RESEARCH AND PRACTICE

Effects of Recreational Physical Activity and Back Exerciseson Low Back Pain and Psychological Distress:Findings From the UCLA Low Back Pain Study

Eric L. Hurwitz, DC, PhD, Hal Morgenstern, PhD. and Chi Chiao, MS. PhD

Low back pain is one of the leading reasonsfor physician visits' and is the most commonreason for use of complementary and alterna-tive medicine in the United States.^ Backsymptoms are frequently accompanied bydepression or anxiety and psychological dis-tress, '^ which are principal reasons for use ofboth eonventional and complementary healthcare. ' Despite the well-known cardiovascularand weight control benefits of regular physi-eal exercise, participation has decreased inrecent years.^ with almost 70% of US adultsre|)orting that they do not engage in regularleisure-time physical activity.̂ Findings fromseveral studies conducted in the Umted Statesand elsewhere have shown associations ofphysical activity with depression and psycho-logical well-being,'""' as well as the potentialbuffering effects of physical fitness or activity

17-21

on stress.Kxercise interventions have been shown in

some randomized clinical trials (RCTs) to helpprevent low back pain in at-risk populations.^^However, tlie bulk of the scientific literatureindicates that specific back exercises (e.g.,flexion, extension, and strengthening exer-cises) and instruction in standardized exer-cises (as advocated by many physical thera-pists) do not appreciably improve low backpain prognoses.̂ "̂ '̂ "*

Although the results of observadona! stud-ies have not been entirely consistent (e.g..Kujala et al-̂ )̂, these investigations have pro-duced evidence pointing to the benefits ofleisure-dme physical activity in preventinglow back pain or improving its prognosis.̂ **"'"In addidon, various published guidelines"'^"recommend that padents with acute low backl>ain stay acdve, and a recent systematic re-view of RCTs provided corroboradon for sueha strategy. •''' Nevertheless, no published stud-ies have followed a clinical popuiadon overan extended period and collected setial data

Objectives. We sought to estimate the effects of recreational physical activityand back exercises on low back pain, related disability, and psychological dis-tress among patients randomized to chiropractic or medical care in a managedcare setting.

Methods. Low back pain patients (n = 681) were randomized and followed for18 months. Participation in recreational physical activities, use of back exercises,and low back pain, related disability, and psychological distress were measuredat baseline, at 6 weeks, and at 6, 12, and 18 months. Multivariate logistic regres-sion modeling was used to estimate adjusted associations of physical activityand back exercises with concurrent and subsequent pain, disability, and psy-chological distress.

Results. Participation in recreational physical activities was inversely associ-ated—botfi cross-sectionally and longitudinally—with low back pain, related dis-ability, and psychological distress. By contrast, back exercise was positively as-sociated—both cross-sectionally and longitudinally—with low back pain andrelated disability.

Conclusions. These results suggest that individuals with low back pain sbouldrefrain from specific back exercises and instead focus on nonspecific physicalactivities to reduce pain and improve psychological health. {Am J Public Health.2005;95:1817-1824. doi:10.2105/AJPH.2004.052993)

on reported use of back exercises and leisuredme physical acdvity to estimate separate andunconfounded associations of physical activityand exercise witli concurrent and subsequentlow back pain, disability, and psychologicaldistress.

We sought to estimate the effects on severaloutcomes of recreadonal physical acdvity andback exercise among low back pain patientspardcipating in a large RCT conducted in amanaged care setting. We hypothesized that(1) pardcipadon in sport or recreadonal physi-cal activities would reduce levels of low backpain, disability, and psychological distress and(2} back exercises would not reduce levels ofpain, disability, and psychological distress.

METHODS

Study Design and Source PopulationIndividuals were enrolled from October

30, 1995, through November 9. 1998;follow-up data were collected through Jtme

2000. Baseline for each padent was the dateof randomizadon, which was also the date oftheir first (baseline) visit and the date that thebaseline quesdonnaire was completed, indi-viduals presendng for care with low backpain at one of the 3 outpadent care facilidesof a muldspecialty network of health careproviders based in Southern Califotnia wererandomized, in a balanced design, to 4 treat-ment groups: chiropraede care with physicalmodalities, chiropractic care without physicalmodalities, medical care v̂ ith physical ther-apy, and tnedical care without physical tlier-apy. Follow-up quesdonnaires were sent toparticipants at 6 weeks and. after completionof the initial (baseline) questionnaire, at 6. 12.and 18 months. The source populadon con-sisted of the approximately 100000 mem-bers of the multispeciaity network.

Eligibility CriteriaPadents were eligible for the study if they

were health maintenance organization (HMO)

October 2005, Vol 95, No. 10 I American Journal of Public Health Hurwitz etal. Peer Reviewed | Research and Practice i 1817

RESEARCH AND PRACTICE

members of any HMO; had sought care atone of the 3 sites between October 30, 1995,and November 9, 1998; had presented with acomplaint of low back pain; had not receivedlow back pain treatment within the pastmonth; and were at least 18 years old. Poten-tial participants were excluded if their lowback pain involved third party liability orworkers' compensation claims or if they

(1) had low back pain from a fracture, tumor,infection, or other nonmechanical cause;

(2) had severe medical comorhidity; (3) werebeing treated with electrical devices, such aspacemakers; (4) had a blood coagulation dis-order or were using corticosteroids or anti-coagulant medications; (5) had progressivelower-limb muscle weakness; (6) had currentsymptoms or signs of cauda equina syndrome(bilateral radicular symptoms and signs in thelower extremities owing to ner\'e root com-pression from large herniated lumbar nucleuspulposus); (7) had plans to move out {)f tliearea; (8) were not easily accessible by tele-phone; or (9) were not fluent in English.

Data Collection

At baseline, participants underwent physi-cal examinations and completed question-naires, yielding data on sociodemographic,clinical, and psychosocial characteristics; pai--ticipadon in recreational physical activities;and use of back exercises. Follow-up ques-tionnaires addressed paiticipants' health aiidfunctional status, low back pain intensity andrelated disability, physical activity levels, anduse of back exercises. Details about patientscreening and enrollment and the treatmentand foltow-up protocols have been providedelsewhere,''

Low Back Pain and Disability Measures

Participants used numerical rating scales(0=no pain. 10—unbearable pain) to assesstheir most severe and average pain intensityduring the past week. Such scales have beenshown to have excellent reliability and valid-ity for measuring back pain. '*' Paiticipantswho assigned to their low back pain ratings of2 or higher were considered to have clinicallymeaningful levels of pain.

Low back disability was assessed via the24-item Roland-Morris Low Back DisabilityQuestionnaire.'' '̂ Patients responded by an-swering yes or no to indicate whether each

statement represented a tiiie description oftlieir current disability owing to low backpain. Scores can range from 0 [indicating nodisability) to 24 (indicating severe disability').This instrument has been shown to be reli-able and valid''"'" and responsive to changeover time-"*""" Participants with scores of 3or higher were considered to have clinicallymeaningful low back disability.'""*"'

Psychological Distress Measure

Psychological distress was measured witlian alternate form of the 5-item menta] healthindex (MHI-5)"̂ '' fi-om the Medical OutcomesSttidy 36-ltem Short-Form Health Survey(SF-36).''""'̂ The MHl-5, which assessesgeneral mental health, including depression,anxiety, behaviorai-emotional coiitrol andgeneral pasitive affect, has been shown to bea reliable indicator of depressive symptoms.The measure is scored on a scale of 0 to 100,wilh lower scores indicating more psychologi-cal distress, i^rticipants scoring below themedian (76) were considered to have psycho-logical distress.

Recreational Physical Activity and BackExercise Measures

At baseline and at each follow-up pointparticipants were asked how many hours perweek, on average, they engaged in walkingand 1 or more light moderate, and strenuoussport or recreational physical activities. Meta-bolic equivalent task (MET) values were as-signed to each activity, and MET scores werecalculated for each paiticipant"*'' The MFTscore measures metabolic enei^ cost ex-pressed as a multiple of tJie resting metabolicrate (the bigher the score, the greater the en-ei'gy expenditure from physical activity).Quaitiles were tlien formed for activity-specific and total MliTs.

At baseline and at each follow-up point,participants were also queried about tbe fre-quency with which they had engaged in backexercises specifically intended to prevent orto deal with low back pain during the pastweek. Response options were never, seldom(less than 1 day), sometimes {1-3 days), andoften (4-7 days).

Other Variables

Otlier variabies measured at baseline lordescriptive or analytic purposes included age.

gender, race/ethnicity, education level, mari-tal stattis, employment status, self-perceivedgeneral health status, scores on 4 additionalSF-36 subscales,""' duration of current lowback pain episode, ntimber of previous epi-sodes, score on the Internal Healtli Locus ofControl scale from the MultidimensionalHealtb Locus of Control Scales,^" and itemsfoaising on patients" social support networksand strategies for coping with pain. At base-line and at each follow-up assessment partici-pants were also asked to report the numberof hours per week they engaged in musclestrengthening and flexibility exercises.

Data Analysis

Linear and logistic regression models wereused to estimate cros.s-sectional and longitudi-nal associations of back exercise and recre-ational physical activity with low back pain,low back disability, and psychological distress.Generalized estimating equations witli robuststandard error estimates were used to takeinto account within-subject correlations dur-ing the 18-month follow-up period.'^'"'' Be-cause findings from the linear and logisticanalyses were similar, we present only the lo-gistic model results. Estimated associationsare described in the form of adjusted odds ra-tios (ORs) and 95'Vo coTifidence intervals (CIs).SAS was used to manage and analyze data;the GENMOD procedure was for generalizedestimating equations (GEF!) estimation.̂ "*'̂ ''

Low back pain, low back disability, andpsychological distress were dichotomized atvarious cutpoints in preliminary models. Be-cause findings did not vaiy appreciably ac-cording to the cutpoint used, results are pre-sented here for the models for the "clinicallymeaningful" cutpoints. We found little evi-dence of different effects according to inten-sity of sport or recreational activity (e.g.. walk-ing or light moderate, and strenuousactivities), and thus we limited our presenta-tion to total MET values derived from all re-ported sport and recreational activities.

Separate models were fit to estimate cross-sectional and longitudinal associations of backexercise and recreational physical activitywith each low back pain, disabiiity, and psy-chological distress outcome. Data from the 6-week and the 6-, 12-. and 18-month follow-upassessments were assessed simultaneously in

1818 I Research and Practice I Peer Reviewed ; Hurwitz el al. American Journal of Public Health I October 2005. Vol 95, No. 10

RESEARCH AND PRACTICE

all analyses. Covariates included in the cross-sectional models were age, gender, baselineduration of low back pain, ntimber of previ-ous episodes of pain, assigned treatmentjjroLip, social support, strategy for coping with[lain, internal locus of control, baseline MHI-5score, baseline outcome variable value, mus-cle strengthening and flexibility exercising,and follow-up week. Back exercise estimateswere also adjusted for physical activity level,and physical activi^ estimates were adjustedlor back exercise.

In addition to this set of covariates, all lon-gitucfinal (transition) models (first-orderMarkov chain regression models^ '̂) also in-cluded previous values of the pain or disabil-ity vanable and previous MHI-5 scores.These models were used to examine associa-tions between activity and exercise and subse-quent levels of pain, disability, and psycholog-ical distress. Back exercise estimates wereatijiisted for previous physical activity level,and physical activity estimates were adjustedfor previous baek exercise. Product terms rep-resenting interactions (deviations from multi-[)licativity) of physical activities and back ex-ercises with gender, age, treatment group,baseline duration of low back pain episode,and follow up week were included in prelimi-nary models: however, because these esti-mated interactions were negligible, they wereexcluded fiijm the final models.

RESULTS

Screening, Enrollment, and Follow-UpA total of 2355 patients were screened.

Eight hundred eighty-six (37.6%) were ex-cluded because they had undei-gone low backpain treatment in the past month (n = 270),their back pain was not primarily in thelumbosaeral area (n— 144). they had fee-for-service or no health insurance coverage (n =119), they had Medi-Cal or Medicare cover-age only (n = 80), they were involved in athird-par^ liability or workers' compensationcase (n = 55), they were not (luent in English(n =46), they were less than 18 years of age(ti = 4.'l). they planned to move out of the ai-ea(n= 18). or they were not easily accessible bytelephone (n = 4)- Prospective participantswere excluded for medical reasons such aslow back pain owing to fracture, tumor, or in-

fection (n=40); severe coexisting disease(n = 37); use of anticoagulant medications(n= 13); ankylosing spondylitis or other rheu-matie disease (n=7); treatment with an elec-trical device (n = 5); progressive or severe uni-lateral tower-limb muscle weakness (n—2);abdominal aortic aneurysm (n= 1); symptomsor signs of cauda equina syndrome (n= 1);and blood coagulation disorder {n= 1).

Of the 1469 eligible patients. 788 (53.6%)declined to participate because of lack of in-terest (n = 345), preference for one or moreof the randomized treatments (n —266), in-convenience (n= 137), or inability to makemultiple copayments (n = 31). Nine otherwiseeligible and willing potential participants werenot enrolled because they did not understandthe informed consent form. Thus, of the1469 eligible patients, 681 were enrolled inthe study. Six hundred seventy-five partici-pants (99-1%) returned 6-week follow-upquestionnaires with complete outcome data;652 (95.7%) and 610 (89.6%) participantscompleted 6- and 18-month follow-up ques-tionnaires, respectively.

Characteristics of Study PopulationTable 1 shows baseline distributions of se-

lected sociodemographic variables. Fifty-twopercent of the participants were lemide, 50%were younger than 50 yeai's, 40% were non-White, and 67"/(i were employed. Almost 3 in10 participants reported engaging in no recre-ational sport or physical activities at baseline.Table 2 shows baseline distinbulions of lowback pain and health status variables. Almosthalf of the participants had been in pain formore than 1 year. Approximately 80*'/o re-ported previous episodes of low back painand at least 1 disability day in the pastmonth. The median low back disability scoreof 11 rellected moderate disability, whereaspain intensi^ scores indicated appreciablelevels ofpain perception. SF-36 seores werelower than US general population norms""'but were consistent with seores observed inother populations of back-pain patients. •'̂ '̂ "̂ *'

Cross-Sectional AssociationsTable 3 presents logistic regression results

for adjusted cross-sectional associations (ORs)of back exercise and recreational physical ac-tivity with average and most severe low baekpain, low back disability, and psychological

TABLE 1-Frequency Distributions of

Seiected Sociodemographic

Characteristics: UCLA Low Back Pain

Study, 1995-2000

Variable

<30

30-49

50-69

>70

Gender

Male

Female

Race/ethnicity

White/non-Hispanic

Latino/Hispanic

Asian/Pacific Islander

African American/Black

Otber

Education level

High school Of less

Some college

Coiiege

Maritai status

Married/involved in relationship

Widowed/divorced/separated

Never married

Employment status

Employed fuli time

Employed part time

On leave/unemployed

Retired

General health status

Excellent

Very good

Good

Fair/poor

Weekly metabolic equivalent

task score," %

0

0.1-10.49

10.5-25.9

>26

Sample

{n-681),%

9.4

40.2

32.2

18.2

48.0

52.0

60.4

29.8

4.5

2.8

2.5

29.6

39.5

30.8

71.6

183

10.1

58.7

8.1

7.5

25.7

8.1 •

37.0

42.0

12.9

27,6

21.9

25.5

25.0

^The mean age was 51.0 years (SO = 16.7); the

medlar was 50 years,

The metabolic equivalent task score measures

metaboiic energy cost expressed as a multiple of

the resting metaboiic rate (the higber the score, the

greater the energy expenditure from physical

activity).

October 2005, Vol 95, No. 10 i American Journal of Public Health HurwHz et al. i Peer Reviewed i Research and Practice i 1819

RESEARCH AND PRACTICE

TABLE 2-Low Back Pain and Health

Status Variables at Baseline: UCLA Low

Back Pain Study, 1995-2000

Duration of back pain episode, %

<3wk

3 wk-3 mo

3mo- ly

>ly

Hoiand-Morris disabiiity score

(0-24 scale)

<5,%

6-10,%

11-15,%

>15,%

Mean (SD)

Median

Most severe low back pain in past

*eek (0-10 scaie)

Mean (SD)

Median

Average iow back pain in past week

(0-10 scaiel

Mean (SD)

Median

Days with restricted activity owing to

iow back pain in past montb

Mean (SD)

Median

At ieast 1, %

History of iow back pain episodes, %

Yes

No

SF-36 pbysicai lunctioning score

Mean (SD)

Median

SF-36 roie limitations score: physicai

problems

Mean (SD)

Median

SF-36 roie limitations score:

emotional problems

Mean (SD)

Median

SF-36 mental heaith index score

Mean (SD)

Median

SF-36 generai beaitfi perceptions score

Mean (SD)

Median

Sampie

(n-681)

26.1

15.6

11.6

46.7

17.5

31.6

27.9

23.1

10.9 (5.4)

11

6.7 (2.1)

7

4,6 (1.9)

5

6.9 (8.0)

4

76.9

82.2

17.8

62.2 (24.6)

65

41.3(40.1)

25

70.0 (39.2)

100

71,2 (16.6)

76

67.8 (18.0)

70

TABLE 3-Adjusted Cross-Sectional Associations of Back Exercise and Recreational

Physicai Activity With Pain Ratings, Disability, and Psychoiogicai Distress

Back exercise

Never

Seldom (<ld/wk)

Sometimes (1-3 d/wk)

Often (4-7 d/wk)

Physical activity (weekiy metabolic

equivalent task score)

0

0.1-10.49

10.5-25.9

>26

Most Severe Pain,"OR (95% Ci)

1.00

1.48(1.09,2.00)

2.13(1.63,2.79)

2.12(1.57,2.85)

1.00

0.78(0.55,1.09)

0.68(0,49,0.95)

0.62(0.44,0.87)

Average Pain,''OR (95% Cl)

1.00

1.49(1.14.1.94)

1.56(1.20,2.01)

1.56 (1.18,2.06)

1.00

0.83(0.60,1.13)

0.63 (0.46,0.85)

0.72(0.52,0.99)

Back DisaOiiity,'Ofi (95% Cl)

1.00

1.59(1.19,2.12)

1.85(1.44,2.38)

1.61(1.22,2.13)

1.00

0,72(0.52,1.01)

0.60(0.44,0.82)

0.48(0.35,0.66)

Psychological Distress,"OR (95% Ci)

1.00

1.05 (0.78,1.42)

0.98 (0.77,1.25)

0.95 (0,73,1.23)

1.00

0.77(0.56,1.06)

0.68(0.50,0.91)

0.60(0.44,0.83)

Wore. OR- odds ratio; Ci - confidence intervai. Average back pain and most severe low back pain were defined as ratings of 2

or higher on a 0-10 numerical scaie; low back disat)iiity was defined as a score of 3 or above on the 0-24 Roland-Morris

Low Back Disabiiity Qjestionnaire: psychoiogjcal distress was defined as a mental health inden score of less than 76.

"Adjusted for age, gender, baseline duration of low back pain episode, number of previous iow back pain episodes, assigned

treatrrent group, social support, strategy for coping with pain, internai health locus ot control, baseline mental health index

score, baseline vaije of outcome variable, muscie strengthening and flexibility exercising, and follow up week. Back exercise

effect estimates were also adjusted for physical activity levei, and physicai activity effect estimates were adjusted for back

exercise.

''Adjusted for age, gender, baseline duration of low back pain episode, number of previous low back pain episodes, assigned

treatment group, social support, strategy for coping with pain, internal locus of control, baseline mental bealth index score,

baseline iow back pain and disability levels, muscle strengthening and flexibility exercising, and follow up week. Back exercise

effect estimates were also adjusted for physical activity level, and physical activity effect estimates were adjusted for back

M e . Roland-Morris-Rolartd-Morris Low BackDisability Questionnaire; SF-36 = Medical OutcomesStudy 36-ltem Short Form Health Suivey.

disti'ess dunng the follow-up period. For ex-ample, relative to paiiicipants reporting nophysical activity, the odds of meaningful lowback disability were more tban halved ajnongparfidpaiits in the top qumlile (>26 ME'IVweek) of recreationai pbysica] activity {0R=0.48; 95% CI = 0.35, 0.66). No associationwas detected between back exercise stalusaiid confiiiTent psychological distress.

By contrast, after adjustment for tlie effectsof back exercise and other covariates, odds ofclinically significant low back pain and dis-ability' decreased as reported physical activitylevel increased (P<.05 for trend). For exam-ple, participants in the top quartile of recre-ational physical activity' (26 MI'Ts per week)were less than half as likely as participantsreporting no physical activity to have experi-enced meaningful low back disability (0R=0.48; 95"/ii CI = 0.35, 0.66). Similarly, oddsof psychological distress decreased as re-ported physical activity increased {P<.05 fortrend). Relative to physically inactive partici-pants, the odds of being psychologically dis-

tressed were 40% lower among participantsin the top quartile of physical activity (0R =0.60; 95% CI = 0.44. 0.83).

Longitudinal AssociationsTable 4 presents logistic regression resulLs

for adjusted longitudinal associations (ORs)of back exercise and recreational physical ac-tivity with subsequent average and most se-vere low back pain, low back disability, andpsychological distress. After control for previ-ous values of the low back pain or disabilityvariables and other covariates, back exerciseincreased the odds of subsequent appreciablelow back pain and disability by 64% and44'Vo, respectively. However, back exercisereduced the odds of subsequent psychologi-cal distress by 22"/,) (OR = 0.78; 95%) CI =0.59. 1.03).

As reported participation in physical activ-ity increased, tlie odds of experiencing clini-cally meaningful low back pain and disabilityat tiie subsequent assessment decreased. Forexample, after controi foi- low back disability

1820 Research and Practice Peer Reviewed Hurwitz et al. American Journal of Public Health October 2005. Vol 95, No. 10

RESEARCH AND PRACTICE

TABLE 4-Adjusted LongittJdJnal Associations of Back Exercise and Recreational PhysicalActivity With Subsequent Pain, Disability, and Psychoioglcal Distress

Previous back exercise

Never

Seldom [< 1 d/wk)

Sometimes (1-3 d/wk)

Often (4-7 d/vrt)

Physical activity (weekly metabolic

equivalent task score)

0

0,1-10,49

10,5-25.9

>26

Most Severe Pain,̂OR (95% Cl)

1,00

1,78(1,23,2.59)

1.46 (1,10,1,95)

1,64(1,21,2,23)

1.00

0.91(0.62,1.34)

0.85 (0.60,1,20)

0.82(0.57,1.17)

Average Psin,̂OR (95% Cl)

1,00

1,48(1,07,2,03)

1,48(1.14,1,93)

1.30 (0,97,1.73)

1.00

0,95(0.69,1.32)

0.85 (0.63,1.15)

0,82 (0.60,1.12)

Back Disability,'OR (95% Cl)

1.00

1.32 (0,95.1.83)

1,19 (0.91,1,56)

1,44 (L07,1.94)

1,00

0,91 (0,64,1.29)

0,70 (0,50,0.97)

0,69 (0.50,0.97)

Psychoiogical Distress,''OR (95% Cl)

1.00

0.87(0,65,1,17)

0,93(0.72,1.20)

0.78 (0.59,1,03)

1.00

1.03 (0.76,1.39)

0,91(0,69,1.20)

0,75(0,55,1.01)

Note. Average back pain and most severe low back pain were defined as ratings of 2 or higher on 0-10 numerical rating

scales; disability was defined as a score of 3 or above cn the 0-24 RolanO-Morris Low Back Disability Questionnaire; and

psychological dratress was defined as a mental health index score of less than 76.

'Adjusted for age, gender, baseline duration of low back pain episode, number of previous lew back pain episodes, assigned

treatment group, social support, strategy for coping with pain, internal locus of control, previous muscle strengthening and

flexibility exercising, previous mental health index score, previous value of outcome variable, and follow-up week. Back

exercise effect estimates were also adjusted for previous physical activity level, and physical activity effect estimates were

adjusted for previous back exercise.

"Adjusted for age, gender, baseline duration of low back pain episode, number of previous low back pain episodes, assigned

treatment group, social support, strategy for coping with pain, internal health locjs of control, previous muscle strengthening

and flexibility exercising, previous low back pain and disability levels, previous mental health index score, and follow-up

week. Back exercise effect estimates were also adjusted for previous physical activity, and physical activity effect estimates

were adjusted for previous back exercise.

at previoiLs assessment and other covariates,the odds of clinically meaningCul disabilitywere 30% lower among participants in theupper 2 quartiies of the physical activity dis-tribution than among inactive participants,i.e., those in the lowest quartile of the physi-cal activity distribution (OR=0.69: 95"/()CI=0.50, 0,97), Compared with inactive par-ticipants, and after conttiDl for psychologicaldistress at the previoas assessment and othercovariates, tlie odds of being psychologicallydistressed at the subsequent assessment wereS.'i'JAi lower among those in the uppermostciuaitile of the physical activity distribution(OR-0.75;95%C!-0.55. 1.01),

DISCUSSION

To our knowledge, this is tbe first longitudi-nal study of primary care low back pain pa-tients to sbow not only tbat recreational phys-ical activity and back exercise are associatedcross-sectionally and longitudinally with lowback pain and disability but also that tbe as-

sociations are in opposite directions and thatphysical activity may be beneficial in copingwith comorbid psychological distress, pain,and disability. We found that (1) participationin recreational physical activities reduces thelikelihood of concurrent and subsequent lowback pain, related disability, and psychologi-cal distress and (2) use of hack exercises in-creases the likelihood of concurrent and sub-sequent low back pain and related disability.We also found tbat the associations of backexercise and physical activity with low backpain outcomes were stronger in tbe cross-sectional analyses than in the longitudinalanalyses, indicating that amount of back exer-cise and physical activity may be afTected bydegree of back pain or related disability (i.e.,reverse causation).

Leisure time physical activitj', physical fit-ness, and exercise have been suggested aspossible risk or prognostic factors for lowback pain in some community-based, occupa-tional, and clinic-based studies'*^"'̂ ' but not in

-^""''' In a systematic review of

11 controlled clinical trials conducted in in-dustrial settings,'"* tbe 3 investigations of ex-ercise interventions were fbimd to demon-strate effects on preventing back pain orwork-loss days, and the researchers con-cluded that evidence for the effectiveness ofexercise was limited. However, the 3 studiesinvolved several methodological problemsthat made interpretation of their results diffi-cult A recent review of 39 studies involvingworkers showed little association of leisuretime physical activity witb low back pain, al-though sedentary leisure activity was found tobe associated with an increased prevalence oflow back pain and related sick leave,''̂

In a systematic review of exercise therapythat identified 39 RCTs, the researchers con-cluded that specific hack exercises shouldnot be recommended for patients with acute{12 weeks or less) or chronic (more than 12weeks) pain but that exercise in general maybe beneficial as pail of an active r-ehabilita-tion program for chronic pain sufferers.̂ "̂They found (1) moderate evidence that flex-ion exercises are not effective in reducingacute pain, (2) strong evidence that extensionexercises are not effective in redticing aaitepain, (3) no evidence that flexion exercisesare effective in reducing chronic pain, (4) noevidence that strengthening exercises are ef-fective in reducing acute pain, and (5) strongevidence that strengthening exercises are notmore effective than other types of exercise.^''In a systematic review of 20 RCTs of physicalexercise and training interventions, the Inter-national Paris Task Force on Back Pain con-cluded that active physical exercise should bepromoted among patients with acute orchronic pain and that no evidence exists to in-dicate the effectiveness of specific exercises orthe relative benefits of one exercise regimenover another.^''

Consistent with those studies, our findingssuggest that specific back exercises may becounterproductive and that restoration of nor-mal functioning should instead be empha-sized.'*'* Our results are also consistent withcuirent guidelines for managing acute lowback pain that recommend low-stress aerobicexercises such as walking and swimming. •'•̂ ••'•'Recommendations to engage in physical activ-ity appear to be appropriate for people withchronic low back pain as well. Brisk walking

October 2005 , Vol 95, No. 10 American Journal of Public Health Hurwitz el al. Peer Reviewed I Research and Practice ; 1821

for at least 3 hours per week is equivalent to10.5 METs, a level we found to be associatedwith reductions in conairrent low back pain,disability, and psychological distress and sub-sequent back disability. As others have loundin regard to back pain"' and depression/'"strenuous activities such as running andswimming do not neccssanly result in betteroutcomes than walking or other less strenu-ous activities. Given the shared neural andbiochemical pathways of physical and psycho-logical pain.'''''" engaging in physical activitymay be a safe and ettective sti'ategy to ame-liorate both pain and depression.

The beneficial effects of physical activityon pain perception and psychological dis-tress may be caused by beta-end orphin lev-els, which have been shown to be lower inphysically active men than in sedentarymen.^''' Higher resting beta-endorphin lev-els have been observed to be associatedwith depression,^'' and physical activity maydecrease resting plasma beta-endorphin andimprove mood.'"* In contrast, circulafingbeta-endoiphin levels have heen shown tobe elevated during physical exercise,' ' ' tohe relatively higher among pregnant womenwho exercise regularly than among thosewho do not,^' and to increase among anginapatients during treadmill testing.̂ *̂ Periph-eral and experimental pain thresholds havealso been shown to rise in response to exer-cisej'^'^ and percepfions of tabor pain maybe lower among regular exercisers thanamong nonexercisers.^' However, withrespect to both physical pain and psycholog-ical distress, it is unclear whether beta-endorphin or other mediators and mecha-nisms are responsible for appai'ent physicalactivity-induced effects.^" Different mecha-nisms may be involved depending on typeor duration of activity, other behavioralor environmental factors, or presence ofcomorbidities. '̂̂ '*'

Several limitations of this study should beconsidered when intcipreting our tlndings.Our [jarticipants were primajy care pafientstaking part in an investigation of treatmentoptions for low back pain. Thus, they maynot be representative of individuals with lowback pain who present for care in otlier kindsof clinical settings (e.g., specialty dinics orcenters focusing on third par^ tiahility or

workers' compensation cases) or who do notseek clinical treatment at all. Because all ofthe parficipanls had low back pain at hase-line, we cannot draw inferences about the as-sociations of exercise and physical activitywith pain or disability and psychological dis-ti-ess in initially pain-frpc populations. Also,hecause the outcomes assessed here are rela-tively common, tlic odds ratios observedprobably tended to overestimate reductions inrelative risks. Furthennore. inlbniiation onspecific types of back exercises was not col-lected. Although little evidence indicating thatsome specific exercise regimens are more ef-fective than othei"s exists in the literature, cer-tain exercise regimens may be more effecfivetlian others. Also, we relied on participants'self-reports of their exercise and physical ac-tivifies, and it was not feasible to validate re-sponses with other strategies such as directobservation.

Because recreational physical activity wasnot randomized, confounding is an additionalconcern, although we considered the most im-portant potential confounders in our analyticstrategy. Residual confounding is possible,however. For example, individuals with morechronic or disahling histories of back painmay be more likely than those with less se-vere or chronic histories to engage in regularback exercise. Although we controlled statisti-cally for baseline duration of low back painepisode and number of previous episodes ofpain, our back exercise estimates could stillbe subject to residual confounding. However,it is unlikely that such confounding would beso substantial as to mask truly protecfive backexercise effects.

Although these limitations weaken our abil-ity to offer linn causal inferences, our studyinvolved a number of strengths that supportthe possibility that the associations observedmay in fact be causal. First, the 18-month fol-low-up rate of almost QC/o diminishes thepossihility that our findings were due entirelyto selection bias. Second, we controlled forseveral factors likely to confound crude asso-ciafions of exercise and physical activity withsubsequent pain, disability, and psychologicaldistress. Third, previously validated measureswere u.sed in assessing all outcomes, and well-established MFT values were assigned tophysical activity categories. Finally, because

our serial assessments involved a large popu-lation of low back pain sufferers, we wereable to conduct analyses that clearly delin-eated the temporal relations of exposures andoutcomes. Tlie stronger CTOSS-sectional thanlongitudinal associations observed suggestpossible reverse causation, a problem inher-ent in inteiprefing estimates derived fromcross-secfional studies of physicai activity andlow back pain.**̂

In summary, in a population of primarycare patients presenting with low back pain,participation in recreational physical activi-ties was inversely associated, both cross-sectionally and longitudinally, with lowback pain, related disability, and psychologi-cal distress. In contrast, back exercise waspositively associated both cross-section allyand longitudinally with low back pain andrelated disability. These results suggest thatindividuals with low back pain, ratlier thanbeing advised to engage in specific back ex-ercises, should instead be encouraged tofocus on nonspecific physical activities tohelp reduce their pain and improve theirpsychological health. Because of the percep-tion that physical acfivity could result inpain persisting for a longer period and fearof pain have been idenfified as possible fac-tors keeping low back pain patients frombeing physically active,"' clinicians maywant to reduce such barriers to patientsmodifying their behavior •

About the AtithorsF.nc L. Huni'iiz is with the Department of Epidemiology,Schnol of Public Health. University nf California, Los An-geles, and the Soiilheni Cutifomia University of HealthSciences, W!iittier Hal Morgenstem is wilh the Depart-ment of Epidemioiof^. School of Puhlic Healtb. Uninrsityof Michigan. Ami Arbor Chi Chian is with the Departmentof Commiinitii Heaith Science.^ School of Public Health,University of Catifontia. Los Angeles.

Requests for reprinm should he sent to Eric L. Hurwitz.DC, PhD, School of Public Health. Department ofEpidemi-ohgy. Universit}/ of California. Box 951772. Los Angeles,CA 90095-1772 (e-mail: ehuiwitz@ucla.edu).

This article was accepted October 12. 2004.

ContributorsE. [.. HurwLty. was responsible for study dcsigii. df vrlop-nient of sludy hypulhescs. supervision of data c()llectjonand analysis, and drafting of the articlr. 11. Morgcnstemcon til bu ted to study design, data interpiftation, andcritical revision of the article. C. Chiao contrihuted tothp literature review and data analysis.

1822 Research and Practice i Peer Reviewed Hurwitz et al. American Journal of Public Health October 2005. Vol 95, No. 10

RESEARCH AND PRACTICE

Acknowledgmentsiliis sliidy was Tunded by grAn\s ln)m the Agency forHt-althraro Hosearcli anti Qualily (RIH HS07755) andthe SfJUtlicni CiiliPoniia University of 1 leallh Sciences.K. 1.. Hurwitz was supported by a frranf from the Na-tional Center for Complementary and Alternative Medi-cine IK23 ATOOO55].

We are j^ateliil to fei Yn and I.ii-May Cliiang fortheir assistance with data managemenl and statisticalanalysis.

Human Participant ProtectionDie study iirotom! was ap|injved by the institutionalreview boards of the LIniversity of California, Los Ange-les, and Ihe piulicipating hcallh care network, f^rtici-[lanLs pruvided written inlbmied consent.

References1. Hing i;, Middleton K. National Hospital Ambula-Hiry Medical Can- Survey: 2002 outpatient dppart-ni<?nt summary. Adv Diilti Vital Jiealih Slal. Jtine 24,2004:345.

2. Barnes PM. Pnwell-Griner E. McFann K, NahinRL. Complementary and altemative medicine usenmiinj; adults United States, 2002, Adv Data Vitalllmkh Slal. May 27, 2004:343,

i. Kinney RK, Gatchel RJ, PolatiTi PB. Fogaity WT,Mayer TC. Prevalence of psyehopalhologj' in acute andchronic low baek pain palitrnLs. / Ocaip Hehub. 1993:3:95-103.

4 Rush AJ, Poladii P. flatdiel RJ. Depression andI limnic low hack pain: establishing priorities in treat-ment. Spine. 2OOO:2S:25(i6-2571.

3. Spariing PB, Owen N. Lambert HV, Haskell \VL.Promoting physical activity: Ihe new imperative forpuhlie health. Ihullh Educ Res. 200();l3:3(i7-37<i.

ii. Banies VW. Sehoenhf>ni CA. Physical acUvityiunong adulls: United States, 2000. .Adv Data VitalHealth Stat. May 14, 2003;333.

7. ['rederick J. Frerichs RR, Clark VA. Personalhealtli habiLs and symptoms of depiession at the com-munity level. Prev Med 19tt8;17:173-182.

8. Ross CE, Hayes D. Fxercise and psychologic well-being in the <romniunit>'. Am J lipiiletnial. 1988;I27:762-771.

9. Stephens T". Physieal activity ant! mental health inthe United States and Canada: evidence from four pop-ulatioi) surveys. Prev Med. 1988:17:35-47.

10. VVeyeRT S. Physical inactivity and depres.sion inIhe community: evidence from the Upper BavarianTieldSluiiy. In! J Sports Med 1992;13:492-49(i,

11. 1 lassnien P, Koiviila \ , Uiitela A. Physical exercisennd ()sych<)logical weil-bi-ing: a population study inI inland. Prn-Med 20()0;30:l7-25,

1 .̂ f'ariiior MK, Locke BZ, Mosdcki KK, Dannenherg\l., Larson I)B, RadlofTLS. Physical activity and depres-

Mve symptoms: the NUANiiS 1 Kpidemiologic Follow-UpStudy. AmJEpulemwt. 1988;128:1340-13ril.

13. Camacho iC, Roberts HH, Lazarus \ B , Kaplan(;A. Cohen RD. Physicai activity and deprewion: evi-dence frtjiu the Alanicda Couniy Study, .-hnj Hpi-dcmiiyl 1991:134:220-231.

14. Lampinen P, Heikkinen R-L, Riioppila 1. Changesin intensity of physical exercise as predictors of depres-

sive symptoms among older adults: an eight-yeai" fol-low-up. Preu Med. 2000:30:371-380,

15. Weyerer S, Kupfer B. Physical exercise aJid |Tsy-chological health, Spurts Med. 1994:I7:l()8-llfi.

16. Tucker LA. Physical fitnes.s and psychological dis-tress. Int J Sport Psychol. 1990;21:185-201.

17. Sinyor D. Schwailz SG. Pcronnel F, Brisson G,Seruganian P. Aerohic fitness level and reactivity topsychosocial stress: physiological, hiochemical, and sub-jective measures. Psf/chosom Med. 198:J;45:2O5-2I7.

18. Hoth DL, Holmes DS. Influence of |)hysica] fitnessin detennining the impact of stressful tile evenLs onphysical and psychologic health. Psi/chiisiim Med. 1985;47:164-173.

19. Rejeski WJ, Thompson A, Bnibaker PH. MillerHS. Acute exercise: buffering psyi-hosoeial sti'css re-sponses ill women, Health Psychai 1992:11:335-362,

20. Carmack CL, Roudreaux E, Amaral-Melendex M.Brantley PJ, de Moor C. Aerobic fitness and leisurephysical activity as iiKKlerators of the stres.s-illTiess re-lation. Ann Hehav Med. 1999:21:251-257.

21. Hurwitz HL. Do asthma and physieal inactivity in-Huenee the associations of personal and Job stressorsmth perceived stress and depression? Findings fnmithe 1998-1999 California Work and Health SurveyAnn Epidemiol. 2003;13:358-368.

22. Linton SJ, van Tlildei- MW. Preventive interven-tions lor back and neck pn)hlcms: what is the evi-dence? Spine. 2001,2(1:778-787.

23. van Tulder M, Malmivaara A, Ksmail R, Koes B.F.xereise therapy for low hack jiaiii: a systematic re-view within the framework of the Cochrane Collabora-tion Back Review Group. Spine. 2000;25:2784-2796.

24. Abenhaim L, Rossignol M. Valat JP, et al. The roleo!" activity in the therapeutic management of back pain:report of the international Paris Task Force on Backi'ain. Spine. 20t}0:25(suppl 4):1S-33S.

25. Kiijala UM, laimela S. Viljanen T, et al. Physicalloatiing anrf performant e as prediclors of back pain inhealthy adults: a 5-year pi-ospective study. Eur] .ApplPhi/siol. 199fi:73:452-458.

2(S. Leino P. Does leisure time physical activity pre-vent low back disorders? A prospective study of metalindustiy employees. Spine. 1993:18:863-871.

27 Harre^hy M. Kjer J, Hesselsoe G, et al. Kpidemio-logical aspecLs and risk factor for low back pain in 38-year-old men and women: a 25-year prospective co-hort study of (i40 school children. Eur Spine J. 199Ci:5:312-318.

28. Hcistaro S, Vaitiainen K, Heliovaara M, Piiska P.Trends of back pain in eastern Finland, 1972-1992, inrelation to socioeconomic status and behavioral riskfactors. Am J Epidemiol. 1998:148:671-682.

29. Salminen JJ. I{rkinlalo M. Laine M, et al. Low baekpain in the young: a prosjieetive three-year follow-upstudy of suliJecLs with and without bw tiack pain.Spim' 1995:20:2101-2108.

30. Taimela S, Diederich C, Hubsch \1, Heinricy M.The folc of physical exercise and inaeti\'ily in pain re-cuiTcnce and absenteeism from work alter active out-patient rehabilitation for recurrent or chronic low hackpain: a follow-up ,study. Spine. 2000;25:1809-'1816.

31. Marfarlane GJ, Thomas F, Croft PR, PapageorgioiiAC, Jayson Ml, Silniaii AJ. IVedictors of early improve-

ment in low hack pain amongst c-onsulters to generalpractice: the infltience of pre-morbid and episode-related factoi-s. Pciin. 1999:80:113-119.

32. Bigos S, Bowyer 0, Braen C}. et al. Anile Low BackProblems in .Adults. Rockville. Md: Agenc-y for HealthCare Policy and Research: 1994, Clinical PracticeGuideline No. 14: AHCPR publication 95-0642.

33. Waddell G, Feder G, Mclntosh A, et ai. Low BackPain Lvidemx Review. London, Ivngland: Royal Collegeof General Practitioners: 1996,

34. Hagen KB, Hilde G, Jamtvedl G. Winnem MFThe Cochrajie review of advice to ,stay active as asingle treatment foi- low baek pain and sciatica. Spine.2002:27:173(1-1741.

35. Hurwitz VA., Morgeiistem 11, Harbei- P, et al, Arandonii/.ed tiial of medical care with and withoutphysical therapy and rhinipractic care with and with-out physical modalities for patients with low-hack pain:six-month follow-up otitcomes from the UCLA LowBack Pain Study, Spine. 2002:27:2193-2204,

36. Strong J. (\sliton R, Chant D, Pain intensity mea-surement in chmnic low hack pain. Clin j Pain. 1991:7209-218.

37. Rolaiid M. Mtirris R, A study of the natural his-tory of back pain—pait I: development of a reliable andsensitive measure of disability in low back pain. Spine.1983:8:141-150.

38. Doyo RA. Compaj-adve validity of Sickness ImpactProfile and shorter scales for functional assessment inlow hack pain. Spine. 198(i:l l :95l-934.

39. Jensen MP. Strom SF,, Turner JA, Romano JM, Va-lidity of the Sickness Impact I'nilTle Roland scale as ameasure of dysHmction in chronic pain patients. Pain.1992:50:157-162,

40. Deyo Ry\, Centor RM. Assessing responsiveness f>fftinctiona! scales to clinical change: analogy to diagnos-tic test performance-y C/imn!f Dis. 1986:39:897-906,

41. Hsieh CJ, Phillips RB, ,'\dams AH. Pope MH.I'unctional outcomes of low baek pain: comparison offour Ireatment giT)ups in a randomizt-d eontrolled trial,JManipukitire Phi/siol Ther. 1992:15:4-9,

42. Sti-atford PVV. Binkley J, Solomon V. Gill C, I inchV. Assessing change over time in patienLs with lowhack pain. Phys Vier 1994:74:528-533,

43. Bouter LM, van Tulder MW, Koes BW, N!ethod-ologic issues in low back j}ain research in primary care,Spme. 1998:23:2014-2020.

44. Hashemi L, Webster BS, Clancy FA. Trends indisability dtiration and cost of workers' compensationlow back pain claims (1988-1996)./Ocni/) EnvironMed 1998:40:1110-1119.

45. McHoniey CA. Ware JE. Construction and valida-tion ol' an alternate fonn general mental healdi scalefor the Medical Outeomes Study Short-f'onii I iealthSurvey. Med Care. l995;33:15-28.

46. Ware JE-:. Sheriioume CD. The MOS :56-ItemShort-Form Health Survey (SF-36). I: conceptual Irame-work and item selection. Med Cure. I992;3O:473-483.

47. Ware JE. SF-3(i Health Survey upilate. Sjnne.2000:23:3130-3139.

48. Berwick DM. Muiphy JM. Goldman PA, et al. Per-fomianee of a five-item mental heaith screening test.Med Core. 1991:29:169-176.

October 2005, Vol 95, No, 10 I American Journal of Public Health Hurwitz et al. \ Peer Reviewed I Research and Practice I 1823

RESEAn.,

49. .'\insworth BE. HaskeU WL, Whitt MC, et al.Compendium of physical activities: aii update of activ-ity codes and MET intensities. Med Sa Spofts Exerc.2000;32(suppi|:S498-S504.

50. Wallslon KA, Wallston BS, DeVellis R. Develop-ment of the Multidimensional Health Locus of Control(MHLC) Scales. Health Educ Mono^: 1978;6:160-170.

51. Zeger SL, Liang K-Y. Longitudinal data analysisfor discrete and continuous outcomes. Biometrics.1986;42:121-I3O,

52. Liang K-Y, Zeger SL. Longittidinal data analysisusing generalized linear models. Biometrika. 1986:73:13-22.

5'3. Diggle PJ, Liang K-Y, Zeger SL. Analysis of Longi-tudinal Data. New York, NY: Oxford Univereity PressInc; 1994.

54. SAS for Windows 8.1. Cary, NC: SAS Institute Inc;2000.

55. SAS/STAT Software: Version 8. Cary. NC: SAS In-stitute Inc; 2000.

56. Coulter ID, Hur^^it? KL, Adams AH, Genovese BJ,Hays R, Shekelle PG. I^tients using chiropractors inNorth America: who are they, and why are they in chi-ropractic care? Spine. 2002;27:291-29B.

57. Carey TS. Garrett J, Jackman A, McLaughlin C,Fryer J, Smucker DR. The outcomes and costs of caiefor acute low back pain among patients seen by pri-mary care practitioners, chiropractors, and orthopedicsurgeons. NEnglfMed 1995;333:913-917.

58. I lays RD, Brown JA, Spritzer KL, et al. Memberratings of healtli care provided by 48 pbysidan groups.Arch Intem Med. 1998;i58:785-79O,

59. Riihimaki H. Low back pain: its origin and risk in-dicators. Scand] Work Environ Health. 1991 ;17:81-90

60. Power C, Prank J, Heitzman C, Schicrhout G, Li L.Predictors of low back paiji onset in a prospective Brit-ish study,/im/Pui/ic WeaM. 2001 ;91:1671-1678.

61- Battie MC, Bigos SJ, Fisher LD. et al. A prospec-tive study of tbe role of cardiovascular risk factors andfitness in industrial back pain complaints. Spine. 1989;14:141-147.

62. Holmstrom KB, Lindeli L, Mortitz U. Low backand neck/shoulder pain in construction workers: occu-pational workload and psychosocial risk factors. Spine.1992;!7:663-671.

63. TroupJDG, Martin JW, i.loyd DCHK Back pain inindustry: a prospective study. Spine. 1981:6:61-69.

64. van Poppel MNM, Koes BW, Smid T, et al. A sys-tematic review of controlled clinical trials on the pre-vention of back pain in industry. Occup Environ Med.1997;54:841-847

65. Hildebrandt VH. Bongere PM. Gui J. van Dijk I J,Kemper HC. The relationship between leisure time,physical activities and musciiloskeletal symptoms anddisability in worker populations. Int Arch Occup EnvironHealth 2000:73:507-518.

66. Waddetl G. ITie Buck Pain Revolution. Edinburgh,Scotland: Churchill Livingston; 1998.

67. Videmaji T, Sama S, Battle MC, et al. The long-term effects of physical loading and exercise lifestyleson back-related symptoms, disabilily, and spinal pathol-ogy among men. Spine, 1995;20:699-709,

68- Blumenthal JA, Babyak MA, Moore KA, et al, Ff-

fects of exercise training on older patients wilh majordepression. Arch Mem Med. 1999;159:2349-2356.

69. Fishbain DA, Cutler R, Rosomoff HL, RosomofTRS. Chronic pain-associated depression: antecedent orconsequence of chronic pain? A review. ClinJ Fain.1997;13:116-137

70. Hurwitz EL, MorgCTistem H. Immediate and long-term effects of immune stimulation: hypothesis linkingthe mimune respoase to subsequent physical and psy-chological wellbeing, Med Hypotheses. 2001 ;56:620-624.

71. Lobstein DD, Rasmussen CL, Dimphy GE, Dun-pby MJ. Beta-cndorphin and components of depressionas powerliil discriminators between joggers and seden-tary middle-aged men. f Psychosom Res. 1989;33:293-305.

72. Lobstein DD, Ismail AH, Rasmussen CL. Beta-en-dorphin and components of emotionality discriminatebetween physically active and sedentary men. Biol Psy-chialry. 1989;26:3-14,

73. Krittayaphong R, Light KC, Golden RN, Pinkel JB,Sheps DS. Relationship among depression scores, beta-eridoiphin, and angina pec-toris during exerase in pa-tients with conmary artery disease. Clin} Pain. 1996;12:126-133.

74. Lobstein DD, Rasmussen CL. Decreases in restingplasma beta-endorpbin and depression scores after en-durance training. / Sports Med Phys Fitness. 1991 ;31;543-551.

75. Schwarz L, Kindermann W, Changes in beta-en-dorphm levels m response to aenjbic and anaerobic ex-ercise. Sports Med 1992;13:25-36.

7b. Goldfarb AH, Jamurtas AZ. Beta-endorphin re-sponse to exercise: an update. Sports Med. 1997;24;8-16.

77 Varrassi G, Bazzano G, Edwards WT. EfTects ofphysical activity on matemal plasma heta-endorphinlevels and perception of laboT' pain. .•Im/ Ohstet Gy-necol. 1989;160:707-712,

78. Janniikli Nl', Iranmanesh A. Russell DC. Effect ofraised |ilasma beta endorphin concentrations on pe-ripberal pain and angina thresholds in patients withstable angina. Heart. 1999;82:204-209.

79. Dn>ste C, Greeiilee MW, Schreck M. Roskamm H.Fvfierimental pain thresholds and plasma beta-endor-phin levels during exercise. Med Sd Sports Exerc. 1991;23:334-342,

80. Nehlsen-Cannarella S. Fagoaga 0. Folz J. GrindeS, Hisey C, 'ITiorpe R. F'ighting, fleeing and having fun:the immunology of physical activity. Intf Sports Med.1997;!8(suppl 1):S8-S2I.

81. Harte JL. Fifert GH, Smith R. The effects of run-ning and meditation on beta-endorphin, corticotropin-releasing honnone and cortisol in plasma, and onmood. Biol Psychol. 1995:40:251-265.

82. Jacob r, Baras M, Zeev A, Epstein L, Pbysical ac-tivities and low back pain: a community-based study.Med Sn Sports Exerc. 2004;36:9-15.

83. Keen S, Dowell AC, Hurst K, Klaber MofTeti JA,Tovey P. Williams R. Individuals with low back pain:how do they view physical activity? Eam Pract 1999;16:39-45.

Case Studies inPublic HealthEthicsBy Steven S. Coughlin, PhD, Colin L

Soskolne, PhD, and Kenneth W.Goodman, PhD

SuitabJe for cJassroom discussions andprofessional workshops, this book of

edited public health case studies illus-trates the ethical concerns and probJemsin pubiic health research and practice-The sixteen chapters cover privacy andconfidentiality protection, informed con-sent, ethics of randomized trials, the insti-tutional re\'iew board system, scientificmisconduct, conflicting interests, cross-cultural research, genetic discrimination,and other topics- An ijistructor's guide isalso provided at the end-

ISBN 0-87553-232-21997 1182 pages I softcover

526,00 APHA Members$37.00 Nonmembers

plus shipping and handling

ORDER TODAY!

American Public Health AssociationPublication SalesWeb: www.apha.orgE-mail: APHA@pbd.comTel: 888-320-APHAFAX: 888-361-APHA ETOI

1824 Researcti and Practice Peer Reviewed Hurwitz et al. American Journal of Public Health , October 2005, Vol 95, No. 10