Activity measurement in pediatric chronic fatigue syndrome

Abstract

Objectives:

Individuals with myalgic encephalomyelitis (ME) and chronic fatigue syndrome (CFS) experience debilitating symptoms, including post-exertional malaise, an intensification of symptoms after physical or cognitive exertion. Previous studies found differences in the activity levels and patterns of activity among individuals with ME and CFS, compared to healthy controls; however, limited research exists on the activity levels of pediatric patients. The objective of this study was to examine differences in activity between healthy children and youth with ME and CFS.

Methods:

The present study examines the objective (i.e., ActiGraphy) and self-reported levels of activity among children (ages 5 to 17) enrolled in a community-based study of pediatric CFS.

Results:

Children with ME and CFS evidenced lower activity levels than healthy control children. Moreover, participants with ME and CFS evidenced increased nighttime activity and delayed initiation of daytime activity. Participants’ self-reported activity data significantly correlated with their ActiGraph data, suggesting that children with ME and CFS are able to accurately describe their activity level.

Discussion:

This study highlights differences in activity level and diurnal/nocturnal activity patterns between healthy children and those with ME and CFS. These differences should be considered in identifying appropriate supports and accommodations for children with ME and CFS.

Myalgic encephalomyelitis (ME) and chronic fatigue syndrome (CFS) affect over one million persons in the United States.¹ The etiology of the illness remains unknown; however, individuals with ME and CFS experience substantial impairment in their daily functioning and quality of life due to their debilitating symptoms, such as post-exertional malaise, sleep dysfunction, and cognitive impairment.²Post-exertional malaise, defined as an exacerbation of symptoms (such as muscle weakness or fatigue) following exertion, is often considered a core feature of the illness.³ This symptom limits individuals’ ability to engage in physical activity, and approximately 25% of patients are housebound.⁴ Evidence of a substantial reduction in activity level (when compared to premorbid functioning) is required by many ME and CFS case definitions,^2,5,6 and individuals with ME and CFS have emphasized that a return to premorbid activity levels should be required when researchers operationalize “recovery” from the illness.⁷ Given the importance of activity level in diagnosing, treating, and defining recovery from ME and CFS, the current study sought to examine and compare objective and subjective measures of physical activity.

ActiGraphs are small devices (5.1 × 3.8 × 1.5 cm) that measure an individual’s movement through recording the amplitude and frequency of the accelerations that occur during activity.⁸ They can be worn around the waist, wrist, or ankle. Several studies have demonstrated the accuracy of ActiGraphs in measuring participants’ activity levels.^8–10 Furthermore, studies have demonstrated that ActiGraphs provide reliable measurements of the activity levels of individuals with ME and CFS.^11–13For example, a study that compared the ActiGraph patterns of adults with ME or CFS and healthy controls found that the participants with ME or CFS engaged in less intense activity and had longer resting periods following activity compared to controls.¹⁴ Despite their accuracy in measuring activity level, ActiGraphs have several limitations: studies may not have the funds required to purchase ActiGraphs for all participants; participants may find it challenging to wear the ActiGraph consistently throughout the study period, resulting in non-compliance; or participants may lose or fail to return the ActiGraph at the end of the measurement period.¹⁵

Due to these limitations, many studies utilize subjective methods to measure activity level, such as self-report surveys, activity logs, or direct observation. However, physical activity is a complex construct to quantify. According to a summary of the primary methods of measuring activity, most fail to capture one or more of the following dimensions: frequency, duration, intensity, or type of activity.¹⁶

While self-report surveys are a cost effective, practical method of obtaining activity data, particularly when working with large samples of participants, these measures are subject to biased reporting. Depending upon the timeframe measured, participants may not be able to accurately recall their activity level.^16,17 Moreover, participants may over-report their activity level as a consequence of social desirability bias.^17,18 Children’s reports of activity level are further complicated by challenges related to telling or estimating time, leading to inaccurate reports of the duration of their activities.¹⁹ The terminology used in child-report questionnaires is particularly important, as children’s concept of activity may differ from adult’s; terms such as “play,” “games,” and “sports” may elicit more accurate information.¹⁹

Activity logs require participants to record the type and duration of their activity shortly after it occurs.¹⁶ These logs provide a detailed account of participants’ activity and result in rich qualitative data.²⁰ The use of activity logs, as compared to general questions about activity level, may reduce recall errors, since they are completed soon after activity takes place.¹⁶ However, activity logs are also subject to social desirability bias and can be burden-some to participants, increasing the likely-hood of attrition.

Direct observation methods can be useful in assessing the activity levels of younger children who cannot easily recall their levels of activity. However, this method is labor- and time-intensive, as it requires multiple coders and high participant cooperation. Additionally, observation is impractical when collecting data from large samples of participants. Furthermore, the data obtained through this method can be affected by the phenomenon of reactivity: altered participant behavior when someone is observing.²¹

These limitations further indicate the need to validate subjective activity measures with objective data. While previous research has shown high correlations between accelerometers and self-report questionnaires, the majority of this research was conducted on adult samples.¹⁷ The current study analyzed and compared the objective and self-report activity data collected from children and adolescents. Analyses examined whether ActiGraph data could differentiate the activity levels of children diagnosed with CFS (diagnosed by the Fukuda et al. CFS criteria,⁵ the most commonly-used diagnostic criteria for the illness that requires severe fatigue and at least four additional symptoms; the term “CFS” is used to describe this group, as “CFS” is the illness name utilized in the Fukuda et al. CFS case definition) from the activity levels of healthy control children. The authors hypothesized that children with CFS would have significantly lower activity levels than control children. Additionally, this study examined the correlation between children’s self-reported activity levels and ActiGraph data in order to evaluate the reliability of a self-reported data among children with CFS.

Method

Participants

Participants were a part of a larger epidemiological study of pediatric ME and CFS. Details of the full study methodology are provided in Jason, Katz, and colleagues²²; a brief overview follows. Ethical approval was obtained from the institutional review boards at [DePaul University and Northwestern University], and informed consent/assent was obtained from children and their parents; no adverse events occurred during this research study. Children aged 5 to 17 were recruited by calling Chicagoland households. During this phone call, parents (or guardians) were asked to answer questions regarding their children’s health. Children who screened positive (i.e., endorsed symptoms of ME and CFS that were not explained by other illnesses or conditions), as well as age- and gender-matched participants who screened negative (i.e., healthy control children) were invited to phase two of the study. Phase two consisted of a medical exam and psychological evaluation; children and parents also completed self-report questionnaires. At the end of the appointment, children were asked to wear an ActiGraph monitor for 24 hours. Children and their parents were compensated $150 for their participation ($75 for the parent and $75 for the child). Following the appointment, an independent panel of physicians reviewed the results of the medical exam, psychological assessment, and self-report measures scores in order to determine whether participants met criteria for ME or CFS. The current study includes children who met the Fukuda et al. (1994) case definition (n = 35) and healthy control children (n = 15) who agreed to wear an ActiGraph.

Measures

ActiGraphy.

Participants were asked to wear ActiGraphs around their waist for 24 hours after the medical and psychological assessments if they lived (or their parents worked) near enough to the authors’ institution such that the ActiGraphs could be returned or collected by research staff without excessive time or cost. Participants were asked wear the ActiGraph at all times, except during activities that would get the device wet (e.g., showering). In these circumstances, participants were asked to record at what times they removed the ActiGraph and put it back on. Children with more than one hour of missing data were removed from the sample. This cutoff was used to reduce the amount of missing ActiGraph data and ensure that participants’ data were comparable, while still allowing for necessary immeasurable activities, such as showering.

ActiGraphs provide “activity count” data in one-minute epochs. The current study analyzed ActiGraphy data in two ways. First, a total activity count was created for each participant by summing their one-minute activity counts. Secondly, a cumulative activity count was created for each participant that spanned from 12:00am to 11:59pm such that the participants’ cumulative activity level at each minute in the day could be examined.

Modifiable Activity Questionnaire (MAQ).

The MAQ is a self-report questionnaire that asks participants to report how many minutes per day, days per week, and months per year they engage in 39 specified activities, such as basketball, dance, or weight lifting.²³ Participants are also able to add activities that are not already listed on the questionnaire. The Modifiable Activity Questionnaire (MAQ) is a commonly-used, and well-studied survey.¹⁶ A recent study has shown that the MAQ is a reliable and valid measure of physical activity among children and adolescents.²⁴

To create a total activity score for each participant, the following formula was used to determine the number of minutes per year the participants engaged in each activity: [minutes per day]*1440*[days per week]*4.25*[months per year]. The resulting values for each activity were then summed to determine the total number of minutes per year the participants engaged in the listed activities.

Fatigue severity scale

The Fatigue Severity Scale²⁵ is a self-report measure that includes nine items related to different aspects and impacts of fatigue (e.g., exercise brings on my fatigue, fatigue is among my three most disabling symptoms, I am easily fatigued, etc.). Participants are asked to reflect upon the past week and rate each item on a Likert-type scale ranging from 1 (strong disagree) to 7 (strongly agree). These scores are summed to create a total score (ranging from 7 to 63). The Fatigue Severity Scale (FSS) was identified as the most commonly-used fatigue scale according to a past bibliographic study of fatigue measurement scales.²⁶ Furthermore, the FSS has been shown to be a useful tool in differentiating between healthy controls and patients with various diseases.²⁷

Results

Demographics

Table 1 displays the demographic characteristics of the study sample. There were no significant differences in age between individuals with CFS (M = 13.7, SD = 2.7) and healthy controls (M=13.9, SD=1.8): t(48)=.74, p=.46. Furthermore, chisquare tests indicated that the percentage of children with CFS did not differ by gender [χ²(1, N=50)=1.24, p=.27] or by race [χ²(1, N=50)=.75, p=.86] from control children. Finally, there were no between-group differences in the time of year during which participants wore the ActiGraphs, χ²(1, N=50)=0.66, p=.42.

Table 1.

Demographic comparison.

	Patient		Control
	M	(SD)	M	(SD)
Age	13.4	(2.7)	13.9	(1.8)
	%	n	%	n
Gender
Male	42.9	(15)	60.0	(9)
Female	57.1	(20)	40.0	(6)
Race
White	62.9	(22)	66.7	(10)
Black or African-American	22.9	(8)	26.7	(4)
Multiracial	11.4	(4)	6.7	(1)
Asian or Pacific Islander	2.9	(1)	0.0	(0)
Time of Year ActiGraph Worn
In School	45.7	(16)	33.3	(5)
Out of School	54.2	(19)	66.7	(10)

None of these comparisons were statistically significant.

ActiGraphy

Figure 1 displays the differences in cumulative activity over the course of 24 hours between individuals who met the Fukuda et al. CFS criteria⁵ and healthy controls. A one-way ANOVA was conducted to compare the total daily ActiGraphy counts of the CFS and control groups. Results indicated a significant difference in activity level between these groups, F(1, 49)=4.31, p=.043. Specifically, individuals who met the Fukuda et al. CFS criteria5 (M=315,716, SD=150,215) were significantly less active than healthy controls (M=447,806, SD=301,683).

Figure 1.

Mean cumulative activity level of patients and controls over time.

Self-report questionnaires

To evaluate the reliability and validity of self-reported activity level, correlational analyses were conducted to examine the relation between ActiGraphy data and two self-report questionnaires (the MAQ and FSS). There was a significant correlation between ActiGraph data and the MAQ, r=.521, p < .001, n = 50, but the correlation between ActiGraph data and FSS score was not significant, r=−.181, p=.212, n=50.

Discussion

Results of this study demonstrate that healthy children are able to participate in higher levels of activity than children with CFS. Moreover, Figure 1 suggests that children with CFS have delayed onset of daytime activity and somewhat more nighttime activity than controls, potentially indicative of sleep disruption. These results are consistent with previous literature that has demonstrated sleep abnormalities among children and adolescents with ME and CFS.^28,29

This study also sought to understand the relation between ActiGraph and self-report questionnaires among children and adolescents. Participants’ ActiGraphy counts significantly correlated with their self-reported activity level, as measured by the MAQ. These findings indicate that child participants are able to provide an accurate account of their activity levels. Although ActiGraph and FSS scores were not significantly correlated, these instruments measure different constructs and demonstrate that children may participate in high levels of activity, regardless of fatigue. This finding is congruent with assertions by clinicians that children may demonstrate paradoxical reactions to fatigue³⁰ and offer counterevidence to suggestions that individuals with ME and CFS experience fatigue due to deconditioning.³¹

This study had several limitations. First, we were unable to collect ActiGraph data from all participants due to logistical constraints; however, a primary purpose of this study was to examine the association between ActiGraphy data and self-report measures, given challenges such as those faced in the current study in utilizing accelerometers in research. Additionally, participants needed to remove ActiGraphs when engaging in activities during which the ActiGraph would become wet. Due to this limitation, the ActiGraph data of participants who engaged in swimming or water sports may not accurately reflect their true activity level. Despite this limitation, the current study found a significant association between ActiGraphy and MAQ data, and the MAQ includes prompts related to a variety of water sports. Future research could use newer, waterproof, ActiGraphs to determine whether the present study’s findings could be replicated. Relatedly, this study utilized an older ActiGraph model, and previous research has shown significant step count differences between older and newer ActiGraph models.^10,32 However, a subsequent study demonstrated that activity counts (the output utilized by the current study), did not differ between older and newer models,¹⁰ suggesting that replication with newer generation ActiGraphs should be comparable to the current study’s results. Finally, participants were only required to wear the ActiGraph monitor for one day. A longer measurement period would likely provide more precise activity data. In spite of these limitations, this study has several strengths. As community-based recruitment methods were used, the study sample’s demographic characteristics are diverse, and access to tertiary care was not required for participation. Furthermore, diagnoses of CFS were made by a team of physicians who reviewed extensive medical, psychological, and self-report data, allowing for greater confidence in the veracity of children’s diagnoses.

The results of this study have implications for the development of child-specific case definitions, as the majority of existing case definitions for ME and CFS require individuals to experience a substantial reduction in functioning or activity level; however, these case definitions do not provide guidelines for operationalizing this reduction.³³ Several studies have demonstrated that minor changes to the measurement of case definition criterion can cause significant fluctuations in the number and type of individuals who meet criteria.^34,35 These small variations occur when researchers have different interpretations of how to measure a criterion. Criterion variance is a primary cause of low diagnostic reliability³⁶; thus, clear measurement guidelines should be defined when case definitions are developed.³⁷ This issue is compounded in the diagnosis of children, as fatigued children may present as hyperactive or fidgety,³⁰ and children have less control over their activities. For example, parents may force children to attend school and participate in extracurricular activities, despite their fatigue.

Prospective research methods could collect activity data to assess premorbid activity level and measure activity reductions among children who develop ME or CFS; these data would allow for an evaluation of the utility of the “substantial reduction” criterion of ME and CFS case definitions. Historically, children with ME and CFS have faced significant stigma from medical providers, family, and friends³⁸; researchers have posited that vague case definition criteria has contributed to this stigma, as lack of specificity leads to inaccurate diagnosis, and children with primary psychological diagnoses may be misdiagnosed with ME or CFS.³⁷Child-specific case definitions that include detailed guidelines related to criterion measurement would help medical professionals in diagnosing the condition, and allow individuals with ME and CFS to receive the medical treatment necessary to relieve their symptoms.

In summary, this study demonstrated that children with CFS evidence lower activity levels than healthy control children. Correlational analyses suggest that children and adolescents can accurately report upon their activity level when using self-report questionnaires (specifically, the MAQ); these questionnaires may be useful for time- or cost-constrained research studies, including prospective studies that require large sample sizes.