Reducing Premature Infants' Length of Stay and Improving Parents' Mental Health Outcomes With the Creating Opportunities for Parent Empowerment (COPE) Neonatal Intensive Care Unit Program: A Randomized, Controlled Trial
OBJECTIVE. Although low birth weight premature infants and parents are at high risk for adverse health outcomes, there is a paucity of studies that test early NICU interventions with parents to prevent the development of negative parent-infant interaction trajectories and to reduce hospital length of stay. Our objective was to evaluate the efficacy of an educational-behavioral intervention program (ie, Creating Opportunities for Parent Empowerment) that was designed to enhance parent-infant interactions and parent mental health outcomes for the ultimate purpose of improving child developmental and behavior outcomes.
DESIGN, SETTING, AND PARTICIPANTS. A randomized, controlled trial was conducted with 260 families with preterm infants from 2001 to 2004 in 2 NICUs in the northeast United States. Parents completed self-administered instruments during hospitalization, within 7 days after infant discharge, and at 2 months' corrected age. Blinded observers rated parent-infant interactions in the NICU.
INTERVENTION. All participants received 4 intervention sessions of audiotaped and written materials. Parents in the Creating Opportunities for Parent Empowerment program received information and behavioral activities about the appearance and behavioral characteristics of preterm infants and how best to parent them. The comparison intervention contained information regarding hospital services and policies.
MAIN OUTCOME MEASURES. Parental stress, depression, anxiety, and beliefs; parent-infant interaction during the NICU stay; NICU length of stay; and total hospitalization were measured.
RESULTS. Mothers in the Creating Opportunities for Parent Empowerment program reported significantly less stress in the NICU and less depression and anxiety at 2 months' corrected infant age than did comparison mothers. Blinded observers rated mothers and fathers in the Creating Opportunities for Parent Empowerment program as more positive in interactions with their infants. Mothers and fathers also reported stronger beliefs about their parental role and what behaviors and characteristics to expect of their infants during hospitalization. Infants in the Creating Opportunities for Parent Empowerment program had a 3.8-day shorter NICU length of stay (mean: 31.86 vs 35.63 days) and 3.9-day shorter total hospital length of stay (mean: 35.29 vs 39.19 days) than did comparison infants.
CONCLUSIONS. A reproducible educational-behavioral intervention program for parents that commences early in the NICU can improve parent mental health outcomes, enhance parent-infant interaction, and reduce hospital length of stay.
- parent-infant/child interaction
- intensive care
- premature infants
- parental beliefs
- randomized, controlled trial
More than 480000 premature infants are born every year in the United States. This incidence is a 27% rise in premature births since 1981.1 Although advances in technology have allowed greater numbers of infants to survive, morbidity remains high and imposes great emotional and financial burdens on families, society, and the health care system.2 Evidence has accumulated to support that low birth weight (LBW) premature infants experience a host of adverse physical and mental health/behavioral problems (eg, delays in cognitive and neurologic development, behavior problems, attention-deficit disorder) that persist well into the school-age and adolescent years.3–8 Furthermore, the costs associated with a NICU stay (approximately $1250–$2000/day) coupled with increased medical utilization postdischarge are exorbitant.9–12
Parents of preterm infants experience high stress levels and feelings of helplessness in the NICU and often lack adequate knowledge of how to parent and interact with their infants during the hospital stay.13–15 As such, they frequently develop misperceptions of their infants and have difficulty interacting with them in a developmentally sensitive manner.16 These negative parent-infant interaction trajectories often lead to adverse parent outcomes after hospitalization, such as depression and anxiety disorders and other dysfunctional parenting patterns.17–19
Despite these adverse outcomes, there has been a paucity of theory-based, parent-focused intervention programs initiated early in the NICU stay to reduce parental stress and empower parents to develop healthy perceptions of and interaction patterns with their premature infants.20 Early intervention may be key to successful programs with parents of preterm infants because it is recognized that once a negative trajectory of parent-infant interaction is initiated in the NICU, changes are difficult to implement and even harder to sustain.21 Some evidence also exists to support that mothers who rapidly engage in the care of their infants in the NICU tend to be more sensitive to their infants' cues and have better relationships with them in later years.22
Findings from our pilot study with 42 mothers of preterm infants that tested the effects of an educational-behavioral intervention program (Creating Opportunities for Parent Empowerment [COPE]) indicated that mothers who were in the COPE program had less stress in the NICU than did mothers who were in a comparison program. In addition, their infants scored almost 1 SD higher in cognitive development than did infants of parents in the comparison group at 6 months' corrected age.14 To confirm these pilot findings and further address this gap in preventive intervention work with parents of LBW premature infants, the purpose of this study was to evaluate the efficacy of the COPE program through a multisite randomized clinical trial conducted in 2 NICUs in the northeast region of the United States.
On the basis of these theories, it was hypothesized that the COPE intervention would strengthen parents' knowledge and beliefs about their preterm infants and their own parenting role and remove barriers that would inhibit them from participating in their infants' care and interacting with them in a developmentally sensitive manner. As a result, parents in the COPE program were expected to have less stress and negative mental health outcomes (eg, depression and anxiety) during the NICU stay and 2 months postdischarge. In addition, because of enhanced parental beliefs in what behaviors and characteristics to expect of their infants as well as more confidence in caring for and interacting with their infants in a developmentally sensitive manner, it was expected that discharge would occur sooner for those in the COPE group than those in the comparison group.
This study was conducted from September 2001 to October 2004 in 2 NICUs (a 52-bed NICU within a 720-bed academic medical center [University of Rochester Medical Center, Rochester, NY] and a 60-bed NICU within a 566-bed medical center [Crouse Hospital, Syracuse, NY]). Home interventions and follow-up visits were completed by research nurses who were rigorously trained in the study protocol. The research subjects review boards at both sites approved the study protocol before its implementation. All mothers and fathers ≥18 years of age who could read and speak English, who had not had another infant admitted to the NICU, and whose infants met the following criteria were eligible for participation: (1) gestational age of 26 to 34 weeks inclusive; (2) birth weight of <2500 g and appropriate for gestational age; (3) anticipated survival; (4) singleton birth; (5) no severe handicapping conditions including grade III or IV intraventricular hemorrhage; and (6) born at the study sites. On the basis of our pilot-study findings,14 it was expected that there would be fathers who were not consistently involved in their children's lives. To be nonexclusionary and in situations where the infants' fathers were uninvolved, mothers could select a significant other who would be playing a supportive role in their children's daily lives to participate in the study with them.
Recruiters screened 1278 premature births, with 562 meeting the eligibility criteria (Fig 1). Of the 562 eligible births, 254 (45.2%) families refused participation, with the majority of parents stating that they were too stressed or tired to participate, they wanted to concentrate only on the new infant, or they believed that participation would take time away from the other children at home. Recruiters were unable to contact the parents of 48 (8.5%) premature infants who met eligibility criteria. Parents who agreed to participate completed baseline measures and were randomly assigned to study conditions by using a blinded system generated by the investigators. Group assignment was concealed by using opaque, sealed envelopes. The envelopes were not opened by the interventionists until after the subjects were enrolled.
The final sample after randomization comprised 260 families, including 258 mothers (147 in the COPE group and 113 in the comparison group) and 154 fathers/significant others (81 in the COPE group and 73 in the comparison group). Although 2 mothers chose not to participate, the fathers of those infants were enrolled. One grandmother also was enrolled as a significant other. A higher number of parents were enrolled in the COPE group because of plans to administer a booster intervention for half of the parents in the COPE program during the toddler years to determine if the effects of the program could be strengthened during the preschool years. After enrollment and randomization of study groups, data for 13 families were eliminated from the study for analyses. Nine families in the COPE program were eliminated from the study because of infant death (n = 3), transfer of child custody (n = 2), or completion of baseline measures only (n = 4). Data for 4 comparison families also were eliminated from the analysis because of findings of a significant congenital anomaly (n = 1) or completion of baseline measures only (n = 3). Therefore, the final sample for data analyses included 245 mothers, 145 fathers, and their 247 preterm infants.
A randomized, controlled trial (RCT) was conducted in which subjects at each of the 2 study sites were randomly assigned to be in either the COPE program or a comparison intervention program (see Fig 1). The random assignment, which was hidden from the enrolling research assistants until after informed consent was obtained, was made by 4-week blocks of time to decrease the probability of staff-to-parent and parent-to-parent contamination (ie, the likelihood that parents in the NICU who were in different study groups would exchange content that was provided to them in their randomly assigned interventions). Four-week blocks of time were chosen because, during the previous year, the mean length of stay (LOS) for LBW premature infants in the 2 NICUs was ∼28 days. The study protocol was implemented to prevent cross contamination between parent groups (eg, parents listened privately to the audiotapes with headphones on or in a private room; both groups received intervention materials packaged in identical notebooks for their personal use). In addition, observationists also were blinded to study group, having no access to data or information regarding group assignment, to avoid contamination or bias.
The COPE program is a 4-phase educational-behavioral intervention program. Each phase provides parents with information on (1) the appearance and behavioral characteristics of premature infants (infant-behavior information) and how parents can participate in their infants' care, meet their infants' needs, enhance quality of interaction with their infant, and facilitate their infant's development (parent-role information) and (2) activities that assist parents in implementing the experimental information. An educational-behavioral intervention strategy was used because information reinforced with behavioral activities has been shown to be more effective in producing change in outcomes than does information alone.26 For the COPE program to achieve widespread diffusion without intensive staff training and time through person-to-person interventions, a decision was made to package the intervention as audiotaped and written information along with prescribed activities so that it could be easily reproduced and administered to all parents of preterm infants in NICUs at low cost.
Phase I of the COPE program occurred 2 to 4 days after the infants' admission to the NICU and consisted of audiotaped infant-behavior and parent-role information. A written copy of the audiotaped information was provided by the research assistant for parents to read as they listened to the audiotape. In addition, parents were given their first set of parenting activities to complete that consisted of identifying the special characteristics of their infants and keeping a record of their milestones in the NICU.
Phase II occurred 2 to 4 days after the first intervention and consisted of audiotaped and written information that reinforced critical content of the initial audiotape and provided supplemental information on premature infants' behaviors and development. Additional suggestions regarding how parents could participate in their infants' care and enhance their development also were provided. The activities in this phase consisted of identifying the special characteristics of their preterm infants and recognizing their infants' stress cues as well as cues that signal readiness for interaction.
Phase III of the COPE program occurred 1 to 4 days before discharge and consisted of developmentally appropriate audiotaped information, including specific details about infant states and the best times for interaction as well as parent-role information about how to smooth the transition from hospital to home and how to continue to foster a positive parent-infant relationship. Parent activities during phase III included continuing to assist parents in recognizing their infants' cues and helping parents to recognize strategies to assist their infants when stressed.
Phase IV of the COPE program was delivered in the parents' home 1 week after infant discharge. Audiotaped and written information contained anticipatory guidance specific to preterm-infant development, suggestions for fostering a positive parent-infant relationship, and specific activities for parents to perform to foster their infants' cognitive development.
To control for time and attention and diminish the likelihood that the effects of the COPE program could be attributed to Hawthorne effects, a comparison intervention program was implemented. The comparison program consisted of a series of audiotapes and written information delivered at the same times as the 4 phases of the COPE program. The first 2 tapes of the comparison program provided information about hospital services, the third tape contained discharge information given to all parents, and the fourth tape had information regarding immunizations.
Data collection occurred at baseline (immediately before delivery of the phase I intervention), at each of the phase II through IV phase interventions (2–4 days after the first intervention, 1–4 days before infant discharge from the NICU, and 1 week post–NICU discharge, respectively), and at 2 months' corrected infant age (see Table 1). At each contact, data were obtained from the parents. In addition, observations of parent-infant interaction were made 1 to 2 days after the second intervention by trained observers who were blinded to study group. The primary care nurses, also blinded to study group, rated the parents on their involvement in their infants' care at 1 to 2 days after the second intervention as well.
Primary Outcome Measures
The following measures were used to assess parental emotional and functional coping outcomes (ie, the primary outcomes for the parents) and parental beliefs (ie, the mediating variable through which the COPE program was expected to positively impact parent outcomes).
Infant LOS was the primary infant outcome. Cronbach's α reliability coefficients were determined and are reported for the scales, where appropriate, in Table 1. On the basis of the effect sizes found on the primary outcomes in our pilot study, we expected to find both short-term and long-term effects from the COPE program. Therefore, the data collected were examined at each time point.
For emotional coping outcome, the valid and reliable State-Trait Anxiety Inventory27 was used to assess anxiety. A person's baseline anxiety “proneness” is measured with the 20-item trait scale (A-Trait), and current feelings of anxiety are measured with the A-State scale.
The well-known valid and reliable Beck Depression Inventory (2nd edition) (BDI-II)28 was used to measure depressive symptoms.
The Parental Stressor Scale-Neonatal Intensive Care (PSS-NICU),29 with its established construct validity, was used to assess parental stress in the NICU arising from 4 dimensions: (1) sights and sounds; (2) infant behavior and appearance; (3) parental role alteration; and (4) staff behaviors and communication. Parents rated their perceived stress on each of 46 items using a 5-point Likert scale ranging from 1 (nonstressful) to 5 (extremely stressful). Higher total scores indicate greater stress levels.
For functional coping outcome, the 15-item Index of Parental Behavior in the NICU30 measured the quality of parent-infant interaction in the NICU. Blinded observers rated each item representing parent behaviors (eg, parent seeks to interact face to face with infant during the quiet alert state) on a dichotomous scale (exhibited the behavior or did not) during a half-hour observation period. The scale items were developed through a thorough review of the literature and observations of parents in the NICU. Content validity was established by 8 maternal-child experts and an expert in cultural competence. These same observers also rated the overall appropriateness of parent interaction with their infant on a 1-item visual analog scale ranging from 0 to 100, with 100 indicating a more positive interaction. Interrater reliability of 90% was established with the blinded observers. Factor analysis revealed 2 subscales: positive interaction with infant in quiet alert state (7 items) and altering environment and interaction with stressed infant (8 items).
The infants' primary care staff nurses (blinded to group) also rated parents on the basis of their general observations of parent involvement in the physical care of their infants and sensitivity to the needs of their infants. These 2 ratings ranged from 0 to 100, with 100 indicating greater involvement and sensitivity.
For the process variable, the 18-item Parental Belief Scale-NICU31 measured parents' beliefs about their premature infants and their parental role during hospitalization. Parents indicated agreement with each item on a 5-point Likert scale from 1 (strongly disagree) to 5 (strongly agree), with higher scores indicating more-positive beliefs. Eight neonatal specialists established content validity of the scale. The 2 subscales are parent role confidence (10 items) and parent knowledge regarding infant behaviors and characteristics (8 items).
Infant NICU LOS was reported as the number of days of hospitalization in the NICU. The total hospital LOS included NICU days plus convalescent care in a transfer hospital or step-down unit before being discharged to home. To estimate the expenses associated with days spent in the NICU, the number of days was multiplied by a cost of $1250/day. This value is a conservative cost estimate based on 1998 dollars and reflects the median treatment costs, which includes accommodation and ancillary costs.12
Manipulation checks were administered to both study groups after each intervention to assess the processing of information. Each 12-item manipulation check included 6 COPE- and 6 comparison-group–specific questions. Higher scores for the COPE program versus comparison group on the COPE-specific information provided evidence of processing the experimental information.
The Clinical Risk Index for Babies (CRIB), a well known valid and reliable measure for infants in the NICU, measured severity of illness. Scores range from 0 to 23, with higher scores indicating increased severity.32
On the basis of the medium to large effect sizes (Cohen's d statistic) associated with group differences in our pilot findings14 (anxiety, depression, parental beliefs), we projected a medium-to-large effect size for key short-term and long-term outcomes in the current study. A final sample size of 144 subjects (72 in each group) was required to achieve a power of 0.8 at the .05 level of significance to test study hypotheses on the basis of the proposed effect sizes and statistical procedures to be used.33 Additional parents were recruited to allow for an estimate of 30% attrition over time, because longitudinal follow-up through the school-age years was planned.
Analyses were conducted on predetermined hypotheses, and subjects were analyzed in the study group to which they were originally assigned (intention-to-treat analysis).34,35 A P value of ≤.05 was set for statistical significance. Descriptive statistics were conducted at baseline to summarize demographic and clinical variables. In addition, t tests for continuous variables and χ2 analyses for categorical variables were performed to determine if the 2 study groups differed on these baseline factors. If the 2 groups differed significantly on baseline variables and those variables correlated with the specified outcomes, they were controlled for statistically in the analyses to determine efficacy of the intervention. In addition, variables were used as covariates if they were theoretically or empirically supported as confounders in the literature. Because there were a few site differences between study groups (ie, trait anxiety, the number of mothers on public assistance), study site and these 2 variables were controlled for statistically through analyses of covariance to determine if group differences existed on the outcome variables.
The 258 mothers ranged in age from 18 to 43 years (mean: 27.8 years; SD: 6.7 years) (see Table 2). The racial and ethnic composition of the mothers included 174 (67.4%) white, not Hispanic, 58 (22.5%) black, not Hispanic, and 26 (10.1%) in “other” categories (10 Hispanic, 8 Asian, 1 American Indian, and 7 >1 race). The majority of mothers (n = 212 [82.2%]) completed at least high school. Household income for families was less than $20001 for 82 (31.5%), between $20001 and $40000 for 52 (20.2%), and greater than $40000 for 103 (39.6%), and 23 (8.8%) families did not report their household income. Eighty-six of the mothers (33.3%) reported being on public assistance. One hundred thirty-eight (53.4%) of the mothers were married, 98 (38.0%) had never been married, 11 (4.3%) reported some other status (divorced or separated), and 11 (4.3%) did not report their status. The ages of the 155 fathers/significant others ranged from 18 to 49 (mean: 30.6 years; SD: 7.4 years). The racial and ethnic composition of the fathers was: 115 (76.2%) white, not Hispanic, 23 (15.2%) black, not Hispanic, and 13 (8.6%) other (5 Hispanic, 4 Asian, 1 American Indian, and 3 >1 race). The majority of fathers (n = 126 [83.4%]) completed at least high school. Ninety six (63.6%) of the fathers were married, 39 (25.8%) had never been married, 7 (3.2%) reported some other status (divorced or separated), and 9 (6.0%) did not report their status.
The mean gestational age of the premature infants was 31.3 weeks (SD: 2.4; range: 26.0–35.0 weeks). The mean birth weight was 1650 g (SD: 476 g), with 102 (39.2%) infants weighing <1500 g, 84 (32.3%) in the 1500- to 1999-g range, and 74 (28.5%) weighing ≥2000 g. One hundred twenty six (48.5%) of the infants were males, and 134 (51.5%) were females. The mean illness severity index level using the CRIB score was 1.7 (SD: 2.5; range: 0–12). Mean weight at discharge from the NICU was 2150.3 g (SD: 440.4 g; range: 955–4710 g). Total LOS in the NICU averaged 35.2 days (SD: 27.3 days; range: 5–163 days). Sixty-one infants (24.3%) were transferred to another hospital before discharge to home.
Participants Versus Nonparticipants
There were 254 infants who qualified for the study whose parents refused to participate. These families did not differ from those who enrolled in regards to maternal age (mean: 31.9 [SD: 2.1] vs 31.3 [SD: 2.4] weeks) or race/ethnicity, although their infants were approximately half of a week older and slightly heavier (mean: 1765.1 [SD: 422] vs 1649.7 [SD: 476] g) at birth and had significantly lower severity-of-illness scores at birth (mean CRIB score: 1.25 [SD: 2.2] vs 1.75 [SD: 2.5]) than infants of the enrolled parents. In summary, participating infants were younger, weighed less, and had a higher severity-of-illness score.
Of the 260 families who were randomly assigned to the 2 study groups, 147 were assigned to the COPE program and 113 to the comparison program (see Fig 1). No significant group differences were found at baseline on key demographic and clinical variables (see Tables 2 and 3). In addition, no baseline differences between groups were found for baseline depression and anxiety (Tables 4 and 5).
A rigorous study protocol was implemented in which research assistants were trained on the intervention using a reproducible format. Manipulation checks were administered after all intervention phases of each program. Subjects in the COPE program had significantly higher total scores on the questions dealing with the experimental COPE-program content as compared with comparison subjects, indicating that mothers and fathers in the COPE program processed the content of their program from all phases of the interventions (see Tables 4 and 5). In addition, comparison subjects had significantly higher total scores on the questions dealing with the comparison-program content versus parents in the COPE program. Completion of the parent activities that were prescribed to the parents in the COPE program also was monitored throughout the study. Parents who had not completed the activities by the time of the next phase of the intervention were prompted to perform them by the next contact.
Nearly all of the parents reported that the COPE information (mothers: 97.7%; fathers: 96.9%) and activities (mothers: 90.2%; fathers: 83.1%) were helpful. Approximately 62.5% of the mothers in the COPE program and 54.7% of fathers in the COPE program reported completing all NICU activities, whereas 99.2% of mothers and 98.4% of fathers reported being comfortable with the information and activities. Parents in the comparison group also stated that the information they received was helpful (mothers: 95.2%; fathers: 96.9%) and that they were comfortable with the information (mothers: 99.0%; fathers: 100%).
A total of 128 mothers and 60 fathers (130 families [88.4%]) in the COPE program and 98 mothers and 64 fathers (98 families [86.7%]) in the comparison program remained active in the study at 2 months' corrected infant age. No significant differences in study attrition by experimental group were found for families at 2 months' corrected infant age after discharge from the NICU (χ12 = 0.85).
Effects of the COPE Program
During the NICU hospitalization, mothers in the COPE program reported significantly less overall parental stress in the NICU than did mothers in the comparison group and specifically related to their general NICU stress, their infants' behaviors, and staff behaviors. There were no differences between the groups on fathers' reports of stress in the NICU. In addition, both mothers' and fathers' state anxiety and depressive symptoms decreased over time, without significant COPE- and comparison-group differences identified.
Within 2 to 4 days after the phase I intervention in the NICU, mothers and fathers in the COPE program reported significantly higher parental beliefs about their role and what characteristics and behaviors to expect from their premature infants than did comparison mothers and fathers. Mothers and fathers in the COPE program continued to report significantly higher parental beliefs 4 to 7 days before their infant's discharge from the NICU than did comparison mothers and fathers.
When controlling for site, trait anxiety, and gestational age, parents in the COPE program (both mothers and fathers) were rated by trained, blinded observers as having more positive parenting interactions with their preterm infants in the NICU on the positive interaction with infant in quiet alert state subscale of the Index of Parental Behavior in the NICU (eg, interacts face to face with infant when in the quiet alert state). Fathers in the COPE program also were more involved in their infants' care and more sensitive to their needs in the NICU than comparison fathers, as rated by blinded staff nurses assigned to the infants' care.
At 2 months' corrected infant age, mothers in the COPE program reported significantly less state anxiety and depressive symptoms than comparison mothers (Fig 2). There were no differences found for fathers' anxiety or depressive symptoms after NICU discharge.
The total NICU infant LOS (n = 241) was 3.8 days shorter for those in the COPE group (mean: 31.86 days; SD: 22.54 days; 95% confidence interval [CI]: 29.32–34.41) than for those in the comparison group (mean: 35.63 days; SD: 26.85 days; 95% CI: 32.80–38.46) (P = .05; see Fig 3). In addition, the total hospital LOS, including days spent in transfer hospitals, was 3.9 days shorter for infants in the COPE group (mean: 35.29 days; SD: 22.40 days; 95% CI: 33.06–37.52) versus comparison infants (mean: 39.19 days; SD: 27.00 days; 95% CI: 36.69–41.69) (P = .02). Although there were no significant overall group or site differences on nearly all baseline and key clinical variables, there were group differences within each site in infant gestational age as well as maternal trait anxiety at baseline. Therefore, gestational age, site, and maternal trait anxiety were controlled for statistically in these LOS analyses. Additional subgroup analyses for LOS based on birth weight revealed that smaller infants in the COPE program had an even shorter NICU and total LOS. Specifically, for very low birth weight (VLBW) infants weighing <1500 g at birth (n = 90), there was a group difference of 8.31 days for NICU LOS (COPE mean: 51.81 days; SD: 20.84 days; 95% CI: 46.31–57.31; comparison mean: 60.12 days; SD: 25.67 days; 95% CI: 54.20–66.04; P = .05) and 7.9 days for total LOS (COPE mean: 57.16 days; SD: 18.16; 95% CI: 52.49–62.84; comparison mean: 65.03 days; SD: 22.36; 95% CI: 60.06–70.00; P = .03), with infants in the COPE program being discharged sooner than comparison infants. To further understand the LOS outcomes in this study, additional secondary analyses were performed. Both mothers' and fathers' early stronger beliefs (at time 2) about their parental role and what to expect from their infants were significantly correlated with a shorter NICU LOS (mothers: r = −0.15, P = .02; fathers: r = −0.20, P = .02). Stronger beliefs of the parents in the COPE program also were correlated with total NICU plus transfer LOS (mothers: r = −0.14, P = .03; fathers: r = −0.17, P = .05). In addition, both mothers' and fathers' positive parenting behavior with their infants in the quiet alert state as rated by trained blinded observers were significantly correlated with NICU LOS and NICU plus transfer LOS (mothers: r = −0.32, n = 211, P < .001, and r = −0.35, n = 210, P < .001; fathers: r = −0.42, n = 109, P < .001, and r = −0.42, n = 108, P < .001, respectively).
To our knowledge, this is the first RCT to demonstrate that a reproducible theory-based intervention with parents of premature infants that commences early in the NICU stay results in (1) less maternal stress in the NICU, (2) stronger parental beliefs, (3) more positive parent-infant interactions in the NICU, (4) less maternal anxiety and depressive symptoms after hospitalization, and (5) a reduced LOS for preterm infants. Because mothers and fathers who were in the COPE program reported stronger beliefs in their ability to understand their preterm infants' behaviors and cues as well as how to care for and relate to their infants, they interacted with them in a more developmentally sensitive manner versus parents in the comparison program. This finding is consistent with self-regulation theory23,24 and data from our previous intervention studies with parents of hospitalized and critically ill children, which have supported that parental beliefs about their children and their role is a potent mediator of the effects of our COPE intervention on both parent and child coping/mental health outcomes.36–39 Launching of this positive parent-infant interactional trajectory shortly after admission in the NICU is critical; previous studies have indicated that early interaction patterns between parents and infants in the NICU are likely to remain consistent over time.40
To further understand the shortened LOS for infants in the COPE program, poststudy discussions were held with unit-based neonatal health care teams involved in discharge planning of all NICU infants. These teams indicated that criteria for discharge not only include that the preterm infant is physiologically stable but that the parents are ready to assume care for their infants. As a result of stronger beliefs/confidence and participation in their infants' care, parents in the COPE program may have been perceived by staff to be ready and able to take their infants home at an earlier gestational age than comparison parents. The shortened hospital stay of those in the COPE group resulted in decreased hospital costs of $5000 per infant (4 days × $1250/day), which is a substantial cost savings to an already stressed national health care system. If this cost (ie, $5000) is multiplied by 480000 LBW premature infants born every year in the United States, it would result in a $2.4 billion cost savings for the national health care system.
Although both groups had fairly low levels of depressive and anxiety symptoms 2 months after hospitalization, mothers in the COPE program reported significantly fewer symptoms on both of these outcomes than comparison mothers. This finding may be related to the fact that the mothers in the COPE program had stronger beliefs/confidence about their infants' characteristics and behaviors and how to interact and care for them as a result of the information and activities in the experimental program. We are continuing to follow these children and their parents to determine if these lower patterns of depressive and anxiety symptoms will continue longitudinally over time through 3 years of age or escalate as developmental changes occur and lags in infant development are discovered. Planned statistical testing also will include the analysis of moderating variables to determine for whom the intervention worked best over time. Longitudinal follow-up of this sample is critical, because parents of premature infants typically report overwhelming feelings of helplessness and inadequacy that contribute to sustained stress and depression over time, well into their children's preschool and school-age years.13,15,17 Interventions, such as the COPE program, that are targeted to lessen depressive symptoms are especially important in that depressed mothers have been found to be less responsive, affectionate, and positive during interactions with their infants,41 which leads to later adverse child outcomes. Specifically, maternal depression has been empirically linked with family violence, marital discord, impaired cognitive development, child abuse and neglect, and childhood mental health and behavior disorders.42–45 Despite the high incidence of maternal depression in women with premature infants, these women rarely seek professional assistance for their condition and are often unaware of their symptoms or minimize them.46
Strengths of this study include the use of an RCT to test the effects of this theoretically based COPE program in >1 institution, with a comparison intervention that controlled for the time and attention being spent with the parents in the COPE program. In addition, both parent self-report and observational measures were used to evaluate the efficacy of the COPE program on key outcome variables to enhance validity of the findings. Although significant group differences for fathers' involvement in their infants' care resulted from staff nurse observers who were not trained on interrater reliability, the findings from the nurse ratings converge with other study findings.
Another strength of this study is the inclusion of fathers. Although fathers did not demonstrate the same significant group differences on emotional outcomes as the mothers in this study, there were important group differences found in fathers' functional outcomes in the NICU (ie, interaction with their infants) and their beliefs about their ability to parent their premature infants. Additional modeling and analyses of more long-term father outcomes will be critical to better understand the impact of the intervention on fathers over time.
Finally, the inclusion of the LOS outcome was important to provide hospital administrators with cost data necessary to make informed decisions regarding whether an intervention program can be feasibly implemented in hospital settings.
One of the limitations of this study is that it was conducted in only 2 level 3 NICUs located in hospitals within the northeast region of the United States. Potential differences in practice may exist in other institutions that may influence generalizability of our findings. Another potential limitation that should be noted is that this sample includes relatively healthy and larger premature infants (∼60% of the infants were >1500 g). However, research has indicated that even larger and healthier preterm infants are likely to be treated differently by their parents on the basis of their perceptions of their preterm infant termed “prematurity stereotyping.”47 Even having a relatively well infant in the NICU can be emotionally disturbing.48
Additional analyses are planned to determine if the effects of the COPE intervention were moderated by gestational age and birth weight, especially because 70.8% of the infants in this study weighed <2000 g at birth and 38.5% were of VLBW (<1500 g). Other research has found that effects of interventions with parents that commenced after NICU discharge were most effective with parents of infants in the 2000- to 2500-g range compared with parents of infants <2000 g.49 Testing COPE with parents of even younger, smaller, and sicker infants as well as parents of multiple births is now needed to determine if the intervention has more potent effects with more vulnerable populations in whom anxiety and stress levels as well as more negative developmental outcomes are known to be higher.50,51 Future analyses also need to determine the influence of moderating variables on this intervention's effects, because research has indicated that some interventions are most effective in families with few resources.52
Effects of the intervention are currently available only through 2 months' corrected infant age. Findings based on longer-term outcomes to measure the potential sustained positive effects of the intervention will be examined. Finally, further understanding of how the COPE program exerts its effects is required. Although it is believed that parental beliefs are a potent mediator of the effects of the COPE intervention, future analyses must be conducted to support this hypothesis.
Interventions to enhance coping and mental health outcomes in parents of LBW premature infants to improve both parent and child health outcomes have lagged behind the rapid technologic advances to sustain survival in NICUs. However, evidence now has accumulated to support that the COPE program not only improves parent coping and mental health outcomes during, as well as shortly after, hospitalization but also reduces hospital LOS.
Continued study of these infants and their parents is currently underway, through 3 years of age, to determine the more long-term effects of the COPE program. Although this study demonstrates considerable cost savings during the NICU stay, sustainable long-term effects could result in additional savings to a burdened health care system. More important is that the COPE program facilitates the development of critical parenting skills that are needed to shape the next generation of a high-risk group of families.
This study was supported by National Institutes of Health/National Institute of Nursing Research grant R01 05077.
We are grateful to Jennifer Anderson, PhD, Karen Cavallo Szewczyk, RN, NNP, Yolanda DeCamilla, RN, Diane D’Eugenio, OT, Sharon Dowd, RN, NNP, Linda Grady, RN, Sara Heidelberger, RN, NNP, Diane Hust, RN, PNP, Karen Juszek, RN, NNP, Christine Kowaleski, RN, NNP, Rachael Kraus, RN, PNP, Lisa Mitchell, OT, Lisa Spath, RN, WHNP, and Jennifer West, PhD, for assistance with data gathering and input at various times during the process.
- Accepted June 8, 2006.
- Address correspondence to Bernadette Mazurek Melnyk, PhD, RN, FAAN, Arizona State University College of Nursing, 500 N 3rd St, Phoenix, AZ 85004. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
As principal investigator, Dr Melnyk had full access to all of the data in the study and takes responsibility for the integrity of the data and accuracy of the data analysis. Drs Melnyk, Alpert-Gillis, Feinstein, Crean, and Gross came up with the study concept and design; Dr Feinstein, Ms Fairbanks, and Drs Sinkin, Small, and Gross acquired the data; Drs Melnyk, Feinstein, and Alpert-Gillis, Ms Fairbanks, and Drs Crean, Sinkin, Tu, and Gross analyzed and interpreted the data; Drs Melnyk, Feinstein, Fairbanks, and Crean drafted the manuscript; Drs Melnyk, Feinstein, and Alpert-Gillis, Ms Fairbanks, and Drs Crean, Sinkin, Stone, Tu, and Gross critically revised the manuscript for important intellectual content; Drs Melnyk and Feinstein, Ms Fairbanks, and Drs Crean, Sinkin, and Tu performed statistical analysis; Drs Melnyk, Alpert-Gillis, and Feinstein obtained funding; Drs Melnyk, Feinstein, and Alpert-Gillis, Ms Fairbanks, and Drs Crean and Gross provided administrative, technical, or material support; and Drs Melnyk, Alpert-Gillis, and Feinstein, Ms Fairbanks, and Dr Gross provided study supervision.
The funding agency was not involved in the design or conduct of the study; in the collection, analysis, or interpretation of the data; or in the preparation of the data or preparation, review, or approval of the manuscript.
- ↵Hamilton BE, Martin JA, Sutton PD; Centers for Disease Control and Prevention, National Center for Health Statistics. Births: preliminary data for 2003. Natl Vital Stat Rep.2004;53(9) :1– 17
- O'Brien F, Roth S, Stewart A, Rifkin L, Rushe T, Wyatt J. The neurodevelopmental progress of infants less than 33 weeks into adolescence. Arch Dis Child.2004;89 :207– 211
- Saigal S, Ouden LD, Wolke D, et al. School-age outcomes in children who were extremely low birth weight from four international population-based cohorts. Pediatrics.2003;112 :943– 950
- ↵Altimier L, Eichel M, Warner B, Tedeschi L, Brown B. Developmental care: changing the NICU physically and behaviorally to promote patient outcomes and costs. Neonatal Intensive Care.2004;17 :35– 39
- Leijon I, Finnstrom O, Sydsjo G, Wadsby M. Use of healthcare resources, family function, and socioeconomic support during the first four years after preterm birth. Arch Dis Child Fetal Neonatal Ed.2003;88 :F415– F420
- Petrou S, Mehta Z, Hockley C, et al. The impact of preterm birth on hospital inpatient admissions and costs during the first 5 years of life. Pediatrics.2003;112 :1290– 1297
- ↵Rogowski J. Using economic information in a quality improvement collaborative. Pediatrics.2003;111(4) . Available at: www.pediatrics.org/cgi/content/full/111/4/SE1/e411
- ↵Pinelli J. Effects of family coping and resources on family adjustment and parental stress in the acute phase of the NICU experience. Neonatal Netw.2002;19(6) :27– 37
- ↵Miles MS, D'Auria J. Parenting the medically fragile infant. Capsules Comments Pediatr Nurs.1994;1 :2– 14
- ↵Wereszczak J, Miles MS, Holditch-Davis D. Maternal recall of the neonatal intensive care unit. Neonatal Netw.1997;16(4) :33– 40
- ↵Minde K. Prematurity and serious medical conditions in infancy: implications for development, behavior, and intervention. In: Zeanah CH Jr, ed. Handbook of Infant Mental Health. 2nd ed. New York, NY: Guilford Press; 2000:176– 194
- ↵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
- ↵Johnson JE, Fieler VK, Jones LS, Wlasowicz GS, Mitchell ML. Self-Regulation Theory: Applying Theory to Your Practice. Pittsburgh, PA: Oncology Nursing Press; 1997
- ↵Lorion RP, Price RH, Eaton WW. The prevention of child and adolescent disorders: from theory to research. In: Schaffer D, Phillips S, Enzer NB, eds. Prevention of Mental Disorders, Alcohol and Other Drug Use in Children and Adolescents. Rockville, MD: US Office of Substance Abuse Prevention; 1989:55– 96. DHHS publication No. ADM 89-1646
- ↵Spielberger C, Gorsuch R, Lushene R. The STAI Manual. Palo Alto, CA: Consulting Psychologists; 1977
- ↵Beck AT, Steer RA, Brown GK. Beck Depression Inventory: 2nd ed. Manual. San Antonio, TX: Psychological Corporation: 1996
- ↵Miles MS. Parents of critically ill premature infants: sources of stress. Crit Care Nurs Q.1989;12(3) :69– 74
- ↵Melnyk B, Feinstein NF, Fairbanks E, Small L. The Index of Parental Behavior in the NICU. Rochester, NY: University of Rochester; 1998
- ↵Melnyk BM. The Parental Belief Scale: NICU. Rochester, NY: University of Rochester; 1997
- ↵Hollis S, Campbell F. What is meant by intention to treat analysis? Survey of published randomized controlled trials. BMJ.1999;319 :670– 674
- ↵DiCenso A, Guyatt G. The principle of intention to treat. In: DiCenso A, Guyatt G, Ciliska D, eds. Evidence-Based Nursing: A Guide to Clinical Practice. St Louis, MO: Elsevier Mosby; 2005:245– 250
- Melnyk BM. Coping with unplanned childhood hospitalization: the mediating functions of parental beliefs. J Pediatr Psychol.1995;20 :299– 312
- ↵Melnyk BM, Alpert-Gillis L, Feinstein NF, et al. Creating Opportunities for Parent Empowerment: program effects on the mental health/coping outcomes of critically ill young children and their mothers. Pediatrics.2004;113(6) . Available at: www.pediatrics.org/cgi/content/full/113/6/e597
- ↵Ugarrizza DN. The assessment and treatment of post-partum depression. In: Melnyk BM, Fineout-Overholt E, eds. Evidence-Based Practice in Nursing and Healthcare. Philadelphia, PA: Lippincott, Williams & Wilkins; 2004:19– 26
- ↵Gennaro S, Brooten D, Roncoli M, Kumar SP. Stress and health outcomes among mothers of low-birth-weight infants. West J Nurs Res.1993;15 :97– 113
- Copyright © 2006 by the American Academy of Pediatrics