitalization.
Because the COPE program created a clear cognitive schema for their children’s likely behaviors and emotions as they recovered from critical illnesses, mothers reported stronger beliefs regarding their ability to understand and predict their children’s responses to hospitalization. As a result, COPE mothers experienced less parental stress after their children were transferred to the general pediatric unit, in comparison with control mothers. The mounting evidence regarding the findings from this study and our previous work6,22,29,30 provides strong support for the self-regulation theory10 in explaining how the child behavior component of the COPE program positively affects maternal coping.
A lack of differences between the COPE and control group mothers with respect to state anxiety and negative mood state during hospitalization may be attributable to the fact that the length of hospital stay for this full-scale clinical trial was approximately one-half of that in our pilot study.6 There might not have been enough time to demonstrate the positive effects of the COPE program on maternal anxiety and mood state during the short course of hospitalization in this trial. The control group did receive an intervention at a high-stress time, which might have decreased anxiety and negative mood, in comparison with a “pure” control group receiving standard care only.
COPE mothers did report less anxiety and negative mood state (specifically depression) at various times after hospitalization. In addition to statistically significant differences, there were clinically meaningful differences in state anxiety; mothers in the control group scored in the 72nd to 73rd percentile of standardized normative values, compared with mothers in the COPE group, who scored in the 58th to 59th percentile 12 months after discharge. Overall, upon reflection, the hospitalization might have been less stressful for the COPE parents than the control parents. Because mothers in the COPE group were reported as being more involved in their children’s emotional and physical care in the pediatric unit, they might have felt better about their ability to cope with the stressful event and to enhance their children’s coping outcomes.
Although there were no significant differences between the study groups with respect to the nurses’ ratings of maternal involvement in the physical and emotional care of the children in the PICU, the children’s nurses, who were blinded with respect to study group, rated the COPE mothers as being significantly more involved in their children’s care in the pediatric unit. Because of the short length of stay, mothers in the COPE program might not have had time to begin to demonstrate higher levels of involvement in their children’s care, compared with control mothers, in the PICU. In the first 2 days in the PICU, children typically undergo numerous treatments and tests, which may not create an environment in which parents feel at ease in performing multiple parenting activities. Findings from earlier studies support this reasoning; emergency admission of a child to the PICU produces a major loss of parenting roles for parents, which is a major source of stress for them. However, after transfer to the pediatric unit, mothers in the COPE program received a booster intervention that suggested strategies regarding how they could continue to participate in their children’s care, as well as activities to perform with their children that could enhance the children’s adjustment. As a result, according to control theory,13 barriers to parent participation were most likely removed, so that these mothers could provide more physical and emotional care to their children.
With respect to PTSD symptoms, there was a trend for COPE mothers to report fewer symptoms 6 months after discharge, which led to a greater divergence of PTSD scores between the 2 study groups 12 months after discharge and resulted in a statistically significant difference. These differences may be related to the fact that COPE mothers reported less parental stress during hospitalization and participated more generally in their children’s emotional and physical care, compared with control mothers. Therefore, the experience might not have been as stressful for them. The role that a person has in a traumatic experience is a major factor that increases the likelihood of developing PTSD.31 Specifically, a previous study found that parents whose children underwent bone marrow transplantation might experience PTSD, largely because of their being unable to be as supportive to their children and as actively involved in their care as they were during outpatient treatments, with high stress levels and feelings of helplessness.5 Those investigators hypothesized that loss of parental roles might be a factor contributing to the development of PTSD among parents whose children undergo bone marrow transplantation. The COPE mothers in this study were more involved in their children’s care in the pediatric unit than were the control parents, which might explain why the former experienced fewer PTSD symptoms after the experience. In addition, COPE mothers were able to anticipate their children’s emotions and behaviors, which might aid in decreasing uncertainty and feelings of helplessness.
Overall, the findings from this study indicated that children of mothers who received the COPE program exhibited significantly fewer adjustment problems with time, compared with children of mothers in the control group. Although they were not statistically significant, small positive effects of the COPE program were noticeable 3 months after hospitalization. At the 6-month follow-up assessments, COPE mothers reported fewer clinically significant behavioral symptoms and withdrawal symptoms, compared with control mothers. Twelve months after discharge, COPE mothers rated their children significantly lower in hyperactivity and higher in adaptability, compared with control mothers. These are very important clinical findings, because there is evidence that children who have been hospitalized in critical care units are at higher risk for psychologic trauma and/or behavior problems that demand psychiatric attention, compared with other hospitalized pediatric patients.3 As hypothesized, COPE mothers experienced improved maternal functional and emotional coping outcomes, which resulted in fewer child adjustment problems.
With the increasing prevalence of attention-deficit/hyperactivity disorder and externalizing problems among children and the documented lack of mental health screening and early intervention services for children in this country,32,33 our COPE intervention could help protect this high-risk population of children from developing these troublesome problems. As a result, the mental health status of children after critical care hospitalization could be improved. In addition, family burdens and costs associated with the mental health treatment of these problems might be substantially reduced.
A limitation of this clinical trial was the rate of attrition through 12 months after discharge, which presents a threat to the internal validity of the study. Even with follow-up telephone calls, monetary incentives for completing questionnaires, repeated mailings to the mothers after hospital discharge, and cultural sensitivity training for research assistants, attrition from the study was sizeable. For the method used for follow-up assessment (ie, mailed questionnaires), however, the return rate was fairly typical.34 Although there was no apparent relationship between the missing data and outcomes, overcoming the challenge of obtaining posthospitalization follow-up data with this high-risk population will be essential for future studies. Home visits in which data are obtained at regularly scheduled intervals may be more successful than mailed questionnaires. In our current intervention trial with high-risk parents and low birth weight premature infants, testing the effects of another version of the COPE program, home visits have resulted in a substantially higher subject retention rate and collection of long-term follow-up data up to 2 years after hospital discharge.
This is the first full-scale, randomized trial to support the effectiveness of an efficient, easy-to-administer, reproducible intervention program for parents of critically ill young children. On the basis of this study’s findings and the accumulating evidence of the benefits of our similar intervention programs with parents of hospitalized children and premature infants,6,23,27,35 the COPE program should be provided routinely to parents of young children admitted unexpectedly to the PICU. Future research is needed to determine 1) the cost-effectiveness of the COPE intervention, 2) whether its effects can be strengthened by including fathers, 3) whether its effects can be strengthened with additional intervention sessions 1 and 3 months after hospital discharge, and 4) whether similar positive effects can be obtained if the program is delivered to parents of younger and older children and parents whose critically ill children have chronic and/or terminal conditions.
This research was supported by the National Institutes of Health, National Institute of Nursing Research grant R01-NR04174.
Special thanks go to Zendi Moldenhauer, MS, RN, and Nancee Bender, PhD, RN, for assistance with this research.
- Received February 2, 2004.
- Accepted February 2, 2004.
- Reprint requests to (B.M.M.) University of Rochester School of Nursing, 601 Elmwood Ave, Box SON, Rochester, NY 14642. E-mail:
- ↵Zimmerman JJ. The pediatric critical care patient. In: Furhman BP, Zimmerman JJ, eds. Pediatric Critical Care. 2nd ed. New York, NY: Mosby; 1998:1–5
- ↵Pollack MM. Outcome assessment. Crit Care Med.1993;21(suppl) :S395– S396
- ↵American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: American Psychiatric Association; 1994
- Wolfer JA, Visintainer MA. Pediatric surgical patients’ and parents’ stress responses and adjustment. Nurs Res.1975;24 :245– 255
- ↵Leventhal H, Johnson JE. Laboratory and field experimentation: development of a theory of self-regulation. In: Woolridge PJ, Schmitt MH, Skipper JK Jr, Leonard RC, eds. Behavioral Science and Nursing Theory. St. Louis, MO: Mosby; 1983:189–262
- ↵VanderVeer AH. The psychopathology of physical illness and hospital residence. Q J Child Behav.1949;1 :55– 79
- Jimerson SS. Patterns of anxiety. In: Haber J, Hoskins PP, Leach AM, Sideleau BF, eds. Comprehensive Psychiatric Nursing. 3rd ed. New York, NY: McGraw-Hill; 1987:545–563
- ↵Spielberger C, Gorsuch R, Lushene R. The STAI Manual. Palo Alto, CA: Consulting Psychologists; 1977
- ↵Lorr M, McNair D. Profile of Mood States. San Diego, CA: Educational and Industrial Testing Service; 1982
- ↵Miles MS, Carter MC. Sources of parental stress in pediatric intensive care units. Child Health Care.1982;11 :65– 69
- ↵Reynolds CR, Kamphaus RW. Behavioral Assessment System for Children Manual. Circle Pines, MN: American Guidance Service Inc; 1992
- ↵Melnyk BM. Coping with unplanned childhood hospitalization: the mediating functions of parental beliefs. J Pediatr Psychol.1994;20 :299– 312
- ↵Cohen J. Statistical Power Analysis for the Behavioral Sciences. Hillsdale, NJ: Lawrence Erlbaum Associates; 1988:407–465
- ↵Schafer JL. Analysis of Incomplete Multivariate Data. London, United Kingdom: Chapman & Hall; 1997
- ↵Allison PD. Missing Data. Newbury Park, CA: Sage; 2002
- ↵Rubin DB. Multiple Imputation for Nonresponse in Surveys. New York, NY: Wiley; 1987
- ↵Wilson J, Smith W, Johnson S. A comparative analysis of PTSD among survivor groups. In: Figley CR, ed. Trauma and Its Wake. New York, NY: Brunner-Mazel; 1985:142–172
- ↵United States Office of the Surgeon General, Department of Health and Human Services. Mental Health: A Report of the Surgeon General. United States Office of the Surgeon General, Department of Health and Human Services; chap 3. Available at: www.surgeongeneral.gov/library/mentalhealth/home.html. Accessed November 12, 2002
- ↵Babbie E. Survey Research Methods. 2nd ed. Bellmont, CA: Wadsworth Publishers; 1997:176–201
- Copyright © 2004 by the American Academy of Pediatrics