|Home | About | Journals | Submit | Contact Us | Français|
To describe the outcomes of children with sickle cell disease (SCD) after discharge from medical care for vaso-occlusive painful events and to test the hypothesis that older age, longer length of hospital stay, and a history of frequent vaso-occlusive painful events will be associated with poor outcomes.
Children ages 2 to 18 years with SCD treated in the emergency department or inpatient unit for a painful event were contacted after discharge to assess: days of pain, days of functional limitations for the child, and days of work/school absenteeism for the caregiver. Descriptive statistics were applied and multivariate logistic regression examined the association between the predictors and outcomes.
Fifty-eight children were enrolled (mean age 10.8±4.8 years, 53.5% female). Post-discharge, 46.5% of children reported three or more days of pain, 54.3% had two or more days of functional limitations, and 24.3% of caregivers missed two or more days of work/school. Children with three or more prior painful events had increased odds of a poor outcome post- discharge (OR 1.79; 95% CI=1.026,3.096).
Acute vaso-occlusive painful events impact the lives of children with SCD and their caregivers, even after discharge to home.
Sickle cell disease is the most common inherited blood disorder in the United States (DeBaun & Vichinsky, 2007). The pathophysiology of the disease is complex and results in broad clinical manifestations and chronic multi-organ disease. Children with sickle cell disease require frequent emergency department visits and hospitalizations to manage pain and other complications, however, sickle cell disease also has an impact on children’s lives at home (DeBaun & Vichinsky, 2007; NIH Pub No.02-2117, 2002; Dampier et al, 2004; Shapiro et al, 1995; Dampier et al, 2002a; Ely et al, 2002; Dampier et al, 2002b). Children with sickle cell disease report poor health-related quality of life in their baseline state of health (Panepinto et al, 2004). In addition, from home pain diary studies, we have learned that painful episodes that occur at home are underreported and impact children’s school attendance (Shapiro et al, 1995; Gil et al, 2000; Dampier et al, 2002a; Ely et al, 2002; Dampier et al, 2002b). This is important in children with sickle cell disease because they are at risk of poor academic attainment (Wang et al, 2001; Steen et al, 2005; Herron et al, 2003; Schatz et al, 2004; King et al, 2008; King et al, 2006) These data underscore the impact that sickle cell disease has on children’s lives at home.
In 2003, there were over 20,000 hospitalizations for children with sickle cell disease and over 16,000 of these were for vaso-occlusive painful events (Whalen et al, 2006). Despite population-based data showing that over 65,000 hospital days per year are for painful events, how these events affect the function of the child and caregiver at home immediately after discharge is not known (Whalen et al, 2006; Panepinto et al, 2005). Since painful events contribute to a significant proportion of the morbidity of the disease, studying the outcomes of these events is crucial to further characterize the impact of sickle cell disease on the child and family.
The objective of this study was to describe the outcomes of children with sickle cell disease after discharge from medical care for vaso-occlusive painful events. Our primary hypothesis was that after children with sickle cell disease are discharged from medical care for an acute vaso-occlusive painful event, there are unrecognized poor outcomes including persistent pain and functional limitations for the child and absence from work/school for the caregiver. Our secondary hypothesis was older age, increased length of hospital stay, and a history of frequent vaso-occlusive painful events will be associated with poor outcomes.
This was a prospective cohort study conducted from June 2006 through May 2007. Children ages 2 to 18 years with sickle cell disease who presented with an acute vaso-occlusive painful event to the Medical College of Wisconsin’s pediatric emergency department were eligible for enrollment. Exclusion criteria included children with sickle cell disease presenting with fever only, stroke, splenic sequestration, those on chronic transfusions, non-English speaking patients, those without access to a telephone, or previous enrollment in this study. Children were consented in the emergency department at presentation or within 12 hours of admission and were contacted at three and seven days post-discharge to assess outcomes. The decision to admit the child to the inpatient unit or discharge the child from the emergency department was at the discretion of the treating physician. Criteria for discharge from the inpatient unit included pain controlled on oral medications.
The institutional review board of the Children’s Hospital of Wisconsin/Medical College of Wisconsin approved the study and informed consent was obtained from the parent or legal guardian and assent from the child when appropriate.
Demographic and medical information were obtained on all children through parental report and from the medical record. Three patient/caregiver-reported outcomes were assessed post-discharge: 1) days of pain, 2) days of functional limitations for the child, and 3) days of work/school absenteeism for the caregiver. We used a priori criteria to designate each outcome variable as a poor outcome (see definitions below). These criteria were based on data from prior outcomes work done in other childhood illnesses such as asthma and fever since there has not been any similar work done in sickle cell disease (Stevens & Gorelick, 2001; Mistry et al, 2007). The outcomes were dichotomized in order to make the results more interpretable and clinically useful for clinicians in identifying a high risk cohort that could be targeted for interventions. We also assessed days of pain before seeking medical care, use of pain medications at home in the days immediately after discharge, and unplanned relapse to care to the emergency department, inpatient unit, or primary medical doctor within seven days of discharge. The outcomes were assessed via telephone interviews at both three and seven days post-discharge from the emergency department or inpatient unit to minimize recall bias. Interviews were conducted using a structured questionnaire with the caregiver and/or the child when appropriate.
The pain outcome was defined as the number of days of self-reported pain present since discharge. The a priori criteria used to designate this variable as a poor outcome was the presence of three or more days of pain post-discharge.
Functional limitations for the child were defined three ways to capture different situations children may be in at the time of study enrollment. First, if the child was enrolled in school, this variable represented the actual number of days of school missed. Secondly, if the child was too young for school but attended daycare, this variable represented the number of days of daycare missed. Lastly, if the child was not in school due to no school scheduled at the time of data collection (summer months, weekends, fall/spring break, etc.) or not attending school for any other reason, this variable represented the number of days a child was unable to play or perform usual daily activities. This variable was modeled after outcomes work done in other childhood illnesses such as asthma and fever (Stevens & Gorelick, 2001; Mistry et al, 2007). The a priori criteria used to designate this variable as a poor outcome was two or more days of functional limitations post-discharge.
Work/school absenteeism for the caregiver was defined as the number days of work or school missed post-discharge, when applicable. The a priori criteria used to designate this variable as a poor outcome was two or more days of work/school absenteeism post-discharge.
We examined age (measured in years), length of hospital stay (measured in days), and the number of previous vaso-occlusive painful events requiring inpatient admission in the year prior to study enrollment as predictor variables for poor outcomes. Data for number of previous hospitalizations for vaso-occlusive painful events were obtained from a comprehensive database maintained by our sickle cell center of all inpatient events, were not reliant on self-report, and thus were not subject to recall bias. Age and length of stay were evaluated as continuous variables. The number of previous admissions for painful events was dichotomized to less than three events per year or three or more events per year based on clinical criteria generally accepted by hematologists for severe sickle cell disease interventions such as hydroxyurea and bone marrow transplantation (NIH Pub No.02-2117, 2002; Charache et al, 1995; Steinberg et al, 2003; Scott et al, 1996; Ferster et al, 1996; Panepinto et al, 2007).
Descriptive statistics were calculated for all demographic and outcome variables. These included mean with standard deviation, median with interquartile range, and proportions, when appropriate. Although the outcome variables were non-normally distributed, means and standard deviations are displayed along with medians and interquartile ranges for clinical utility and interpretability. Mann-Whitney U Test was used to compare differences in outcomes among various sickle cell genotypes (SS, SC, SB+thal).
Multivariate logistic regression analysis was used to examine the effect of the primary predictor variables: age, length of stay, and the number of previous vaso-occlusive painful events on the outcomes. However, since the use of hydroxyurea and the interaction between age and previous vaso-occlusive events (Panepinto et al, 2005) could also affect the outcome, we examined the effect of both of these variables on the outcomes. It is not known whether hydroxyurea moderates the severity of a single painful event, thus the inclusion of hydroxyurea as a covariate in the model was important. For the regression analysis, we dichotomized our outcome into two groups: poor overall outcome or good overall outcome to make the data more clinically useful. A child was designated as having a poor overall outcome if he or she met the a priori criteria for a poor outcome (as previously defined) in one or more of the outcomes assessed (pain, functional limitations, or work/school absenteeism). All other children were designated as having a good overall outcome. A p-value of ≤ 0.05 was considered statistically significant. Analysis was conducted with SAS, version 9.1.3 (SAS Institute, Cary, NC).
Since the true proportion of children with poor outcomes is unknown, to determine sample size, we hypothesized that 30% of children would meet the a priori criteria for a poor outcome for each variable of interest (pain, functional limitations, work/school absenteeism). Therefore, in a two-sided analysis, at a confidence level of 0.95, expected proportion of 0.30, and distance from proportion of 0.15, a sample size of 36 children was needed.
A total of 95 eligible children presented to the emergency department during the study period. Of those eligible children, a total of 74 children (78%) were approached, 61 children (64%) were enrolled and 13 children (14%) refused to participate. Reasons children were not approached included legal guardian not present or missed opportunities. Reasons for refusal included “already in a research study,” “don’t like studies,” and “just not interested.” The enrolled and non-enrolled children did not differ by age or gender. Three children were lost to follow-up. Thus our final study population included 58 children and the demographics for this sample are shown in Table I.
The mean days of pain post-discharge was 2.53 (±2.01 SD) days (median of 2 days; IQR: 1.0–3.25). Figure 1 depicts the proportion of patients that experienced days of persistent self-reported pain post-discharge. Almost half of the children reported three or more days of pain post-discharge (Table II). In addition, 44.8% of children reported continued use of pain medication (ibuprofen, tylenol with codeine, oxycodone, and other) at three days post-discharge and 19% of children reported use of pain medication at seven days post-discharge. There were 9 children (14.8%) that reported unplanned return to care for pain within seven days post-discharge and of these children, 4 (6.6%) required readmission to the inpatient unit for pain control. The days of pain did not differ based on genotype of sickle cell disease (data not shown).
There were 35 children (60.3%) who were enrolled in school or daycare at the time of data collection. There were 23 children (39.7%) who were not actively enrolled in school or daycare or whom data collection occurred when school days were not scheduled. Age, reported days of pain, and days of functional limitations did not statistically differ between those attending and not attending school/daycare at the time of data collection (data not shown). The mean days of functional limitations post-discharge was 2.17 (±1.79 SD) days (median of 2 days; IQR: 1.0–3.0). Figure 1 depicts the proportion of children that experienced persistent days of functional limitations post-discharge. Over half of the children reported three or more days of functional limitations post-discharge (Table II). The days of functional limitations did not differ based on genotype of sickle cell disease (data not shown).
There were 33 caregivers (56.9%) that reported working or attending school at the time of study enrollment. The mean days of work/school absenteeism was 1.15 (±1.20 SD) days (median of 1 day; IQR: 0–1.5). Figure 1 depicts the proportion of caregivers that experienced persistent days of work/school absenteeism post-discharge. Almost one-fourth of caregivers reported missing two or more days of work/school post-discharge (Table II). The days of work/school absenteeism did not differ based on genotype of sickle cell disease (data not shown).
Children with a history of three or more painful events requiring inpatient admission in the year prior to study enrollment had significantly increased odds of a poor overall outcome, independent of the other predictor variables. Unexpectedly, older age (continuous variable) was associated with decreased odds of a poor overall outcome. Therefore, we further analyzed pain and functional limitations in older (≥ 9 years) versus younger (< 9 years) children. We found older and younger children reported similar days of pain post-discharge (median=2 in both groups), however, older children reported less days of functional limitations post-discharge than younger children (median=1 vs. 2, respectively). Length of hospital stay, use of hydroxyurea, and the interaction between age and previous vaso-occlusive painful events did not have a statistically significant effect on the overall outcome. There was no difference in poor overall outcome between the children discharged from the emergency department versus those discharged from the inpatient unit (data not shown).
Acute vaso-occlusive painful events impact children with sickle cell disease and their caregivers. The morbidity of a painful event that brings a child to medical attention does not end once the child is discharged to home. The data from our study show that children experience persistent pain at home after discharge from medical treatment for vaso-occlusive painful events and this pain likely contributes to further functional limitations, including school absenteeism. Our findings further expand the previously documented relationship between pain experienced at home and school absenteeism (Shapiro et al, 1995; Gil et al, 2000; Dampier et al, 2002a; Dampier et al, 2002b). In addition, our findings show that children have almost two days of pain at home prior to seeking medical care and even seeking medical care in the emergency department or inpatient unit does not mean the end of a painful episode. These data underscore the fact that sickle cell pain that occurs at home is a problem that disrupts school attendance and likely other activities of daily living.
The impact of vaso-occlusive painful events at the school level is not trivial. Previous studies have shown that children with sickle cell disease are at risk of poor academic attainment (Wang et al, 2001; Steen et al, 2005; Herron et al, 2003; Schatz, 2004; King et al, 2008; King et al, 2006). Poor academic attainment in children with sickle cell disease is a multifactorial problem; nevertheless, the number of missed school days while children are hospitalized for painful events has been shown to be predictive of poor academic attainment (Herron et al, 2003; Schatz, 2004; Taras & Potts-Datema, 2005). In our study, on average, children were hospitalized for three days resulting in school absenteeism with an additional two days of school absenteeism post-discharge. Therefore, a single painful event can result in five missed school days. For children with frequent painful events, who also had the worst outcomes in our study, painful events alone have an impact on their ability to attend school. In addition, these data do not account for other issues that may impact the school attendance of children with sickle cell disease such as admissions for fever or other sickle cell disease related complications, and admissions for other non-sickle cell disease related acute or chronic medical conditions, such as asthma.
We also found that vaso-occlusive painful events affect the caregiver’s ability to work or attend school which can in turn impact the socioeconomic status of the family. This is extremely important for children with sickle cell disease because they are more likely to live in poverty (Barbarin et al, 1999; Hill, 1994), and many caregivers of these children are single parents, thus they need to work outside the home because they are the sole provider for the entire family (Hill, 1994). In addition, people of lower socioeconomic status are more likely to have jobs that do not give them sick leave or flexible work hours that allow them to care for a sick child (Smith et al, 2002). Caregivers of children with chronic diseases have identified high child health care use leading to missed work days as a barrier to employment, thus parents of children with chronic medical conditions are more likely to be unemployed (Smith et al, 2002; Kuhlthau & Perrrin, 2001). For children with sickle cell disease, the unpredictability of painful events can lead to multiple absences from work for caregivers which could ultimately lead to unemployment. Thus, when a chronic disease, such as sickle cell disease, is introduced to an already impoverished family, the impact of missed work and potential unemployment on socioeconomic status is crucial to the welfare of the entire family. Our study illustrates that a single painful event results in missed days of work/school for caregivers and for those children with frequent painful events, this impact is magnified and could further negatively impact the socioeconomic status of the entire family.
When new interventions are developed, it is important to identify high risk patient groups that could benefit from these interventions. Our data show that children with a history of three or more vaso-occlusive painful events requiring inpatient admission in the year prior to study enrollment had significantly increased odds of experiencing poor outcomes post-discharge. Therefore, our data provide further evidence that these children make up a high risk cohort and should be targeted for interventional studies to improve their outcomes in the post-discharge period and novel new therapies to treat acute events such as intravenous magnesium (Brousseau et al, 2004). In addition, these children should be evaluated for known preventive therapies for pain such as hydroxyurea (NIH Pub No.02-2117, 2002; Charache et al, 1995; Steinberg et al, 2003; Scott et al, 1996; Ferster et al, 1996), bone marrow transplantation (Panepinto et al, 2007), chronic blood transfusions, or other future clinical trials developed to decrease the frequency of painful events.
Interventional studies for vaso-occlusive painful events have used length of hospital stay as an outcome endpoint (Brousseau et al, 2004; Weiner et al, 2003). However, our study shows the morbidity of painful events does not end the day a child is discharged to home. Using length of hospital stay as a primary outcome measure of a vaso-occlusive painful event does not account for what happens when the child is home. Therefore, when measuring the effectiveness of new interventions for vaso-occlusive painful events, it is imperative to assess the morbidity children experience at home in the post-discharge period in addition to length of hospital stay. In addition, this study provides a model for hospital follow-up that could be utilized in a multi-center observational study aimed at longitudinally assessing the outcomes of painful events.
Unexpectedly, older age portended slightly decreased odds of experiencing poor outcomes post-discharge. Since the number of vaso-occlusive painful events and length of hospital stay has been shown to increase with age (Panepinto et al, 2005) we hypothesized older children would have worse outcomes, the opposite of what we found in our study. Upon further analysis of pain and functional limitations in older (≥ 9 years) versus younger (< 9 years) children, we found similar days of pain post-discharge in both groups, however, older children had fewer days of functional limitations (including school absenteeism) post-discharge than younger children. This suggests that perhaps older children have adapted or adjusted to the pain from their disease, thus reintegrating themselves into their lives sooner than younger children. These data could also suggest that younger children may miss more school as a function of greater influence of parental decision making on their return to school. Finally, these data may suggest that older children have more chronic pain that follows the epidemiological pattern of pain in adults that has recently been described (Smith et al, 2008). This observation warrants additional longitudinal work to further elucidate the affect of age on the outcomes of individuals.
This study is limited by data that are self-reported. Recall bias of self-reported days of pain prior to presentation is also a potential limitation. In addition, the outcomes assessed post-discharge are also subject to recall bias. However, this was likely minimized by contacting families at two points in time post-discharge. Like many studies, this sample reflects the experience of a single center and outcomes may vary from center to center. Therefore, multi-institutional studies would be important to evaluate the pattern of outcomes at other centers. Our study did not find a difference in overall outcome between children discharged from the emergency department versus those discharged from the inpatient unit and did not find a difference in outcomes based on genotype of sickle cell disease. However, this study was not powered to find such differences and a larger study would be required to test this hypothesis. In addition, our sample size only allowed for dichotomization of our outcome into two groups (good and poor). A larger study would allow dichotomization into additional groups, such as a moderate outcome category. There is currently no standard classification for severity of sickle cell disease and there is inherent variability of the severity of an individual painful event. This lack of severity classification and lack of a validated risk adjuster for length of pain in sickle cell disease make risk adjusted analysis in the study difficult, therefore, this remains a limitation to the study. Finally, this study assessed a single vaso-occlusive painful event per child and was not designed to be a longitudinal study to follow children over time.
In conclusion, we provide new evidence that vaso-occlusive painful events impact the function of children with sickle cell disease and their caregivers even after discharge from medical care. Children experience poor outcomes at home that manifest as persistent pain and functional limitations for the child, including school/daycare absenteeism, and work/school absenteeism for the caregiver. A history of three or more vaso-occlusive painful events is a risk factor for persistent pain, functional limitations for the child and missed days of work/school for the caregiver. Therefore, this study provides further evidence that these children make up a high risk cohort and should be targeted for medical, social, and school interventions. In addition, when examining the impact of interventional studies on vaso-occlusive painful events, post- discharge assessment is necessary in determining the length of painful events. Finally, interventions to improve outcomes in the post-discharge period are imperative to improve the quality of life of children affected by this chronic debilitating disease.
Funding: This work was supported in part by a grant from the National Heart, Lung, and Blood Institute-K23 HL80092 (JP).
The authors would like to thank Aniko Szabo, PhD and Shuyuan Mo, both in the Division of Biostatisics at the Medical College of Wisconsin, for their statistical assistance in preparation of this manuscript.
Conflicts of Interest: None for any of the authors