Theory Paper Big Tank ONLY

SymptomClusterasaPredictorofPhysicalActivityinMultipleSclerosisPreliminaryEvidence-1-s2.0-S0885392409000682-main.pdf

Home >Nursing homework help >Theory Paper Big Tank ONLY

Original Article

Symptom Cluster as a Predictor of Physical Activity in Multiple Sclerosis: Preliminary Evidence Robert W. Motl, PhD, and Edward McAuley, PhD Department of Kinesiology and Community Health, University of Illinois at Urbana-Champaign, Urbana, Illinois, USA

Abstract Thepresent studyexamined thesymptomcluster of fatigue, pain, anddepression, and itsdirect and indirect prediction of physical activitybehavior in a sampleof individualswith multiple sclerosis(MS) usingaprospectiveresearchdesign andtheTheoryof Unpleasant Symptoms. The sampleincluded 292 individualswith a definitediagnosisof MS. Theparticipantscompleted self-report measuresof fatigue, depression, pain, self-efficacy, and functional limitationsat baselineand six monthslater, worean accelerometer for seven daysand completed a self-report measureof physical activitybehavior. Thedata analysisindicated that: 1) fatigue, depression, and pain represented a symptom cluster; 2) thesymptom cluster had a strong and negative predictiverelationship with physical activitybehavior; and 3) functional limitations, but not self-efficacy, accounted for thepredictiverelationshipbetween thesymptomcluster and physical activitybehavior. Such findingsprovidepreliminarysupport totheimportanceof considering symptomclustersasameaningful correlateof physical activitybehavior in personswithMS. J Pain Symptom Manage 2009;38:270e 280. Ó 2009 U.S. Cancer Pain Relief Committee. Published byElsevier Inc. All rightsreserved.

Key Words Symptom cluster, fatigue, depression, pain, physical activity behavior, multiplesclerosis

Introduction There is accumulating evidence that physi-

cal activity behavior is associated with desirable consequences in persons with multiple sclero- sis (MS). Two recent meta-analyses have dem- onstrated that physical activity behavior is

associated with improvements in walking mo- bility1 and quality of life 2 in persons with MS. Nevertheless, there is considerable evi- dence that individuals with MS are largely sed- entary, as demonstrated by a literature review3 and a meta-analysis.4 One objective of current research has been the identification of vari- ables that correlate with physical activity behav- ior among those with MS. This is based on the assumption that such variables might serve as targets of a well-designed intervention for in- creasing physical activity behavior in this population. Symptoms are perceived indicators of change

in normal functioning, sensation, or appear- ance,5 and have been identified as a correlate of

Funded by the National Institute of Neurological Diseases and Stroke (NS054050) . Address correspondence to: Robert W. Motl, PhD, De- partment of Kinesiology and Community Health , University of Illinois at Urbana-Champaign, 350 Freer Hall, 906 S. Goodwin Avenue, Urbana, IL 61801, USA. E-mail: [email protected] Accepted for publication: August 17, 2008.

0885-3924/ 09/ $e see front matter doi:10.1016/ j.jpainsymman.2008.08.004

270 Journal of Pain and Symptom Management Vol. 38 No. 2 August 2009

physical activitybehavior among personswith MS in cross-sectional analyses. For example, one study reported that the number of symptoms ex- perienced during the past 30 days was negatively associated with physical activity behavior,6 whereas a second study reported that worsening of overall symptomsacross a three-to five-year pe- riod was independentlyand negativelyassociated with self-reported levels of physical activitybehav- ior.7 Another studyreported that higher levels of overall symptoms were directly and indirectly as- sociated with lower levels of physical activity behavior, and the indirect pathwayinvolved diffi- culty walking (i.e., functional limitations8) . One final studyreported that the frequencyof overall symptomsand motor symptomswere directlyand indirectly associated with physical activity behav- ior bywayof self-efficacy.9 These previous studies have generally focused

on either a single dimension of overall symptoms or a single specific symptom as cross-sectional or temporally proximal correlates of physical activ- ity behavior in MS. There are definitional, con- ceptual, and theoretical bases for considering a symptom cluster rather than an overall or spe- cific symptom as a prospective or temporally dis- tal correlate ofphysical activitybehavior in MS.By definition, a symptom cluster represents ‘‘three or more concurrent symptoms (e.g., pain, fa- tigue, sleep in sufficiency) that are related to each other.’’ 10, p. 465 This definition underscores the two primary features of a symptom cluster, namelythe existence of three or more symptoms and an interrelationship, either through a com- mon etiology or statistically as a cluster or latent variable.11 Conceptually, the studyof a symptom cluster recognizes that: 1) multiple symptoms of- ten occur concurrently and 2) co-occurring symptoms likely provide a better prediction of consequences (e.g., behavior, function, or quality of life) than a single symptom. The concept of a symptom cluster and its

possible influence on performance outcomes, including physical activitybehavior, is the central theme of the Theory of Unpleasant Symptoms.5 This theory suggests that symptoms can occur as separate entities or concurrently as a symptom cluster, and the symptoms have antecedents (e.g., physiological, environmental, and per- sonal factors) and temporally proximal and dis- tal consequences (e.g., functional limitations and physical inactivity) . Significantly, one central tenet of the Theoryof Unpleasant Symptoms5 is

that concurrent symptoms likelyhave a stronger effect on consequences compared with a single symptom. This stronger effect is based on the fact that co-occurring symptoms likely catalyze each other (e.g., pain is considerably worse when one is fatigued) , thereby resulting in a dis- proportionately more severe and disruptive symptom experience.5 We further note that the inclusion of both temporally proximal (e.g., cross-sectional) and distal (e.g., prospective) per- formance consequences in this theory provides a basis for considering a symptom cluster as po- tentially having a prospective association with physical activity behavior. One commonlyreported cluster of symptoms

includes fatigue, depression, and pain . This symptom cluster has been identified in persons undergoing treatment for cancer,12 and the same symptom cluster might exist in persons with MS and influence physical activity behav- ior. The symptoms of fatigue, depression, and pain often co-occur in persons with MS,13e 17 and are likelyto have synergistic effects.14 These symptoms are interrelated through common neuropathic consequences, including co-occur- ring and diffuse axonal damage ( i.e., lesions) across different regions of the central nervous system.18 To our knowledge, researchers have neither established the existence of this symp- tom cluster in persons with MS nor examined the prospective or temporally distal relation- ship between the symptom cluster and physical activity behavior in th is population. The present study examined the symptom

cluster of fatigue, pain , and depression, and its direct and indirect association with physical ac- tivitybehavior in a sample of individualswith MS using a prospective research design and the TheoryofUnpleasant Symptoms.We first exam- ined the existence of fatigue, pain , and depres- sion as a prespecified symptom cluster in persons with MS, and secondly, examined the possibility that th is symptom cluster would pre- dict levels of physical activitybehavior after a six- month period. This second purpose examined the possibility of a temporally distal association between the symptom cluster and physical activity in persons with MS. We subsequentlyex- amined the possibility that the symptom cluster predicted physical activity behavior directly or indirectly through a pathway that included self-efficacy and functional limitations, in part, consistent with our previous research.6e 9

Vol. 38 No. 2 August 2009 271Symptom Cluster

Methods Participants This study included a convenience sample of

persons with MS. The sample was recruited through: 1) research announcements mailed to previous study participants; 2) advertisements placed in MSConnection; and 3) e-bursts distrib- uted to registered members of three state chap- ters of the National MS Society. There were 511 individuals who expressed interest in the study, and 387 underwent screening for possible inclu- sion. The screening criteria were: 1) a definite di- agnosis of MS; 2) relapse free in the last 30 days; and 3) ambulatory with minimal assistance. Of those who underwent screening, 27 individuals did not satisfyour inclusion criteria, and 16 indi- viduals declined participation. We sent an in- formed consent document and verification letter to the remaining 344 individuals, and the forms were returned by 300 of the individ- uals (87% response rate) . Of those who re- turned the forms, eigh t did not continue with participation, resulting in the final sample of 292 individuals with MS (3% attrition) . The cat- egorical descriptive characteristics of the sam- ple are provided in Table 1. The mean age of the sample was 48.0 years (standard deviation

[SD] ¼ 10.3) , and the mean duration since diag- nosis of MS was 10.3 years (SD ¼ 7.9) .

Instruments

Physical Activity. We measured physical activity behavior using an ActiGraph accelerometer (model 7164 version; Health One Technology, Fort Walton Beach, FL) and the Godin Lei- sure-Time Exercise Questionnaire (GLTEQ).19 Both the accelerometer and the GLTEQ have evidence of validity in individuals with MS,20,21 and the combined use of self-report and objective measures has been recognized as ideal by experts,22 and allowed for modeling physical activity behavior as a latent variable. The ActiGraph accelerometer has a vertical axis piezoelectric bender element that gener- ates an electrical signal that is proportional to the force acting on it. The positive and negative acceleration signals are digitized by an analog-to-digital converter, numerically in- tegrated over an epoch interval, and the inte- grated value of movement counts is stored in random access memory and the integrator is reset. The accelerometer is programmed for start time and epoch interval, and the move- ment counts are retrieved for analysis by means of a personal computer in terface and software provided with the ActiGraph. The downloaded data from the accelerometers are then entered into Microsoft Excel for pro- cessing. Regarding processing, participants re- corded the time that the accelerometer was worn on a log, and this was verified by inspec- tion of the minute-by-minute accelerometer data. We further examined the accelerometer data for long periods of continuous zeros as a check of compliance with wearing the device, and we used a criterion of 60 minutes of con- tinuous zeros for noncompliance. We based the judgment of a valid day of measurement on 10 hours of wear time during the waking hours, defined as the moment of getting out of bed in the morning through the moment of getting to bed in the evening. We consid- ered the data to be spurious when counts ex- ceeded 20,000 per minute, and we required the participants to have three valid days of data for a reliable estimate of weekly physical activity behavior. There were four participants with three valid days, seven with four valid days, 13 with five valid days, 12 with six valid

Table1 Categorical Demographic Characteristics of the Sample of 292 Individuals with MS

Variable n %

Type of MS Relapsing-remitting 239 82 Secondary progressive 34 12 Primary progressive 12 4 Benign 7 2

Sex Female 245 84 Male 47 16

Race Caucasian 272 93 African American 20 7

Marital status Married 199 69 Single 45 15 Divorced/ separated 39 13 Widow/ widower 9 3

Employment status Employed 154 53 Unemployed 138 47

Education High school 41 14 Some college 83 28 College graduate 168 58

272 Vol. 38 No. 2 August 2009Motl and McAuley

days, and 229 with seven valid days of acceler- ometer data; 27 participants had missing accel- erometer data based on either unit malfunction or not wearing the unit. The movement counts for each day were summed and then averaged across the period of valid days of data. This resulted in accelerometer data expressed in total movement counts per day ( i.e., usual physical activity behavior) . The intraclass correlation for those with seven days of accelerometer data was 0.92. The GLTEQ is a self-administered two-part

measure of usual physical activity behavior; we only included the first part in th is study, consistent with previous research .20,21 The first part has three items that measure the fre- quency of strenuous (e.g., jogging) , moderate (e.g., fast walking) , and mild (e.g., easy walk- ing) exercise for periods of more than 15 min- utes during one’s free time in a typical week. The weekly frequencies of strenuous, moder- ate, and mild activities are multiplied by 9, 5, and 3 metabolic equivalents, respectively, and summed to form a measure of total leisure ac- tivity. This study used the previous week as a time frame for the GLTEQ, and participants completed the GLTEQ after wearing an accel- erometer for the seven-day period.

Fatigue. Fatigue was measured with the Fa- tigue Severity Scale (FSS23) . The FSS has nine items that are rated on a 7-point scale of 1 (strongly disagree) and 7 (strongly agree) . The item scores are averaged, and the overall scores range between 1 and 7,with higher scores forming an overall measure of fatigue’s impact on activities. This scale has good evidence of in- ternal consistency, test-retest reliability, and score validity.23 Coefficient alpha for the FSS was 0.93 in the present study.

Depression. Depression was measured using the Hospital Anxiety and Depression Scale (HADS24) . The HADS contains 14 items; seven items measure anxiety symptoms and seven items measure depression symptoms. The items are rated on a 4-point scale of 0 (most of the time) and 3 (not at all) , and items are reverse-scored and then summed. HADS de- pression and anxiety scores range between 0 and 21, and higher scores indicate more fre- quent depression and anxiety symptoms. This scale has good evidence of score reliability

and validity.24 Coefficient alpha for the depres- sion component of the HADS was 0.82 in the present study. We only included the measure of depression because of our a priori focus on its role in the symptom cluster.

Pain. Pain was measured with the short-form McGill Pain Questionnaire (SF-MPQ25) . The SF-MPQ contains a 15-item adjective checklist that assesses sensory (e.g., ‘‘stabbing,’’ ‘‘sharp’’) and affective (e.g., ‘‘sickening,’’ ‘‘tiring-exhaust- ing’’) dimensions of typical whole bodypain in- tensity. The items are rated using a 4-point in tensity scale of 0 (none) and 3 (severe) , and the item scores are summed. Scores on the SF- MPQ range between 0 and 45, and higher scores indicate more intense overall pain . The scores from the items are summed to form a pain-rat- ing index. The SF-MPQ is internally consistent, reliable across time, and has evidence of score validity.25 Coefficien t alpha for the SF-MPQ was 0.88 in the present study.

Self-Efficacy. Self-efficacy was assessed by the Exercise Self-Efficacy Scale (EXSE) .26 The EXSE scale has six items that assess an individ- ual’s beliefs in his or her ability to engage in 20þ minutes of moderate physical activity be- havior three times per week, in one-month in- crements, across the next six months. The first item on the EXSE was: ‘‘I am able to participate in physical activity behavior three times per week at a moderate intensity, for 20þ minutes without quitting for the NEXT MONTH.’’ The next five items on the EXSE progressively in- creased the length of the physical activitybehav- ior period in one-month increments from two through six months. The items on the EXSE scale are rated based on a 100-point percentage scale comprised of 10-point increments, rang- ing from 0% (not at all confiden t) to 100% (highly confident) . An overall exercise self-effi- cacy score is computed by averaging the item scores from the EXSE scale with higher scores represen ting greater efficacy for engaging in physical activity behavior; scores range from 0 to 100. This scale is in ternally consistent and has evidence of score validity.26 Coefficient al- pha for EXSE was 0.99 in the present study.

Functional Limitation. Functional limitation was measured using the function component of the abbreviated version of the Late-Life

Vol. 38 No. 2 August 2009 273Symptom Cluster

Function and Disability Instrument (LL- FDI) .27 The function component of the abbre- viated LL-FDI contains a 15-item self-report measure of functional limitations ( limitations in a person’s ability to perform discrete actions or activities) that correspond to advanced lower extremity function, basic lower extremity function, and upper extremity function. The 15 items were rated using a 5-point scale of 1 (none) and 5 (cannot do) , and were reverse- scored and then summed to form a composite measure of functional limitation; higher scores represent better functioning. Researchers have provided evidence for the reliability and valid- ity of LL-FDI scores in individuals with MS.28

Procedure All participants provided written informed

consent, and the procedures were approved by a University Institutional Review Board. The baseline and follow-up materials for the study were delivered and returned through the U.S. postal service, and all participants received $40 on returning the study materials. The participants completed self-report mea- sures of fatigue, depression, pain , self-efficacy, and functional limitations at baseline and six months later, wore an accelerometer for seven days and completed the self-report measure of physical activity behavior after the week of ac- celerometer data collection. The six-month follow-up of physical activity behavior allowed for an examination of the prospective or distal relationship between the symptom cluster and physical activity behavior.

Data Analysis The data were primarily analyzed using

covariance modeling with Full Information Maximum Likelihood (FIML) estimation in Mplus 3.0 (Muthen & Muthen , Los Angeles, CA).29 The FIML estimator was selected be- cause there were missing accelerometer (9% missing data) and GLTEQ (5% missing data) data, and the rate of missing data was different per measure. The FIML estimator is a theoreti- callybased method for the treatment of missing data in covariance modeling.30 This approach is standard in covariance modeling programs, and does not require any manipulation of the data (e.g., mean centering) ; it estimates the model and its parameters using all available data from the participants, and has resulted in

unbiased estimates of parameters and model fit with up to 25% of simulated missing data.30 We initially tested a measurement model, whereby FSS, HADS, and MPQ scores served as indicators of the symptom cluster latent vari- able. We then performed a cluster analysis in SPSS, version 15.0 for Windows (SPSS, Chicago, IL) as a method of clustering persons with MS into groupsbased on experienceswith the three symptoms. After establishing the symptom clus- ter as a latent variable, we conducted an analysis, whereby the symptom cluster latent variable predicted physical activity behavior as a latent variable with accelerometer counts and GLTEQ scores as indicators. The final analysis involved testing EXSE and LL-FDI scores as mediators of the relationship between the symptom clus- ter and physical activity behavior latent vari- ables. We based a good-model data fit on a nonsignificant Chi-square value and combina- tory rules of standardized root mean squared residual (SRMR) # 0.08 and comparative fit index (CFI) $ 0.95.31

Results DescriptiveStatistics The descriptive statistics and correlations

among the variables are provided in Tables 2 and 3, respectively. We did not transform the physical activity behavior data as the estimates of skewness and kurtosis for the accelerometer were 1.3 and 2.6, respectively, and for the GLTEQ were 1.6 and 3.1, respectively, indicating a reason- able approximation of a normal distribution. The correlations among FSS, HADS, and MPQ scores, in particular, were moderate in magni- tude,32 supporting the notion of a symptom clus- ter.10,33 There was a moderate correlation between GLTEQ scores and accelerometer counts, consistent with previous research20,21

Table2 Descriptive Statistics for the Measures

in the Sample of 292 Individuals with MS

Measure Mean Score

Standard Deviation

Range of Scores

FSS 5.0 1.4 1e 7 HADS 6.0 4.2 0e 18 MPQ 10.7 7.8 0e 33 EXSE 72.1 32.9 0e 100 FDI 55.0 11.8 28e 75 GLTEQ 26.1 23.9 0e 133 Accelerometer 207,475 105,419 26,541e 694,878

274 Vol. 38 No. 2 August 2009Motl and McAuley

and our expectations. We expected a moderate rather than strong association given that accel- erometer counts provided an overall measure of usual physical activity behavior, whereas the GLTEQ provided a measure of leisure-time physical activity behavior. There were no significant differences be-

tween men and women on FSS, HADS, MPQ, EXSE, LL-FDI, GLTEQ, or accelerometer scores. There were significant differences on FSS, HADS, MPQ, EXSE, LL-FDI, and acceler- ometer scores as a function of employment status. Those who were unemployed reported higher levels of fatigue, depression, pain , and functional limitations, and had less self-efficacy and physical activitybehavior on the accelerom- eter. Age was not significantly correlated with FSS, HADS,MPQ, or EXSE scores, but it was cor- related with LL-FDI, GLTEQ, and accelerome- ter scores. Those who were older reported more functional limitations and had less physi- cal activitybehavior on both the GLTEQ and ac- celerometer measures. Therefore, we controlled for employment status and age in the analyses examining symptom cluster as a predictor of physical activity behavior.

Model 1: Symptom Cluster Latent Variable The single-factor measurement model in

Fig. 1 provided an excellent fit for the data (c 2 ¼ 0, df ¼ 0, P ¼ 0.73, SRMR ¼ 0.00, CFI ¼ 1.00) . The factor loadings for the indicators of the symptom cluster laten t variable were all statistically significant and sufficiently large in magnitude.33 The additional cluster analysis identified three groups or clusters of individ- uals with relatively low (n ¼ 58; 20%), moder- ate (n ¼ 140; 48%), and high (n ¼ 94; 32%) scores across the measures of FSS, HADS, and MPQ. This is displayed in Fig. 2.

Model 2: Direct Association Between Symptom Cluster and Physical Activity The second model that we tested had a di-

rect path between symptom cluster and physi- cal activity behavior latent variables, and it represen ted a good fit for the data (c 2 ¼ 4.98, df ¼ 4, P ¼ 0.29, SRMR ¼ 0.03, CFI ¼ 1.00) . The path coefficient in Fig. 3 between the symptom cluster and physical activity be- havior was statistically significant (g ¼ 0.49) , and indicated that those who reported

the symptom cluster of worse fatigue, depres- sion, and pain were less physically active. This relationsh ip was unaffected in an additional analysis that accounted for employment status and age. Figure 4 provides the mean acceler- ometer and GLTEQ scores across the three groups of individuals with low, moderate, and

Table3 Correlations Among Variables for the Sample of 292 Individuals with MS

Latent/ Manifest Variable 1 2 3 4 5 6 7

1. FSS d 2. HADS 0.50 d 3. MPQ 0.42 0.35 d 4. EXSE 0.43 0.30 0.24 d 5. FDI 0.55 0.42 0.44 0.53 d 6. GLTEQ 0.23 0.16 0.07 0.30 0.29 d 7. Accelerometer 0.35 0.18 0.19 0.25 0.47 0.43 d All correlations are statistically significant (P < 0.05) with the exception of the correlation between MPQ and GLTEQ.

Symptom Cluster

FSS HADS MPQ

.77 .65 .54

Fig. 1. Single-factor model tested using confirma- tory factor analysis for establishing the symptom cluster of fatigue, depression, and pain in the sam- ple of 292 individuals with multiple sclerosis. All coefficients are standardized estimates.

Vol. 38 No. 2 August 2009 275Symptom Cluster

high symptom experiences based on the symp- tom cluster of fatigue, depression, and pain .

Model 3: Indirect Association Between Symptom Cluster and Physical Activity The th ird model that we tested had an indi-

rect path between symptom cluster and physi- cal activity behavior latent variables by way of self-efficacy and functional limitations as man- ifest variables. The model represented a good fit for the data (c 2 ¼ 21.79, df ¼ 11, P ¼ 0.03, SRMR ¼ 0.04, CFI ¼ 0.98) . The

statistically significan t path coefficients are provided in Fig. 5, which indicate that those who reported the symptom cluster of worse fa- tigue, depression, and pain had lower function (g ¼ 0.71) and self-efficacy (g ¼ 0.51) , and those who reported lower function (b ¼ 0.55) , but not self-efficacy (b ¼ 0.07) , had less physi- cal activity behavior. Therefore, the relation- ship between the symptom cluster and physical activity behavior was significantly and indirectly accounted for by functional limita- tions (gb ¼ 0.39) , but not self-efficacy (gb ¼ 0.04) . These relationsh ips were unaffected in an additional analysis that accounted for employment status and age.

Discussion Using a prospective research design and the

Theory of Unpleasant Symptoms, the current study examined the symptom cluster of fa- tigue, depression, and pain as a predictor of physical activity behavior among persons with MS. Our results indicated that: 1) fatigue, de- pression , and pain represented a symptom cluster based on bivariate correlations, covari- ance modeling, and cluster analysis; 2) th is symptom cluster had a moderate and negative predictive relationsh ip with physical activity be- havior; and 3) functional limitation, but not self-efficacy, accounted for the predictive rela- tionship between the symptom cluster and physical activity behavior. Such findings

FSS HADS MPQ Measure

M ea

n ±

St an

da rd

E rr

Low (n=58) Moderate (n=140) High (n=94)

Fig. 2. Mean scores and standard errors for the measures of fatigue, depression, and pain based on the low, moderate, and high symptom clusters identified in the cluster analysis with the sample of 292 individuals with multiple sclerosis.

Symptom Cluster

FSS HADS MPQ

Physical Activity

GLTEQ ACCEL

−.49

.54 .80.84 .60 .51

Fig. 3. Model tested using covariance modeling for understanding the association between symptom cluster and physical activity behavior in a sample of 292 individuals with multiple sclerosis. All coefficients are standardized estimates. Accel ¼ accelerometer counts.

276 Vol. 38 No. 2 August 2009Motl and McAuley

provide preliminary support for the impor- tance of considering a symptom cluster as a meaningful correlate of physical activity be- havior in persons with MS. There is a growing body of research that has

identified symptoms as a cross-sectional or tem- porally proximal correlate of physical activity behavior in MS.6e 9 All these studies were af- fected by a set of limitations that included the general focus on the overall frequency or

intensity of symptoms, lack of a guiding symp- tom-based theoretical framework, and a cross- sectional research design. The present study extended previous research by focusing on a symptom cluster of fatigue, depression, and pain as a temporally distal predictor of physical activity behavior after a six-month period using a prospective design and based on the Theoryof Unpleasant Symptoms.5 Our findings are con- sistent with the Theory of Unpleasant Symp- toms, that is, 1) the symptom cluster of fatigue, depression, and pain was moderately and negatively associated with physical activity behavior; and 2) the relationship between the symptom cluster and physical activity behavior was indirect through functional limitations. The present study further extended previous research6e 9 by considering both self-efficacy and functional limitations as possible interven- ing variables between the symptom cluster and physical activity behavior in MS. To that end, our results provided evidence that functional limitations, rather than self-efficacy, repre- sented the stronger intermediate variable in the temporally distal association between the symptom cluster and physical activity behavior. The results of previous research and the cur-

rent studymight have implications for the pro- motion and maintenance of a physically active lifestyle in persons with MS. Indeed, individuals

Low (n=58) Moderate (n=140) High (n=94)

Accelerometer GLTEQ

Fig. 4. Mean scores and standard errors for the ac- celerometer counts and GLTEQ scores across the three groups of individuals with low, moderate, and high symptom experiences identified in the cluster analysis. Accelerometer data are expressed in units of original values 10 4 so that the units are easily graphed and compared in magnitude with GLTEQ scores.

Symptom Cluster

FSS HADS MPQ

Physical Activity

GLTEQ ACCEL

.53 .81.79 .62 .56

EXSE

FDI

−.51

−.71 .55

.07

Fig. 5. Model tested using covariance modeling for understanding the associations among symptom cluster, self- efficacy, functional limitation, and physical activity behavior in a sample of 292 individuals with multiple sclerosis. All coefficients are standardized estimates. Solid lines represent statistically significant paths, and dashed lines represent nonsignificant paths. Accel ¼ accelerometer counts.

Vol. 38 No. 2 August 2009 277Symptom Cluster

with MS are often physically inactive and seden- tary,3,4 and emerging evidence indicates that symptoms might be a determinan t of inactivity among persons with MS. The management of symptoms, therefore, represents a possible strat- egy for the promotion and maintenance of physical activity behavior among those with MS. This might be accomplished, in particular, through a program that includes provisions for the management of fatigue, depression, and pain as a cluster of symptoms. To that end, there is a growing bodyof evidence regard- ing programs for managing symptoms, such as fatigue, depression, and pain in MS.34e 36We be- lieve that one direction for future research in- volves incorporating elements of these programs into a multidimensional in tervention for promoting physical activity behavior in per- sons with MS. An additional important finding of this

study is the confirmation of a symptom cluster of fatigue, depression, and pain in persons with MS. The symptom cluster has been identi- fied in persons undergoing treatment for can- cer,12 and was established in the current study based on three sets of analyses, including 1) the bivariate correlations among the three vari- ables; 2) the fit of a single latent variable for the three variables; and 3) the cluster analysis that yielded three distinct groups of individ- uals with relatively low, moderate, and high scores across the cluster of three variables. This is consistent with the recommendations in the literature for statistically identifying a symptom cluster,10,33,37 particularly the re- cent call for using two conceptual approaches for identifying symptom clusters ( i.e., bivariate correlations and factor analysis) and sub- groups of individuals based on experiences with a symptom cluster ( i.e., cluster analysis38) . We further note that the symptoms of fatigue, depression, and pain are linked through a com- mon etiology based on neuropathic conse- quences, including co-occurring and diffuse axonal damage ( i.e., lesions) across different regions of the central nervous system.18 There- fore, we provide preliminary evidence for the existence of a symptom cluster of fatigue, de- pression, and pain in individuals with MS, and this is consistent with the literature on persons with cancer.12 There are several limitations of the current

study. One limitation is that we focused on

a six-month follow-up for measuring symptoms, and the magnitude of the relationship between the symptom cluster and physical activitymight differ with shorter or longer follow-up periods. The second limitation is that we did not control for baseline activityin the analysis nor did we ex- amine the relationship between changes in the symptom cluster and physical activity behavior across time. The focus on only three symptoms of fatigue, depression, and pain is a third limita- tion, and future researchers might consider ad- ditional symptoms within this cluster and other clusters of symptoms in MS. The symptom di- mensions that were measured varied across the FSS, HADS, and SF-MPQ. For example, the FSSmeasured the impact of fatigue on activ- ities, whereas the HADS measured the fre- quency of depressive symptoms. This is a the fourth limitation of the present study, and fu- ture researchers should be cautious in using measures that share common symptom dimen- sions. The final limitation is that the relation- ship between the symptom cluster and physical activity behavior was only studied in a sample of persons with MS, and future re- searchers might consider extending our find- ings into those with other disease conditions, including cancer. In conclusion, the results of the present

study provide a preliminary basis for consider- ing the role of the symptom cluster of fatigue, depression, and pain within the growing body of knowledge that symptoms play a role in mul- tiple behavioral outcomes, including physical activity behavior, in persons with MS. The role of symptoms in physical activity behavior of individuals with MS, in particular, is an area substantial with research potential for promoting and main taining physical activity behavior. The promotion and maintenance of an activity lifestyle is an important compo- nent for enhancing the lives of persons with MS.1,2

References 1. Snook EM, Motl RW. Effect of exercise training on walking mobility in multiple sclerosis: a meta-- analysis. Neurorehabil Neural Repair 2009;23: 108e 116.

2. Motl RW, Gosney JL. Effect of exercise training on quality of life in multiple sclerosis: a meta-analy- sis. Mult Scler 2008;14:129e 135.

278 Vol. 38 No. 2 August 2009Motl and McAuley

3. Motl RW. Physical activity and its measurement and determinants in multiple sclerosis. Minerva Med 2008;99:157e 165.

4. Motl RW, McAuley E, Snook EM. Physical activ- ity and multiple sclerosis: a meta-analysis. Mult Scler 2005;11:459e 463.

5. Lenz ER, Pugh LC, Milligan RA, Gift A, Suppe F. The middle-range theory of unpleasant symptoms: an update. Adv Nurs Sci 1997;19:14e 27.

6. Motl RW, Snook EM, McAuley E, Gliottoni R. Symptoms, self-efficacy, and physical activity among individuals with multiple sclerosis. Res Nurs Health 2006;29:597e 606.

7. Motl RW, Arnett PA, Smith MM, et al. Worsen- ing of symptoms is associated with lower physical activity levels in individuals with multiple sclerosis. Mult Scler 2008;14:140e 142.

8. Motl RW, Snook EM, Schapiro RT. Symptoms and physical activity behavior in individuals with multiple sclerosis. Res Nurs Health 2008;31: 466e 475.

9. Snook EM, Motl RW. Physical activity behavior in individuals with multiple sclerosis: roles of overall and specific symptoms, and self-efficacy. J Pain Symptom Manage 2008;36:46e 53.

10. Dodd MJ, Miaskowski C, Paul SM. Symptom clusters and their effect on the functional status of patients with cancer. Oncol Nurs Forum 2001;28: 465e 470.

11. Miaskowski C, Dodd M, Lee K. Symptom clus- ters: the new frontier in symptom management research. J Natl Cancer Inst Monogr 2004;32:17e 21.

12. Fox SW, Lyon DE. Symptom clusters and quality of life in survivors of lung cancer. Oncol Nurs Forum 2006;33:931e 936.

13. Crayton HJ, Rossman HS. Managing the symp- toms of multiple sclerosis: a multimodal approach. Clin Ther 2006;28:445e 460.

14. Krupp LB. Fatigue in multiple sclerosis: A guide to diagnosis and management. New York: Demos Medical Publishing, 2004.

15. Minden SL, Frankel D, Hadden L, et al. The Sonya Slifka Longitudinal Multiple Sclerosis Study: methods and sample characteristics. Mult Scler 2006;12:24e 38.

16. National Multiple Sclerosis Society. Multiple sclerosis information sourcebook. New York: Infor- mation Resource Center and Library of the National Multiple Sclerosis Society, 2005.

17. O’Connor AB, Schwid SR, Herrmann DN, Markman JD, Dworkin RH. Pain associated with multiple sclerosis: systematic review and proposed classification. Pain 2008;137:96e 111.

18. Lublin FD. Clinical features and diagnosis of multiple sclerosis. Neurol Clin 2005;23:1e 15.

19. Godin G, Shephard RJ. A simple method to assess exercise behavior in the community. Can J Appl Sport Sci 1985;10:141e 146.

20. Gosney JL, Scott JA, Snook EM, Motl RW. Phys- ical activity and multiple sclerosis: validity of self-report and objective measures. Fam Commun Health 2007;30:144e 150.

21. Motl RW, McAuley E, Snook EM, Scott JA. Valid- ity of physical activity measures in ambulatory indi- viduals with multiple sclerosis. Disabil Rehabil 2006;28:1151e 1156.

22. Dishman RK, Washburn RA, Schoeller DA. Mea- surement of physical activity. Quest 2001;53:295e 309.

23. Krupp LB, LaRocca NG, Muir-Nash J, Steinberg AD. The fatigue severity scale. Application to patients with multiple sclerosis and systemic lupus erythematosus. Arch Neurol 1989;46:1121e 1123.

24. Zigmond AS, Snaith RP. The Hospital Anxiety and Depression Scale. Acta Psychiatr Scand 1983; 67:361e 370.

25. Melzack R. The short form McGill Pain Ques- tionnaire. Pain 1987;30:191e 197.

26. McAuleyE. Self-efficacy and the maintenance of exercise participation in older adults. J Behav Med 1993;16:103e 113.

27. McAuley E, Konopack JF, Motl RW, Rosengren K, Morris KS. Measuring disability and function in older women: psychometric properties of the late-life function and disability instrument. J Gerontol A Biol Sci Med Sci 2005;60A:M901e M909.

28. Motl RW, Snook E,GosneyJL,McAuleyE.Measur- ing disabilityand function in individualswith multiple sclerosis: psychometric properties of the Late-Life Function and Disability Instrument. In: Goldfarb PM, ed. Psychological tests and testing. Hauppauge, NY: Nova Science Publishers, 2007: 251e 263.

29. Muth én LK, Muth én BO. Mplus. Los Angeles: Author, 1998e 2003.

30. Enders CK, Bandalos DL. The relative perfor- mance of full information maximum likelihood esti- mation for missing data in structural equation models. Struct Equat Model 2001;8:430e 457.

31. Hu L, Bentler PM. Cutoff criteria for fit indices in covariance structure analysis: conventional crite- ria versus new alternatives. Struct Equat Model 1999;6:1e 55.

32. Cohen J. Statistical power analysis for the behav- ioral sciences, 2nd ed. Hillsdale, NJ: Lawrence Erl- baum, 1988.

33. Barsevick AM. The elusive concept of the symp- tom cluster. Oncol Nurs Forum 2007;34:971e 980.

34. Kos D, Duportail M, D’hooghe MB, Nagels G, Kerckhofs E. Multidisciplinary fatigue management programme in multiple sclerosis: a randomized clin- ical trial. Mult Scler 2007;13:996e 1003.

Vol. 38 No. 2 August 2009 279Symptom Cluster

35. Mathiowetz VG, Matuska KM, Finlayson ML, Luo P, Chen HY. One-year follow-up to a random- ized controlled trial of an energy conservation course for persons with multiple sclerosis. In t J Re- habil Res 2007;30:305e 313. 36. Mohr DC, Hart S, Vella L. Reduction in disabil- ity in a randomized controlled trial of telephone-ad- ministered cognitive-behavioral therapy. Health Psychol 2007;26:554e 563.

37. Barsevick AM, Whitmer K, Nail LM, Beck SL, Dudley WN. Symptom cluster research: conceptual, design, measurement, and analysis issues. J Pain Symptom Manage 2006;31:85e 95. 38. Miaskowski C, Aouizerat BE, Dodd M, Cooper B. Conceptual issues in symptom clusters re- search and their implications for quality-of-life as- sessment in patients with cancer. J Natl Cancer Inst Monogr 2007;37:39e 46.

280 Vol. 38 No. 2 August 2009Motl and McAuley