Intraindividual Pain Variability Metrics for Youth with Sickle Cell Disease: Relations to Health Outcomes

Abstract

Background:

While the majority of pediatric sickle cell disease (SCD) research has used mean pain intensity as the only pain metric, recent evidence suggests this metric alone is inadequate in describing the intraindividual variability in SCD pain experiences and subsequent impact. There is limited information on other intraindividual pain metrics in youth with SCD, or how they relate to health outcomes in this population. The aims of this study were to describe differing patterns of intraindividual pain metrics derived from ecological momentary assessments (EMAs) of youth with SCD and to characterize the unique relationships between these metrics and health outcomes.

Methods:

Eighty-eight youth with SCD aged 8 to 17 (mean age = 11.6) were recruited from three regional pediatric SCD clinics in the US. At baseline, youth and their guardians reported on demographic and disease information. Then youth completed twice daily EMAs for up to 4 weeks. Pain metrics derived from EMA data were calculated including mean daily pain intensity (DP), SD-DP (standard deviation of DP), proportion of pain days (PPD), and 90^th percentile of DP (p90). Pearson correlations were calculated between pain metrics and health outcomes.

Results:

High DP and SD-DP were correlated with more anxiety symptoms, while high SD-DP and p90 were correlated with more depression symptoms. High SD-DP was correlated with low self-esteem, and high DP and PPD were correlated with low sickle cell self-efficacy. For healthcare utilization due to pain, high p90 was correlated with more emergency department visits, while high DP, p90, and PPD were correlated with more health care contacts.

Conclusion:

There are distinct associations between pain variability metrics beyond DP and health outcomes. Collectively, the patterns of associations suggest the utility of these pain metrics for determining risk in relation to specific health outcomes for youth with SCD.

Background

Pain in children and adolescents with SCD, especially painful vaso-occlusive crises (VOC), lead to disability, suffering, and high healthcare utilization.^1–3 One of the earliest discovered predictors of mortality in SCD was the annual “pain rate,” defined as the hospitalization rate due to VOC.⁴ Most pain is managed at home, resulting in only the most severe VOCs presenting to medical attention.⁵ The annual pain rate does not consider that SCD pain experiences vary widely nor that the majority of sickle cell pain is experienced outside of the context of hospitalizations. To better consider the broader nature of sickle cell pain, which varies so much from day to day within and between individuals, researchers have tried to develop methods to comprehensively describe sickle cell pain over a specified period of time (e.g., 6 months to 1 year), using either frequency in days or mean pain intensity.⁶ They have then tried to relate either of these SCD pain measures or utilization due to VOC to health-related quality of life⁷, daily mood and stress⁸, or coping.⁹ The implied intent, to conceptualize or define SCD pain phenotypes using summary pain measures^10,11 and then use them to predict pain-related outcomes or design treatment interventions has met with only modest success based on recent reviews or treatment guidelines.^12,13 Describing SCD pain using these gross summary measures is simply not detailed enough to improve treatment, create better prognoses, or increase our understanding of mechanistic processes, information essential for guiding our intervention and prevention efforts. What is needed are better systems for describing and phenotyping SCD pain, an area of intense investigation.^14–18

One novel descriptive method investigators studying adults with SCD have begun to explore is the use of intraindividual variability metrics of SCD pain over time.¹⁹ Intraindividual variability of pain refers to an individual’s dynamical pain experience and how one’s individual experience of pain fluctuates over time.²⁰ Studying intraindividual pain variability could better characterize how pain relates to impairment. A typical intraindividual measurement strategy employs repeated assessment of pain-relevant variables in real-time, natural settings, a form of ecological momentary assessment (EMA)^20,21 EMAs of pain using electronic diaries allows for the investigation of daily fluctuations in self-reported pain. It is exceptionally positioned in evaluating a patient’s pain experience over time with high precision. EMA provides rich data on the temporal dynamics of pain and the effects of the specific contexts of a patient’s pain experience.^22,23 The majority of EMA studies in SCD research have utilized mean pain intensity as the sole indicator to describe pain patterns and fluctuations. However, patients with SCD have significant variability in pain experiences, including in how pain levels are spread out across the sample (i.e., interindividual) and how pain fluctuates among individuals over time (i.e., intraindividual).^19,20,24 EMA, therefore, attempts to fully capture the variability of SCD pain, to better delineate how dynamic SCD pain can be.

A conceptual review described different statistical metrics of intraindividual variability that can be calculated from EMA reports and that may encapsulate more refined and important aspects of pain examined on a daily basis for various chronic pain conditions, including SCD.²⁰ Table 1 provides examples of EMA derived pain metrics representing daily pain severity, variability, and frequency. To date, there is limited information on EMA derived intraindividual pain variability metrics in youth with SCD and how these metrics may relate to health outcomes. This is of interest given how the severity and frequency of pain vary greatly among youth with SCD.^3,25 Also, typically beginning in adolescence or early adulthood, pain becomes more frequent and severe and often transforms into chronic pain.^4,25 Thus, childhood may be a particularly salient developmental period for understanding and evaluating intraindividual variability metrics of SCD pain and identifying how they may uniquely relate to health outcomes. The aims of this study are to describe different patterns of pain intraindividual variability metrics derived from EMA reports of youth with SCD and to characterize the unique relationships between these metrics and health outcomes.

TABLE 1.

Pain Metrics in Current Study

	Pain Metric	Description
DP	Mean Daily Pain Intensity	Mean amount of daily pain severity reported across the diary period for each youth.
p90	90th percentile of pain	The point where on 10% of pain scores are above. Thus, it measures the upper levels or most severe pain usually experienced on a daily basis by each youth.
SD-DP	Standard deviation of Daily Pain	Standard deviation of DP scores. It is an indicator of how much daily pain varies for each youth.
PPD	Percentage of Pain Days	Percentage of any days where pain scores were above 0 (VAS > 0), providing the number of days with pain divided by the total number of reported days. PPD is an indicator of pain frequency.

Method

Recruitment

We conducted a secondary analysis of EMA data collected as part of a larger study investigating sleep, pain, and related factors in youth with SCD. Participants included families of youth (aged 8–17 years) receiving care at outpatient SCD clinics across three different clinics with independent centers in the southeastern US. A detailed description of recruitment and assessment procedures can be found for the primary study in a previously published manuscript.²⁶ Eligibility criteria for the larger study included 123 youth (aged 8–17 years) with SCD having experienced at least one SCD pain episode in the past year (i.e., at least 20 min of SCD pain) and being English-speaking. Youth who had a diagnosis of a comorbid pain condition, a neurological impairment that would impede completion of the surveys, a history of extreme non-compliance as indicated by the healthcare provider, were on chronic blood transfusions, or were currently receiving a sleep intervention were excluded. For the current investigation, we also required that the youth had to have provided at least 2 weeks (24 entries) of daily EMA data. Out of the 123 participants in the larger study sample, 88 of them met the eligibility criteria. There were no differences in age, sex, SCD genotype, or whether the youth were prescribed hydroxyurea between the larger sample and the final sample used in the current investigation.

Procedure

The research was approved and monitored by the institutional review boards at all three sites. Data was collected between 2011 and 2015. Families were screened at their regularly scheduled SCD pediatric clinic visits and completed written parental informed consent and child assent forms. Youth and their guardians completed measures assessing their demographic and disease information (i.e., youth’s age, sex, race, and SCD genotype pubertal status), healthcare use related to pain (i.e., number of hospitalizations, doctor visits, and ER visits), and health outcome related measures (e.g., Behavior assessment system for children (BASC) and Sickle Cell-Efficacy Scale (SCSES)) at the study entry during the baseline interview. Youth were also asked to complete an EMA, which consisted of brief survey administered via an app, twice a day (i.e., in the morning and evening) for up to 4 weeks on either an Apple or Android device they owned or that was provided to them as part of the study.

Measures

Demographic and Disease Information

Guardians reported on the youths’ age, sex (0 = male and 1 = female), and SCD genotype. For the primary analyses, SCD genotype was coded as 1 for severe genotypes (e.g., HbSS and HbS/b0), and 0 for moderate genotypes (e.g., HbSC and HbSbþ, or other). Medical chart reviews were done to confirm SCD genotypes and to assess whether the youth were currently prescribed hydroxyurea. Of note, at the time of the study, hydroxyurea was prescribed only to patients with sickle cell presenting with more severe sickle cell-related complications, which was consistent with the CDC recommendations at that time.²⁷

Pubertal Development

Youth completed the 5-item Self-Rating Scale for Pubertal Development.²⁸ Changes in specific physical indicators of pubertal development are on a scale ranging from 1 = not yet started to 4 = seems complete. Boys and girls completed separate forms, with girls completing an item regarding whether they have begun menstruation, scored as no = 1 and 4 = yes. The scores are then averaged across items to produce an overall pubertal development score. This measure is reliable and has been validated against both parent and pediatrician reports of pubertal development.²⁸

EMA of Daily Pain

As part of the evening survey of the daily EMA, youth were asked to report whether they experienced any SCD pain during the day. If so, the youth were asked to rate their average pain severity that day on a 100mm horizontal visual analog scale (VAS) ranging from “not hurting at all” (0 mm) to “hurting a whole lot” (100 mm). If the youth indicated no pain on that day, their pain severity rating was automatically coded as 0. Research has provided support for the reliability and validity of the VAS in assessing daily SCD pain in children and adolescents.^29,30 The following pain variability metrics were calculated utilizing the daily pain rating severity : (1) mean daily pain intensity (DP); (2) the 90th percentile of DP scores (p90); (3) standard deviation (SD-DP) of DP scores; and (4) proportion of pain days (PPD). See Table 1 for an additional descriptions of the pain metrics.

Health Outcomes

Mental Health.

The Behavior assessment system for children, second edition (BASC-2;) was used to assess depression and anxiety symptoms. The BASC-2 is a validated and reliable standardized assessment that evaluates the behavior and emotions of individuals between 2 and 25 years of age. As part of it, youth completed the self-report BASC depression and anxiety subscales. The depression subscale assesses feelings of stress, sadness, and unhappiness and the ability to carry out everyday activities. The anxiety subscale assesses feelings of fearfulness, nervousness, or worry. T-scores (M=50, SD=10) on the depression and anxiety subscale were used, with subscale scores between 60 to 69 considered to be at risk, and scores greater than 70 considered to be clinically significant.

Intrapersonal Factors.

Self-reliance, self-esteem, and sickle cell self-efficacy were assessed as indicators of intrapersonal factors that aid in coping and managing sickle cell.^31–33 Youth completed the self-report BASC-2 self-reliance and self-esteem subscales. The self-reliance subscale assesses for a child’s confidence in their ability to solve problems and their belief in their personal decisiveness and dependability. The self-esteem subscale assesses feelings of self-acceptance, self-respect, and self-esteem. T-scores (M=50, SD=10) on the self-reliance and self-esteem subscale were used, with subscale scores between 31 to 40 considered to be at risk, and scores less than 30 considered to be clinically significant.

Also, adolescents (aged 12 to 17) completed the Sickle Cell-Efficacy Scale (SCSES), a nine-item scale developed specifically to assess adolescents’ and adults’ self-appraisals of their ability to participate in everyday activities associated with sickle cell disease.³⁴ The SCSES is widely used in sickle cell research and has shown high reliability and validity^34–36, with a Cronbach alpha of 0.77 in the present study.

Healthcare Utilization.

Guardians reported healthcare utilization during the Structured Pain Interview for Parents (SPI-P). To assess healthcare utilization due to pain within the past year, three items were used: (1) “How many times have your child gone to see a doctor?”, (2) “Number of times your child has been hospitalized”, and (3) “Number of emergency room visits”. The SPI-P measure has been used extensively for assessing health care use in youth with SCD and has acceptable interrater and test-retest reliability over a 9- to 12-month period.^30,37,38

Data Analysis Plan

Analyses were completed using R. Upon calculating pain metrics described in the Measure section, descriptives of demographic, disease characteristics, pain metrics, and health outcomes were calculated. Pearson correlations and t-tests were calculated to examine the various relationships between demographic and disease variables (i.e., age, sex, SCD genotype), pubertal development, pain variability metrics, and health outcomes. Specifically, Pearson correlations were calculated to examine how age and pubertal development were related to pain variability metrics and health outcomes. T-tests were calculated to examine group differences in pain variability and health outcomes based on sex, SCD genotype severity, and hydroxyurea use. To examine associations among pain variability metrics, Pearson correlations were calculated between the different variability metrics. Lastly, Pearson correlations were calculated to examine associations between pain variability metrics and health outcomes. Missing data were handled using list-wise deletion for t-tests and correlations. The healthcare utilization (HCU) variables were found to be skewed (all > 3) and to have high kurtosis (all > 13) and thus were log-transformed for Pearson correlation and t-test analyses. All other study variables appeared normally distributed.

Results

Descriptive statistics for demographic and disease characteristics, pain variability metrics, and health outcomes are summarized in Table 2. On average, youth were aged 11.66 years, and the majority of the sample were female (59%), currently prescribed hydroxyurea (54%), and has sickle cell genotype HbSS (50%). Age and pubertal development were not significantly correlated with pain variability metrics or health outcomes (p >.05 for all correlations). Differences in pain variability metrics and health outcomes based on gender, SCD genotype, and hydroxyurea use are reported in Table 3. No significant differences based on gender were found across pain metrics and health outcomes. Participants with a severe genotype reported lower self-reliance. Youth prescribed hydroxyurea reported lower self-reliance. No other significant differences based on genotype severity or hydroxyurea use were found.

TABLE 2.

Demographic and Disease Data (N=88)

Factor	n	%		N
Female	52	59%		88
Genotype
HBSS	44	50%		88
HBSC	27	31%		88
HbSβ+	12	14%		88
HbSβ0	3	3%		88
Other	2	2%		88
Hydroxyurea	47	54%		88
	Mean	SD	Range	N
Age (years)	11.66	2.99	8 to 17	88
Pubertal Development	2.23	0.91	1 to 4	88
Pain Metrics
DP	13.96	19.65	0 to 88.39	88
p90	34.44	37.38	0 to 100.00	88
SD-DP	17.41	14.22	0 to 49.57	88
PPD	21.85%	26.61%	0 to 100%	88
Mental Health
Anxiety	51.46	10.49	35 to 82.00	87
Depression	48.64	7.92	40 to 73.35	88
Transitional Factors
Self-esteem	53.99	6.39	33.27 to 62.00	87
Self-reliance	50.63	9.23	27 to 67	88
Sickle Cell Self-efficacy	30.58	6.37	18 – 41	40
Healthcare utilization in the past year due to pain
# of Hospitalizations	1.15	4.44	0 to 40	87
# of Doctor Visits	2.74	4.82	0 to 36	87
# of Emergency Depart Visits	2.05	2.64	0 to 12	87

Note. HBSS = Hemoglobin SS; HBSC = Hemoglobin SC; HbSβ⁺ = Hemoglobin Sβ⁺; HbSβ⁰=Hemoglobin Sβ⁰; DP = Daily pain; p90 = 90th percentile of pain; SD-DP = Standard deviation of Daily Pain; PPD = Percentage of pain days; Depart = Department. Descriptive statistics for healthcare utilization variables are prior to log transformations of those variables.

TABLE 3.

Comparison of Pain Metrics and Health Outcomes by Sex, SCD Genotype, and Hydroxyurea

Sex	Female M	Female SD	Male M	Male SD	t	df	p-value
Pain Metrics
DP	16.54	21.52	10.23	16.16	−1.49	86	0.14
p90	38.72	39.40	28.26	33.84	−1.30	86	0.20
SD-DP	18.40	14.68	15.96	13.59	−0.79	86	0.43
PPD	25.27%	30.11%	16.92%	19.92%	−1.46	86	0.15
Health Outcomes
Anxiety	52.90	11.26	49.31	8.96	−1.58	85	0.12
Depression	49.38	8.61	47.58	6.78	−1.05	86	0.29
Self-esteem	53.42	6.62	54.81	6.04	1.00	85	0.32
Self-reliance	50.60	8.95	50.67	9.75	0.04	86	0.97
Self-efficacy(N=40)	30.40	6.55	30.87	6.27	0.22	38	0.83
Emergency Department Visits	2.06	2.54	2.29	3.85	−0.37	85	0.71
Hospitalization	0.63	1.09	1.91	6.87	0.22	85	0.83
Doctor Visits	2.96	5.79	2.41	2.88	−0.13	85	0.90
Genotype	Severe M	Severe SD	Non-Severe M	Non-Severe SD	t	df	p-value
Pain Metrics
DP	11.81	15.76	16.42	23.30	1.10	86	0.27
p90	33.85	35.43	35.11	39.93	0.16	86	0.88
SD-DP	17.40	13.55	17.42	15.11	0.01	86	0.99
PPD	19.64%	22.73%	24.39%	30.55%	0.83	86	0.41
Health Outcomes
Anxiety	50.87	9.66	52.12	11.44	0.55	85	0.58
Depression	48.26	8.18	49.08	7.69	0.48	86	0.63
Self-esteem	53.77	7.15	54.24	5.49	0.34	85	0.73
Self-reliance	48.51	8.84	53.05	9.17	2.36	86	0.02
Self-efficacy(N=40)	31.33	6.04	29.44	6.86	−0.92	38	0.36
Emergency Department Visits	2.48	3.72	1.76	2.18	−0.42	85	0.67
Hospitalization	1.77	5.96	0.43	0.78	−1.42	85	0.16
Doctor Visits	2.78	6.11	2.70	2.69	1.69	85	0.09
HU Use	Yes M	Yes SD	No M	No SD	t	df	p-value
Pain Metrics
DP	12.95	17.74	13.86	20.45	0.22	85	0.82
p90	31.22	36.42	35.94	37.85	0.59	85	0.55
SD-DP	16.14	13.92	18.33	14.65	0.71	85	0.48
PPD	22.48%	27.47%	19.66%	23.70	−0.51	85	0.60
Health Outcomes
Anxiety	51.38	9.91	51.13	10.82	−0.11	84	0.91
Depression	48.15	7.89	48.76	7.82	0.36	85	0.72
Self-esteem	53.60	6.97	54.47	5.89	0.63	84	0.53
Self-reliance	47.73	9.35	52.96	8.56	2.72	85	0.01
Self-efficacy(N=40)	30.50	6.57	30.69	6.26	0.09	38	0.93
Emergency Department Visits	2.85	4.10	1.55	1.85	−1.11a	68	0.27
Hospitalization	2.05	6.51	0.40	0.85	−1.98a	63	0.05
Doctor Visits	3.12	6.64	2.40	2.59	1.27a	72	0.21

Notes. For genotype, severe = HBSS or HbSβ⁰; moderate = HBSC or HbSβ⁺ or other. Bolded p < 0.05

SCD = sickle cell disease; DP = Daily pain; p90 = 90^th percentile of pain; SD-DP = Standard deviation of Daily Pain; PPD = Percentage of pain days.

^{a =}Represents t-tests that violated homogeneity of variance assumption and completed a Welch t-test

Pearson correlations among pain variability metrics are summarized in Table 4. All of the pain variability metrics were positively correlated to one another. Pearson correlations among pain variability metrics and health outcomes are summarized in Table 5. For anxiety symptoms, High DP and SD-DP were correlated with more anxiety symptoms. No significant correlations were found between p90 and PPD and anxiety. For depression symptoms, high SD-DP and p90 were correlated with more depression symptoms. No significant correlations were found between DP and PPD and depression symptoms. For self-efficacy, high DP and PPD were correlated with low sickle cell self-efficacy. No significant correlations were found between p90 and SD-DP and sickle cell self-efficacy. Self-reliance was not significantly associated with any of the pain variability metrics. For self-esteem, high SD-DP was correlated with low self-esteem. No significant correlations were found between DP, p90, PPD, and self-esteem. No pain variability metrics were significantly associated with hospitalizations due to pain. For emergency department visits in the past year due to pain, high p90 were correlated with more emergency department visits. No significant correlations were found between DP, SD-DP and PPD and emergency department visits. For doctor visits due to pain in the past year, high DP, p90, and PPD were correlated with more healthcare contacts. No significant correlation was found between SD-DP and doctor visits.

TABLE 4.

Pearson Correlation of Pain Variability Metrics

	DP	p90	PPD	SD-DP
DP	---
p90	0.81**	---
PPD	0.94**	0.78**	---
SD	0.59**	0.84**	0.56**	---

Note.

^**p < 0.01;

SCD = sickle cell disease; DP = Daily pain; p90 = 90^th percentile of pain; SD-DP = Standard deviation of Daily Pain; PPD = Percentage of pain days.

TABLE 5.

Pearson Correlations of pain variability metrics and health outcomes

	DP	p90	SD-DP	PPD
Mental Health
Anxiety	0.22*	0.18	0.22*	0.18
Depression	0.17	0.26*	0.27*	0.17
Transitional Factors
Self-Efficacy (N=40)	−0.45**	−0.29	−0.09	−0.39*
Reliance	−0.14	−0.19	−0.20	−0.13
Self-Esteem	−0.11	−0.09	−0.22*	−0.06
Healthcare Utilization in the Past Year Due to Pain
Hospitalization	0.05	0.12	0.06	0.15
Emergency Department Visits	0.14	0.27*	0.11	0.20
Doctor Visits	0.27**	0.30**	0.18	0.23**

Notes.

^**p < 0.01;

^*p < 0.05;

DP = Daily pain; p90 = 90^th percentile of pain; SD-DP = Standard deviation of Daily Pain; PPD = Percentage of pain days.

Discussion

This is the first study to examine how indicators of intraindividual pain variability metrics beyond mean pain intensity alone relates to health outcomes in youth with SCD. Moreover, to date, there are only two studies utilizing EMA data to assess the intraindividual variability of pain in an SCD population.^19,24 In addition, this study has the largest pediatric SCD sample examining intraindividual pain variability, as one of the prior studies consists of an adult SCD population¹⁹, and the other with a sample size of 20 participants aged 13 to 21.

Consistent with the earlier study using a primarily adult sample, findings indicated pain variability metrics were related to worse health outcomes. In the current study, DP was related to more anxiety symptoms, doctor visits due to pain, and for the subgroup of adolescents, it was also related to low self-efficacy – spanning the different domains of health outcomes being examined. Of note, p90 was related to more symptoms of depression, and SD-DP was related to more symptoms of both anxiety and depression, and low self-esteem, suggesting patient’s worst levels/episodes of pain (i.e., p90) and fluctuation of pain (i.e., SD-DP) is particularly salient in understanding mental health among youth with SCD. In contrast, p90 and PPD were both related to more health care use due to pain (e.g., ER visits, and/or doctor visits), indicating patient’s worst levels/episodes of pain (i.e., p90) and pain frequency (i.e., PPD) may be particularly salient for understanding how youth and their families manage pain. These patterns of findings suggest the potential applicability of intraindividual pain metrics in informing pain interventions. For example, given how SD-DP is an indicator of pain fluctuation, interventions targeting a measurable reduction in SD-DP may in turn improve mental health-related outcomes. Conversely, given how PPD is an indicator of pain frequency and potential transition to chronic pain, interventions targeting a measurable reduction in PPD may in turn improve health care use-related outcomes.

These findings contrast with the earlier primarily adult study findings that found mean daily pain and SD-DP were both related to poor quality of life, high health care use, and high psychological dysfunction.¹⁹ This suggests differential patterns of interindividual variability of pain and how they relate to one’s health between youth and adult populations with SCD. For example, adults with sickle cell may experience greater dispersion and exacerbation of pain from one-time point to the next, resulting in strong associations found in the primarily adult study between SD-DP and a host of poor health outcomes. It may also be mean daily pain and SD-DP are more likely to relate to a host of health outcomes in a population where pain patterns are more indicative of chronic pain. Of the 139 participants in the primarily adult study, 55 (40%) participants reported pain greater than 50% of diary days.¹⁹ In contrast, only 11 (13%) participants in the current sample reported pain greater than 50% of diary days. Thus, it is possible that in a sample reporting more frequent pain, such as an adult SCD population, there may be more robust associations between intraindividual pain metrics and health outcomes across domains, compared to a sample reporting less frequent pain.

While the current study addresses a number of gaps in the literature, it also has a number of limitations that may have impacted the findings. First, though the overall sample size of the project was large in comparison to other pediatric SCD projects, a lack of variability of pain frequency limited specific analyses, such as stratifying the sample by the percentage of pain days (PPD). Research with larger sample sizes and more variability of pain frequency investigating associations among pain variability metrics and health outcomes within subgroups based on pain frequency in youth may in turn elucidate how the transition from acute to chronic pain may impact intraindividual variability metrics and how they relate to health outcomes. In addition, we acknowledge that a large number of hypothesis tests were performed, however, given how the current study is novel preliminary work with a relatively small sample size, we were not able to adjust for multiple hypothesis testing, thus underscoring the need for future research with larger sample sizes. Second, due to the design of the EMA, pain intensity data was collected at the same time of the day. It may be informative to assess pain intensity at different time points (e.g., morning, afternoon, evening) to further capture the within-day intraindividual variability of pain in pediatric SCD.

Third, healthcare utilization outcomes (i.e., emergency department visits, hospitalizations, and doctor visits) were self-reported and not confirmed with electronic medical records due to data being collected across three separate regional locations. In addition, lab values were not collected for hydroxyurea usage as part of this study. However, pediatric patients with sickle cell often are seen by multiple healthcare systems/locations, and thus self-report allows for fully capturing healthcare-related experiences across healthcare systems/locations. Fourth, due to the large age range of the current study sample (i.e., age 8 to 17), we did not extensively examine developmental differences associated with age and pubertal changes. Future studies investigating the intraindividual variability of pain within a pediatric sample should investigate developmental differences in intraindividual pain and subsequent health outcomes. Lastly, we acknowledge the data was collected between 2011 and 2015, and that standards of care for children and adolescents with SCD has changed over the past 10 years. Thus, it may be informative for future studies to investigate the intraindividual variability of pain in a more recent pediatric sample.

Beyond the limitations detailed above, the current study highlights areas for future directions and potential clinical implications. An important implication of the current findings is the utility of calculating and examining multiple indicators of intraindividual variability of pediatric SCD pain. Historically, SCD research has focused on assessing mean pain intensity to measure pain burden or to assess potential pain interventions. However, our results reveal the limitations of mean pain intensity as the sole summary measure and suggest the utility of other pain metrics in addition to mean pain intensity for deepening our understanding of how pain impacts the lives of youth with SCD. In addition, given our findings of distinctive associations between intraindividual pain metrics and health outcomes, these pain metrics may be useful in delineating phenotypes of pediatric pain in SCD. Phenotypes derived from intraindividual variability pain metrics may allow for better understanding pain processes connected to poor health outcomes and the development of interventions to that target the most disruptive pain features; thus, improving treatment development and long-term outcomes for SCD patients.

Future research should also focus on whether pain intraindividual variability indicators remain stable during childhood over long periods of time, and investigate if these metrics may be useful for identifying youth with SCD who may be at risk for developing chronic pain earlier. Given how lab values were not collected as part of this study, future research should examine the relationship between EMA-derived intraindividual pain metrics and lab values. Moreover, in the current study sample, approximately 50% had HbSS, a somewhat lower percentage than in other pediatric sickle cell samples. However, it may be that patients with HbSS are more like to have not met eligibility criteria, such as having a comorbid pain condition or neurocognitive issue that impaired their ability to answer questions, which may have contributed to the lower percentage in the current study. Thus, future studies should re-examine our findings in samples with a higher percentage of youth with HbSS. Lastly, intraindividual pain metrics may have the potential to facilitate the assessment of the effectiveness of therapeutic interventions on various aspects of pain and pain-related health outcomes within a youth population. Future studies should further evaluate the role of pain variability metrics for the assessment of pain impact as well as the potential utility of these indicators in clinical trials for youth with SCD.

In conclusion, the current study focused on addressing gaps in the previous literature by investigating intraindividual pain variability metrics and health outcomes utilizing EMA data in youth with SCD. There were distinct associations between pain variability metrics beyond DP and health outcomes. Collectively the patterns of associations suggest the utility of these pain metrics for determining risk in relation to specific health outcomes for youth with SCD. Pain variability metrics derived from EMAs have the potential to better describe intraindividual pain variability within the pediatric SCD population and other pediatric pain populations. Thus, they may be valuable in informing studies of the transition from acute to chronic pain. Lastly, calculating metrics that capture the intraindividual variability of pain may contribute to the development and testing of targeted pediatric SCD pain interventions to improve specific health outcomes among subgroups of youth with SCD.

Glossary

SCD

Sickle cell disease

VOC

Vaso-occlusive crises

EMA

Ecological momentary assessments

DP

Mean daily pain intensity

SD-DP

standard deviation of DP

PPD

proportion of pain days

p90

90^th percentile of DP

SCSES

Sickle Cell-Efficacy Scale

SPI-P

Structured Pain Interview for Parents