Research Report

Effects of an Exercise Program on Sick Leave

Due to Back Pain

The purposes of this study were to evaluate the effect of a weekly exercise program
on short-term sick leave (<50 days) attributable to back pain and to determine
whether changes in absenteeism were related to changes in cardiovascularfitnesr;.
Subjects were randomly assigned to an exercise group (n =58) and a control
group (n =53). Sick leave attributable to back pain was determined in the intervention period of 12/2years and a comparable 1z/2yearperiod prior to the study.
In the exercise group, the number of episodes of back pain and the number of
sick-leave days attriburable to back pain in the intervention penod decreased by
over 50%. Absenteeism attributable to back pain increased in the control group.
The decrease in sick leave in the exercise group was not accompanied by any
change in cardiovascularfitnesr;.Suggestionsfor establishing exercise program
are given. [KellettKM, Kellett DA, Nordholm LA. Effects of a n exercise program o n
sick leave due to back pain. Phys Ther. 1991;71:283-293.1

Karin M Kellett
David A Kellett
Lena A Nordholm

Key Words: Absenteeism; Backache; E~ercise,general; Neck and trunk, back.

Back pain is a major health problem

in the industrialized countries, aflicting about 80% of the population at
some time during their lives.' In
1983, in Sweden, which then had a
population of 8.5 million people, 24%
of sick-leave days were due to pain in
the back, neck, and shoulders. In
1988, 54,000 people took early retirement in Sweden, 10,500as a result of
back pain (National Social Insurance
Board, Stockholm, Sweden; personal
communication). In 1986, the cost of
sick leave and early retirement because of back pain amounted to 5.1
billion Swedish crowns (approximately $777 million in US curren~y).~
A

major aim for the health care services,

therefore, is to prevent back pain. In
Sweden, 3% to 4% of the population
suffer from chronic back pain leading
to work incapacity.3 The 10% of patients with back pain having the longest periods of sick leave account for
80% of the total number of sick-leave
days attributable to back pain.4 Thus,
preventing acute back pain from progressing to chronic back pain must
also be our aim.
The effects of conservative treatment
techniques on patients with back pain
are still uncertain.l~5Many patients
have been on sick leave too long be-

K Kellett, BSc, is Superintendent Physiotherapist, Department of Physiotherapy, The Health Cenrre,

Box 305, 522 00 Tidaholm, Sweden. Address correspondence to Ms Kellett.
D Kellett. MB. ChB, is General Practitioner. The Health Centre.

L Nordholm, PhD, is Associate Professor, Gothenburg College of Health and Caring Sciences, Guldhedsgatan 6, 413 20 Gothenburg, Sweden.

fore rehabilitation methods such as

work training are considered. This
delay can cause back pain sufferers to
lose confidence in their ability to
carry out activities such as walking,
standing, sitting, lifting, and carrying.
It becomes increasingly difficult to
convince patients of their physical
capabilities and work capacity when
delays occur.6 Furthermore, it takes a
long time to learn an effective work
technique and to carry out a training
program to attain appropriate
strength, coordination, and flexibility.
Consequently, few patients manage to
carry out a training program for any
length of time by themselves.'
Positive effects from participation in a
back school have been reported in
patients with acute and subacute back
pain. Whether the back school has a
preventive effect on future back pain,
however, is u n ~ e r t a i nIn
. ~ a study of
patients with chronic back pain,
Hembom and Holmstrom9 found

This article was submitted April 24, 1989, and was accepted November 20, 1990

Physical Therapy /Volume 71, Number 4 /April 1991

283 131

some change in the participants' attitudes toward back pain following participation in a back school, but this
attitudinal change had no effect on
the amount of sick leave the participants took.
There is evidence that physical activity
is beneficial for both prevention and
treatment of back pain. Cady and associates1 found among fire fighters
that the most physically fit individuals
had fewer and less costly back injuries than the least physically fit individuals. Activity of the major muscle
groups is generally thought to increase the level of endorphins in
cerebrospinal fluid and to increase
pain tolerance under certain conditions.12 Delivery of nutrients to the
intervertebral disk and articular cartilage is stimulated by physical exercise.l2 Nachemson6 contends that,
when there is damage to muscles,
ligaments, and tendons, physical exercise improves symptoms and enhances healing. He concludes that
progressive physical activity is of benefit to patients with back pain.
Mayer and associates13 have had success in facilitating the return to work
of some patients with long periods of
sick leave. Their approach utilizes
principles of functional training and
measurement and of disregarding
pain. Their patients perform physical
exercise and simulated work tasks.
The focus is on what patients can do
functionally. Patients' efforts in physical activity are encouraged, rather
than the specific treatment of their
symptoms.l3
The Volvo Project in Gothenburg,
Sweden, demonstrated that an activity
program for patients with subacute
lumbar pain (sick leave of 6-8 weeks)
resulted in an earlier return to work
for male participants than for female
participants. The program consisted of
a functional assessment; visits to the
patients' work place; a back school;
and individualized, progressive physical activity. Improved function and
increased activity were encouraged,
rather than a specific treatment of
pain (I Lindstrijm, C ~ h l u n dA,
Nachemson; unpublished research).

Exercise programs to complement a

back school have been introduced in
Sweden. Of the participants in an exercise program specifically designed
for people with back pain in Stockholm, 78% reported an improvement
in their symptoms.14The improvement was independent of the duration of their symptoms and was the
same for all age groups. In a similar
study,l5 the majority of the participants reported a reduction in back
pain; however, it was difficult for
them to continue the exercise program independently.
Although physical activity has been
shown to have these positive effects,
there are no reports of the effects of
physical exercise on short-term absenteeism attributable to back pain (<50
days). Ninety percent of patients with
acute back pain return to work within
6 to 7 weeks16 The major purpose of
this study was to investigate the effect
of an exercise program on short-term
sick leave attributable to back pain. A
secondary purpose was to investigate
whether changes in sick leave are
related to effects on cardiovascular
fitness.

Method
Subjects
The participants in this study were
employees of Marbodal Al3, the biggest
employer in Tidaholm, Sweden, and
Scandinavia's major producer of
htchen units. The management
showed considerable interest in the
project and gave their agreement that
about 60 employees could take part in
an exercise program 1 hour a week
for 1% years during paid working
hours. Information about the project
was sent to all employees of Marbodal
Al3. Criteria for inclusion in the study
were self-reported current or previous
back pain, written commitment to participate in the exercise program during
workmg hours, and a willingness to
exercise at least once a week outside
working hours for 1% years. Exclusion
criteria were any period of sick leave
longer than 50 days, irrespective of
cause, during the 1% years prior to
the study, and medical reasons affect-

ing the employee's ability to participate in the exercise program.

Of 850 employees, 143 volunteered to
participate in the study. Eighteen volunteers were rejected according to
the exclusion criteria. From the remaining 125 volunteers, 58 individuals were randomly selected to form
the exercise group, and the remaining
67 individuals formed the control
group. Ten individuals were not interested in continuing as part of the control group because they would not be
participating in the exercise program,
and 4 control group subjects did not
participate in the initial cardiovascular
fitness test. Thus, the exercise group
consisted of 58 subjects and the control group consisted of 53 subjects.

Research Design
The study used a prospective, randomized, controlled research design
to evaluate the effect of physical exercise intervention on the participants'
sick leave attributable to back pain.
The number of days of sick leave attributable to back pain and the number of episodes of back pain were
recorded during the intervention period (period 2 =November l, 1986April 30, 1988) and compared with
data recorded during a period prior
to intervention (period 1=November
1, 1984April 30, 1986). The timing of
the intervention and baseline periods
was chosen to eliminate the effects of
seasonal variation in sick leave. The
data relating to sick leave were obtained from the register of the National Social Insurance Board in Tidaholm, Sweden. Participants were
guaranteed anonymity. The project
was carried out in Tidaholm by personnel from The Primary Health Care
Centre and the industrial nurse at the
factory of Marbodal AB.

Procedure
Exercise program. The exercise
group was divided into two subgroups, and each subgroup participated in the exercise program once a
week during working hours. The program was conducted to music, and, to
ensure variety, the program was

Physical The:rapy /Volume 71, Number 4 /April 1991

changed every 6 months. One of the

three different exercise programs is
described in the Appendix. During
the first 10 weeks, the program was
carried out at low intensity. Participants were encouraged to perform
the exercises at a speed and intensity
level that did not feel strenuous or
cause increased pain after the session.
The main aim during these first 10
weeks was for the participants to become accustomed to movements of
the arms, legs, and trunk in different
positions (eg, standing upright; standing on hands and knees; lying on
back, side, and front). Emphasis was
placed on the harmony and rhythm of
the movements. The subjects were
allowed to stop and rest if an exercise
was painful, and they were encouraged to find similar alternative movements that felt better for them. This
meant that jumping and jogging exercises could be included.
When the participants had learned an
individual intensity level that did not
cause discomfort after the training
session, the intensity of the program
was progressively increased. The exercise intensity was progressively increased by instructions such as: "It is
now time for you to increase your
physical efforts so that you become
warm and sweaty and feel a bit tired
in your muscles afterward. If you feel
pain anywhere after the session, you
have done too much. On the other
hand, if you do not feel tired afterward, you have done too little." There
were no assessments of heart rate to
estimate exercise intensity. Instead, it
was emphasized that the participants
must find an intensity level that suited
them best. Consequently, considerable differences were observed in the
participants' exercise intensity levels.
Each session started with a warm-up
and gentle stretching exercises and
continued with alternate strengthening and cardiovascular fitness exercises. Coordination exercises were
included throughout the session. The
session ended with a "warm-down"

and specific stretching exercises

(Appendix).
Each exercise session lasted 30 to 35
minutes and was followed by 5 to 10
minutes of relaxation, during which
relaxing music was played. The participants were then instructed to tense
their muscles for about 5 seconds, to
relax for about 10 seconds, and finally
to feel the diffeence. This procedure
was performed for muscle groups
throughout the body.
Two thirds of the sessions were led
by a physical therapist assistant and a
physical fitness instructor. The remaining third were led by a physical
therapist (KMK), who started each
class with 10 minutes of lectures relating to theories associated with traditional back schools.'9 Emphasis was
placed on achieving a good prognosis
for acute back pain. In the event of a
recurrence of acute back pain, the
participants were encouraged to be
active as quickly as possible, starting
with walking and swimming. Bed rest
was not suggested for more than a
few hours at a time. The program
continued for 1% years, with breaks
for the industrial holidays in 1987 (4
weeks) and Christmas 1986 and 1987
(3 weeks each). There was an average
of 77% attendance at each session.

Cardiovascular fitness. Before and

after the project, the cardiovascular
fitness of both the exercise and the
control groups was tested using a
computerized exercise cycle.*Each
subject's oxygen uptake capacity (in
milliliters of oxygen per kilogram of
body weight per minute) was estimated according to the method described by htrand.18 This is a submaximal exercise test from which the
maximal oxygen uptake is estimated.
The subjects cycled at a measured
exercise intensity level while the heart
rate was continuously monitored until
a steady state of heart rate was
reached above a target level of 120
bpm. The oxygen uptake capacity was
then correlated to the subject's age,
sex, and weight.18All fitness tests
were administered by a specially
trained health worker with experience in fitness training.
Self-report data. At the end of the
project, participants in the exercise
group reported on the subjective effect of the exercise program on their
back pain by anonymously answering
the question, "How has the exercise
program affected your back pain?"
There were five possible answers:
"much better," "better," "unchanged,"
"worse," and "much worse."
Data Analysis

Additional exercise. In addition to

the organized exercise program, participants gave a written commitment to
exercise at least once a week on their
own outside working hours. Each exercise session was expected to consist
of walking 3 km or for 30 minutes,
cycling 10 km or for 30 minutes, skiing
3 km or for 30 minutes, jogging 2.5
km or for 20 minutes, or performing a
comparable activity for 30 minutes.
The participants chose their own level
of exercise intensity, and they kept a
diary of their activities, which was
handed in at the end of the project. No
further control of their additional exercise was instituted.

'Dynavit, Scheelevagen, S-302 39 Halmstad, Sweden

Physical Therapy/Volume 71, Number 4/April 1991

To establish the equivalence of the

exercise group and the control group
prior to intervention, Student's t tests
for independent groups were carried
out on the following variables: age,
body mass index (BMI) (calculated as
weight [in kilograms] divided by
length [in square meters]), cardlovascular fitness, sick days attributable to
back pain in period 1, and episodes
of back pain in period 1. Comparisons of the sick-leave data between
period 1 and period 2 were performed using paired t tests for both
the exercise and the control groups.
Changes in cardiovascular fitness following intervention were also tested
by paired t tests in both groups. In
addition, the change scores (period
2 -period 1) for the variables sick
days attributable to back pain and episodes of back pain were compared

Table 1. Comparison of Means, Standard Deviations, and Results o f t Tests to Ascertain Group Equivalence Prior to Internention
Exerclse Group (n=37)
Variable

SD

Range

Control Group (n=48)

X
SD

Age (Y)
BMla

41.24
24.65

10.22
2.79

26.00-58.00
19.6131.10

42.13
25.06

9.95
2.87

25.0CL62.00
19.96-32.37

-.40
-.65

83
83

NS
NS

Cardiovascular
fitnessb

43.28

9.79

27.00-64.00

44.36

12.30

23.0G65.00

-.41

83

NS

Range

df

OBMI=budy mass index (in kilograms per square meter)

b~ardiovascularfitness=estimated maximal oxygen uptake (in milliliters of oxygen per kilogram of body weight per minute), according to the method
described by &trand.lN

between the exercise group and the

control group, using t tests for independent groups.

each group) and the ratio between

managerial and shop floor workers
(eg, 1:2).

Results

During the intervention period, the

total dropout rate for the two groups
was 23%. Of the 58 participants in the
exercise group, 6 individuals withdrew from the exercise program for
medical reasons (ie, high blood pressure, spondylolisthesis, epicondylitis,
fracture, angina pectoris, and pyelonephritis). An additional 4 subjects
found it too inconvenient to leave
their work, and 5 subjects withdrew
from the program because of an increase in pain (2 because of back
pain and 3 because of neck pain).
One participant started a study course
that prevented continued attendance,
1 left the firm, 1 commuted with co-

The exercise group and the control

group were first compared to ascertain the equivalence of the two
groups prior to intervention
(Tabs. 1 and 2). As shown in Table 1,
there were no significant differences
between the two groups for the variables age, BMI, or cardiovascular fitness. Table 2 shows that the groups
did not differ significantly for either
sick days attributable to back pain or
episodes of back pain prior to intervention (period 1). The groups were
equivalent with respect to sex distribution (ie, women composed 30% of

workers and found it difficult to be

ready in time after the session, 1
started working overtime and did not
want to reduce his working hours by
continuing to participate in the exercise program, 1 started football training and thought he had sufficient exercise, and 1 lost interest in the
program. By the end of the project,
37 participants (64%) remained in the
exercise group.
Of the 53 subjects in the control
group, 3 were excluded because of
long periods of sick leave (>50 days)
in period 2. Two control group s u b
jects withdrew because of pregnancy.
The control group thus consisted of
48 subjects (91%) by the end of the
project.

Table 2. Comparison of Means, Standard Deviations, and Results o f t Tests of Sick-Leave Variablesfor Exercise and Control
Groups
Exercise Group (n=37)

Control Group (n=48)

TZ

SD

Range

5.59

12.54

0.0-65.00

Period 2

2.73

7.63

0.0-41 .OO

Difference

2.86

7.62

Variable

iz

SD

Range

2.50

5.98

0.0-32.00

4.13

9.32

0.0-40.00

df

1.50

83

NS

-.73

83

NS

2.55

83

<.02

1.27

83

NS

83

NS

83

1.05

Sick days attributable

to back pain
Period 1

-1.63

8.40

Episodes of back pain

Period 1

0.54

0.93

0.03.00

0.33

0.60

0.0-2.00

Period 2

0.27

0.61

0.0-3.00

0.52

1.07

0.0-4.00

Difference

0.27

0.96

34 / 286

-0.19

1.10

-1.27
2.01

Physical Therapy /Volume 71, Number 4 /April 1991

Sick Leave
Thirty-two percent of the exercise
group took sick leave because of back
pain during period 1, and 2 1.6% took
sick leave during period 2. Twentyseven percent of the control group
took sick leave because of back pain
in period 1, and 29% took sick leave
during period 2. Sick-leave data for
the two groups and the two datacollection periods are presented in
Table 2.
Table 2 shows a significant difference
between the 51.2% decrease in sick
days attributable to back pain in the
exercise group and the 65% increase
in the control group. A similar significant difference in change scores is
noted fc~rthe other sick-leave variable,
episodes of back pain (Tab. 2). From
the data presented in Table 2, the
changes in each group were examined separately using paired t tests. In
the exercise group, the number of
sick days attributable to back pain
decreased (t=2.29, df=36, P < .05),
but there was no significant change in
the number of episodes of back pain
(t= 1.79,df=36, P=NS). The control
group dld not change significantly in
either sick days attributable to back
pain (t= - 1.34,df=47, P=NS) or episodes of back pain (t= - 1.18, df=47,
P=NS). 'Thus, it appears that the significant difference in change scores
on the sick-days variable can mainly
be attributed to changes in the exercise group. With respect to the sickness episode variable, the difference
in change scores was produced because the groups changed in different
directions, even though each such
change was not significant.
Records of the National Health Insurance Ofice revealed no difference in
the patterns of absenteeism in exercise group subjects who withdrew
from the project because of increased
pain and the four subjects who withdrew because they found it inconvenient to leave work compared with
the rest of the exercise group. Their
reduction in number of sick-leave
days attributable to back pain in period 2 w~~67%. The reduction in the
number of episodes of back pain for

these nine participants was 50% in

period 2.

Additional Exercise
The exercise group performed exercise on their own at least once a
week during the whole of period 2.
The majority (n=25) chose walking
combined with either cycling, jogging,
or skiing. A few participants chose
football, dancing, tennis, or other
exercise.

Cardiovascular Fitness
The cardiovascular fitness test revealed no significant change in estimated oxygen uptake capacity between the two data-collection periods
in the exercise group @reinternention: X=43.78 mL 0,-kg-'-min-l,
SD=9.79; postintervention: X=43.78
mL O,.kg-l.min-l, SD= 11.31,
t = -.48, df=35, P=NS). A significant
decrease in cardiovascular fitness,
however, was observed in the control
group (preintervention: X=44.36 mL
~ , . k ~ - ' - m i n - SD=
~ , 12.31;postintervention: X=41.86 mL ~ , - k ~ - ~ - m i n - ' ,
SD=10.74, t=2.46, df=47, P<.02).
With consideration taken for age and
sex, both groups were found to have
average to above-average cardiovascular fitness on both test occasions.ls

Self-Report Data
All participants in the exercise group
replied to the question on the subjective change in their back pain. Fortysix percent of the respondents reported that they felt much better, 35%
reported that they felt better, 14%
reported no change, and 5% reported
that they felt worse. Thus, 81% of the
participants reported improvements
in symptoms following participation
in the exercise program.

This study has shown that an exercise

program carried out once a week for
1% years by people with back pain
reduced their number of sick-leave
days and their number of episodes of
sickness attributable to back pain by
over 50% compared with a 1%-year

Physical 'Therapy /Volume 71, Number 4 /'April 1991

period prior to intervention. In contrast, the control group's amount of

sick leave attributable to back pain
increased during the same period.
The analysis of nine exercise group
subjects who withdrew from the program (four stating inconvenience to
leave work and five stating increased
pain) revealed that they had a similar
pattern of absenteeism attributable to
back pain compared with the other
exercise group subjects. This finding
might be explained by the fact that six
of these subjects participated in the
exercise program for more than 1
year and three of these subjects participated between 6 and 12 months.
'I'hus, had they remained in the exercise group, the results apparently
would not have been altered.
Although there are no readily available statistics on the local change in
sick leave attributable to back pain
between the two data-collection periods studied in this project, the total
level of absenteeism in Tidaholm,
Sweden, has increased. In 1985, the
average number of sick days per employee each year was 16.9; in 1987,
the average number of sick days had
risen to 17.8 (National Social Insurance Board, Tidaholm, Sweden; personal communication). Despite this
general increase in absenteeism, the
exercise group showed a decrease in
sick leave during this time.
In the exercise group, the cardiovascular fitness of the participants remained at an average to high level
after the project period. One training
session a week was not enough to
increase the level of fitness for this
group. The participants' own training,
which was mainly walking, apparently
was not carried out at a sufficient intensity to raise the level of fitness.
Cady and associates1 have shown an
inverse relationship between cardiovascular fitness and subsequent back
injuries. In our study, although the
participants did not significantly alter
their oxygen uptake capacity, they did
reduce their amount of sick leave attributable to back pain.
The control group's cardiovascular
fitness decreased significantly. We con-

Appendix Exercise Program"

Big steps from side to side. Bend

Stand with legs apart and awns

stretched upward. Lilt left knee up
and meet with right elbow.

together. Right awn follows upward

'Nu lever jag igen'

Beat: 130/mn

anns down
GENTLE
STRETCHING
Music:
This Is My Life'
3.20 min
Beat: 56/mn
Bend foot up. Put heel down in
fmnt, keeping body weight on
s6ghtly bent [ear leg. Lean fomard
with straight back to feel the
stretch in hamstring musdea. Hold
e few seconds. Bend knees and
awns. Change bp.
STRENGTHENING
Music:
'Une Amore
Grande'
3.0 min
Beat: Wmin

fE3tLa
Lie down, and put one foot on the
opposite thigh. Raise head and
shoulders. Lower and repeat.

CARDIOVASCULAR
FITNESS
Music:
'Hb;gt aver hay&
245 m n
Beat: 1I l n i n

Big steps from side to side. Cross

left leg in front when stepping to
the right and vice versa Swing
anm sideways.

Clap hands under knee. Straighten

Jogging on the spot. Arms relaxed.
arms upward. Change legs.

sideways.

h~i!
Stand with legs apart and knees
bent. L i i m,and hrrn rp and
b d w a r d to the right. Clap knees.
Turn up and backward to the left.

Stand with legs apart and arms up.

b n d one knee at a fime to streah
the oppasite adductors. Arms to
the side on bendng, and dap over
head when straightening.

&&

&%
Lie with bent knees. Tghten
butt&
and lift pehs. Lower and
repeat.

One leg badtward. Touch lkor

with hands. Streighten up, and
raise arms. Change legs.

?$qf&

Lie with legs fully bent. Raise head

Stand on hands and knees. DO
and shoulders. shaighten knees,
pushups.
and stretch arms between legs.

h+J

Streth one leg bdward. keeping

weight on fmnt bent leg. Sketch
arms forward and clap hands.
Change legs.

Big steps from side to side. Cross

(eft leg behindon stepping to the
%ht and vica versa. Swing e m
sideways.

R
Stand with legs spart. Bend knees
and touch b o r . Straighten up.

Physical Therapy /Volume 71, Number 4 /April 1991

Appendix Exercise Programa

STRENGTHENING
Music: 'My o Mf
2.55 min
Beat: 74lmin

Ue on *M side with knees fully

bent and ight arm under head.
M w t left ebow and knee.
Straighten out arm and leg.
Change side halfway through.
CARDK)VASCULAR
FITNESS
Music: 7.L.C.'
2.55 min
Beat: 144/&n

&&
Stand on left knee and hand with
right b a n d arm Sbtched out.
Bling tight knee and ebow
together. Stretch again. Change
side h a h a y through

SPECIFIC
STRETCHING
Music:
(background)
V~valdi'
4.45 min

Sit with bent knees with arms

supporting behind. Lilt peke.
Lwer and repeat.

Hop and lwist in pairs.

WARMING DOWN
Music: TIAmo'
3.0 n i n
Beat: W n i n

Stand on hands and knees. Lift left

leg and rigM am. Change sides.

--

Hop and kidc alternate leg

sideways in pairs.

Lie on *M side with knees fully

bent and hands behind the neck
Turn left ebow, head, and
shoulders badward. Return.
Repeaf and change side halfway
through.

Lie on front. Stretch

alternately.

fowd

$4
Hop around arm in arm. Raise and Holding hands in pairs. Bend down
bwer free arm. Change d i r d o n t o w d the fbor. keeping badc
h e h a y through.
shaigM. Reach up high.

a
Sit on fulty bent knees. Bend
fornard and relax abrut 10
seconds.

A A ha
Stretch ilopsoas muscle lor about
20 seconds each side. Keep
bultocksljghtened to avoid
Stretch hamtn'ng muscles for
extending M a r spine.
about 20 seconds each side.

Lift left arm over the head. and

bend the head toward the tight.
Keep head in this position and
r d r n arm behind b& stretch
trapedus me
upper pa
for
about 20
R~~~~~with
other side.

Stre*
b m r d and hold a
few seconds. Relax, and let head
bend forward, restinga m on
knees for about 20 m e .

SPECIFIC
STRETCHING
(continued)

=
=
&
-3

Clasp arms around knees. and pull

up head to form a ball. Hold a few
seconds. Stretch out to fun length.

"All exercises are carried out in time to the music. The time for each song is divided equally among the four exercises. The exercises are adapted
from a program made by "Friskis and Svettis," Sodra Hamngatan 19-21, 411 14 Gothenburg, Sweden. Illustrations by Gunilla Sundstrom.

sidered this decrease in mean oxygen

uptake capacity from 44.4 to 41.9 mL
0,.kg-'.min-',
however, to be minor.
The control group was still found to
have average to above-average cardiovascular fitness. We believe that this
decrease is of minor importance to the
control group's increase in sick leave
compared with the general trend of
increased sick leave in Tidaholm as
reported previously.
We did not make any objective measurements of changes in back pain
inthis study; nevertheless, 81% of the
participants in the exercise group reported a subjective improvement in
back pain. This finding is in concordance with the results of a study by
Garderud,14who reported an improvement in 78% of the participants
in a similar exercise group. Although
we took no objective measurements
to determine why the participants'
sick leave attributable to back pain
decreased, several factors in addition
to the exercise protocol may have
been influential. We believe it is unlikely that either the back school o r
the increase in general exercise levels
could have had any major effect. The
back school has not been shown to
have any preventive effect on sick
leave attributable to back pain,8 and
the participants did not increase their
exercise levels to a degree that affected their cardiovascular fitness.
The effects of group support, improved attitudes toward the employer,
reassurance about prognosis, and
nonspecific intervention effects (eg,
life-style changes) may have contributed to the positive results obtained
in our study. We believe, however,
that the exercises had a major influence on the outcome. Although the
exercises were not specific, because
they could have been varied in many
different ways, we believe that, because of the design of the exercise
program and the way in which it was
carried out, the results are attributable
to the program. The participants were
given responsibility to adapt exercises
and to find an exercise intensity level
that suited them. We believe that this
responsibility helped the participants
develop a better capacity to adapt

postures in different functional activities, such as walking, standing, lifting,

carrying, and sitting, which may have
improved their body mechanics in
their daily activities. In our opinion,
the participants probably learned postures and movements that caused less
strain on their back.
Traditional management of patients
with back pain with rest and passive
treatment has not been shown to reduce the length of sick leave or to
prevent the recurrence of back
pain.l>5Our study demonstrates that
physical activity is beneficial for reducing short periods of sick leave
attributable to back pain (<50 days).
There is also evidence that physical
activity is beneficial in the management of subacute and chronic back
pain, as shown in the Volvo project
(I Lindstrom, C ~ h l u n dA, Nachemson; unpublished research) and the
work of Mayer and associates.l3

Clinical implications
We believe that therapeutic measures
be diverted away from passive
treatment and instead directed at mobilizing- patients
more actively. We
.
suggest that an exercise program can
be integrated into the traditional back
school and that this program should
continue for 2 to 3 months so that
participants can establish a training
routine.
Physical therapists can interact with
instructors of physical fitness classes
to establish cooperation so that patients can continue with regular exercise in organized forms. Most people
do not succeed in continuing with a
training program on their own,' and
it is therefore important that they are
given support and encouragement to
participate in group activities. It is
difficult to motivate many people on a
voluntary basis; therefore, the possibility of establishing exercise programs during
.
.
working hours
be considered.

Conclusions
This study has shown that a weekly
exercise program has resulted in a

reduction of sick leave for people

with relatively short (<50 days) episodes of back pain. The majority of
patients who develop chronic back
pain have previously had repeated
short episodes of back pain. This
study demonstrated that it is possible
to reduce sick leave by 50%; therefore, it is probable that the number of
patients developing chronic back pain
can also be reduced. An investment in
exercise programs for people with
back pain could lead to considerable
benefits for the employer, society, and
individuals with back pain.
Acknowledgments

We gratefully acknowledge the contributions of Elisabeth Carlsson, EwaBritt Carlsson, Siw Gustavsson, Agneta
Hendrikson, Maria Jansson, and h e
Lindstrom for their collaboration.
References
1 Waddell G. A new clinical model for the
treatment of low-back oain. SDine.
1987;12:632444.
2 Axelsson S, Bruzelius N. Samhallsekonorniska kostnader for siukfrinvaro och fonid.
spensionering orsakade av b e s ~ fran
r
landrygg. ZHE-information. 1988:2.
3 Sandstrom J. On Chronic Low Back Pain
Gothenburg, Sweden: University of Gothenburg; 1985. Dissertation.
4 Spengler D, Bigos S, Martin N, et al. Back
injuries in industry-a retrospective study, 1:
overview and cost analysis. Spine. 1986;ll:
241-245.
5 Spitzer W, LeBlanc F, Dupuis M, et al. Scientific approach to the assessment and management of activity-related spinal disorders. Spine.
1987;12(suppl 1):7S.
6 Nachemson A. Work for all, for those with
low back pain as well. Clin Orthop.
1983;179:77-85.
7 Henrysson S, Wedman J. Regelbundna motionsvanor-umeiexperiment
1: att forandra
levnadssatt. RiksbankensJuhileums/ond RJ.
1982:3.
8 Berauuist-Ullman M. Larsson U. Acute low
back pain in industly: a controlled prospective
study with special reference to therapy and
confounding factors. Acta OrthoD Scand
[suppl]. 197?;170:1-109.
9 Hemborn B, Holmstrom E. Utvardering av
ryggskolan: en litteraturgenomgang. Liikartidningen. 1988;50:44434444.
lo cadv L, BischoE D, 03Connell E, et al,
strength and fitness and subsecluen; back iniuties i n firefighters. J Occup ~ e d1979;21:269.
272.
11 Terenius L. Endorphins and pain. Front
Horm Res. 1981;8:162-177.
-

- -

Physical Therapy /Volume 71, Number 4 /April 1991

1 2 Holm S, Nachemson A. Variations in the

nutrition of the canine intervertebral disc induced by motion. Spine. 1983;8:866874.
13 Mayer TG, Gatchel RJ, Kitchino N, et al.
Objective assessment of spine function following industrial injury. Spine. 1985;10:482-493.

14 Garderud A. Friskis och Suettis ryggympa:

uilka deltar och med uilken effekt? Stockholm,
Sweden: Metropol Wretagshalsov2rd; 1986.
15 Linderoth L. Traning vid ryggbesvar. SPRI
U 20849; 1986.
16 Nachemson A. Advances in low back pain.
Clin Onhop. 1985;200:266278.

17 Zachrisson-Forsell M. The back school.

Spine. 1981;6:104-106.
18 k t r a n d I. Aerobic work capacity in men
and women with special reference to age. Acla
Physiol Scand [Suppll. 1960;49:169.

Intervention was conducted over a

1%-year period, which is justifiable
and commendable in allowing an adequate time period to demonstrate
treatment effect. Control subjects and
those assigned to the exercise program were selected from the same
population of workers, all sharing a
current or past history of self-reported
back pain. Individuals with a "healthy
back" (ie, no previous or current
back pain) were not participants in
this study. We therefore do not know
whether a similar program would be
helpful in reducing sick days attributable to back pain in workers with no
prior history of back pain.

paid toward medical and disability

expenses.3 Because one's prior history
of back pain is critical to outcome
and there exists a wide range of back
pain disorders, a more detailed description of the subject population in
terms of physical impairment, disability, or job category would have been
desirable in this report. A selfreported history of back pain as conducted in this study is likely to include an extremely wide range of
pack pain conditions. Further studies
are needed to address the unanswered question of exactly which segment of the population might benefit
from such a program in terms of either their back pain history or their
occupation.

Commentary

The authors of this article are to be

commended for conducting their
study using a research design that
provides the best assurance of concluding that their results are attributable to the intervention. They used a
randomized clinical trial with concurrent control, which has not been selected often enough in other studies
of back pain, even though it offers the
highest likelihood for concluding causality. According to a recent report
from an international study group
convened to formulate policy for future research in back pain,' it was
strongly urged that (1) more experimental studies be conducted to address the limitations of purely observational studies and (2) clinical trials
@referably randomized) with independent concurrent controls be implemented to reduce the confusion
surrounding efficacy of treatment. Previous research conducted to identify
the effect of clinical intervention on
outcome of back pain has more often
than not been seriously flawed by
poor research design or too small a
sample size. These flaws result in inconclusive data that are not generalizable to a larger population. It is worth
noting that the authors of this study
have attempted to correct these limitations, even though the design of the
study is often difficult and costly to
conduct.
The authors have chosen to study
whether a weekly exercise program
can reduce sick days attributable to
back pair1 in a population of workers.

It is not clear from the authors' description of the work force what percentage of Marbodal AB employees
have a positive history of back pain or
what the back pain incidence rate is
among workers with no previous history of back pain. Future studies
might address this question of
whether similar exercise programs
should include all employees and not
just those with an identifiable back
problem. It should be pointed out
that a number of studies have identified prior history of back pain as a
significant risk factor for back pain.*
The results of this study are therefore
impressive in that they emanated
from a high-risk group that is likely to
be resistant to intervention. From a
purely economical point of view,
workers with recurrent episodes of
back pain also command a disproportionately greater percentage of money

Physical 'Therapy /Volume 71, Number 4 / April 1991

In addition to the important question

of who might benefit from intervention of this kind, it is also of interest
to question how treatment outcome is
measured and interpreted. In this
study, both the number of episodes of
back pain and number of sick days
attributable to back pain were analyzed. Data were collected during the
intervention period as well as during
an equivalent period prior to intervention. The measurements are appropriate for a study of this kind;
however, I am less enthusiastic about
the way comparisons were made
across groups or as a function of time
period within a group. The most serious flaw is that comparisons between
means were conducted on data that
were not normally distributed. I am
as lenient as the next person when it