You are on page 1of 7

Clinical Rehabilitation 2005; 19: 538 /543

Psychometric evaluation of the Energy Conservation


Strategies Survey
Priyanka S Mallik College of Nursing, University of Illinois at Chicago, Marcia Finlayson Department of Occupational
Therapy, University of Illinois at Chicago, Virgil Mathiowetz Program in Occupational Therapy, University of Minnesota,
Minneapolis and Louis Fogg Department of Psychology and College of Nursing, Rush University, Chicago, IL, USA

Received 28th February 2004; returned for revisions 15th April 2004; revised manuscript accepted 8th June 2004.

Objective: To examine the internal consistency and test /retest reliability of the
Energy Conservation Strategies Survey (ECSS), a tool that was specifically designed
to determine if people with multiple sclerosis who attended the six-week energy
conservation course actually implemented the strategies taught to them.
Design: The instrument ECSS was administered at six weeks and seven weeks post
intervention to evaluate the test /retest reliability, while data from six weeks post
intervention were utilized to assess the internal consistency of the ECSS.
Setting: Community locations in Illinois and Minnesota, USA.
Subjects: Data from 53 participants with multiple sclerosis living in the community
were utilized for the psychometric study.
Measures: Energy Conservation Strategies Survey.
Results: The internal consistency of 0.92 reveals that all the items are consistent
with each other and measure a single construct. Inferential analyses using intraclass
correlation coefficient indicates good test /retest reliability (intraclass correlation
coefficient (ICC) /0.79).
Conclusion: Findings of the study suggest that the ECSS exhibits high internal
consistency and good test /retest reliability. The study implies that the ECSS could
be valuable in measuring changes in behaviour over time among people with multiple
sclerosis after the completion of the six-week energy conservation course provided
by occupational therapists.

Introduction symptoms of multiple sclerosis.2  4 The prevalence


of fatigue among individuals with multiple sclero-
Fatigue is a common symptom for individuals with sis has resulted in extensive research on ways to
multiple sclerosis and is reported as a significant lessen the disabling effects of this difficult
problem by about 80/97% of the individuals with symptom.5,7  10
this disease.1 Fatigue is a complex phenomenon As a part of the therapeutic approach, occupa-
because of its subjectivity, unpredictability and the tional therapists routinely deliver energy conserva-
way it exacerbates other physical and mental tion education to reduce the impact of fatigue on
the lives of the individuals with multiple sclerosis.11
The ultimate goal in teaching the energy conserva-
Address for correspondence: Marcia Finlayson, Department of
Occupational Therapy University of Illinois at Chicago, 1919 tion techniques is to facilitate behaviour change.
West Taylor, Chicago, IL 60612, USA. e-mail: marciaf@uic.edu However, few studies that have examined the use of
# 2005 Edward Arnold (Publishers) Ltd 10.1191/0269215505cr789oa
Evaluation of Energy Conservation Strategies Survey 539

energy conservation techniques3,12 have adminis- study reflected the amendment; so all the partici-
tered an instrument to confirm if participants pants provided informed consent to the psycho-
actually modified their lifestyle in response to metric study by agreeing to volunteer for the
what they learned.13 In addition, those instruments original study.
that have been used are neither widely available in The psychometric study utilized data from 53
the published literature nor have they been sub- participants. A brief summary of key demographic
jected to psychometric evaluation. variables and multiple sclerosis-related character-
The Energy Conservation Strategies Survey istics of these 53 individuals are outlined in
(ECSS) (available from the authors) is designed Table 1.
to identify the energy conservation principles that
individuals with multiple sclerosis utilize after
going through a six-week energy conservation Instrument
education course by Packer et al .14 Since it is a The ECSS is a 14-item self-administered instru-
relatively new survey, no study has yet evaluated ment that measures behaviour outcomes among
its psychometric properties. The purpose of this individuals undergoing the six-week energy con-
study is to evaluate the internal consistency and servation education course by Packer et al .14 ECSS
test/retest reliability of the ECSS. The specific has a cascading format with two layers of ques-
questions addressed in this paper are: tions (Figure 1). The first question asks the
participant whether he or she carried out a
1) Do individuals’ responses to the questions on particular strategy learned during the energy con-
the ECSS remain stable over a one-week servation course. For this question, the participant
period? responds ‘yes’ or ‘no’. The second part of the
2) Do all the items on the ECSS measure the questioning for a given strategy provides the
same construct? participant the opportunity to elaborate on his or
her yes/no response in more detail. If the ‘yes’ is
selected in the first part, the participant then
chooses the number of changes made or the
frequency of utilization of this strategy. Finally,
Methods Table 1 Summary of demographic characteristics of the
participants (n /53)
The participants were recruited for this study from Characteristics N %
an on-going randomized control trial entitled
‘Energy conservation course for persons with Gender
multiple sclerosis’, funded by the National Multi- Male 11 20.8
Female 42 79.2
ple Sclerosis Society (NMSS) through a grant Ethnic group
awarded to Mathiowetz, Finlayson and Matuska White 51 96.2
in 2001. African-American 1 1.9
Although a number of instruments were admi- No response 1 1.9
nistered in the original study, the psychometric Type of multiple sclerosis
Relapsing/remitting 35 66.0
study only utilized the demographic information at Primary progressive 1 1.9
baseline and the ECSS at six weeks post interven- Secondary progressive 11 20.8
tion. In addition, the psychometric study added a Unknown 6 11.3
second administration of the ECSS one week after Employment status
Full time 11 20.8
the original study administered the tool. The Part time 15 28.3
Institutional Review Board approved this addi- Unemployed 22 41.5
tional data collection point (seven weeks post Retired 5 9.4
intervention) based on an amendment to the Other demographic variables Mean (SD) Range
original study protocol. Informed consent for the Age 49.0 (8.3) 29 /70
Years since symptoms first started 14.0 (9.3) 0 /39
original study was obtained at the face-to-face Years since diagnosed 8.1 (6.3) 0 /24
screening session. The consent form of the original
540 PS Mallik et al.

Figure 1 Model of the ECSS questionnaire.

the participant rates how effective this strategy was that are rated, and the dividing this sum by the
in reducing fatigue. total number of implemented items.
If ‘no’ is selected for the initial question, the Only the variable ‘mean effectiveness ’ was de-
participant is asked to identify the reason for not rived to assess the test /retest reliability because it
carrying out the particular strategy from the is important to identify the level of effectiveness of
list that is provided. An ‘other response’ is offered the strategies utilized by the participants rather
in case the participant does not find the appro- than the number of strategies that were implemen-
priate explanation on the list that is provided on ted (count of number of ‘yes’ responses from the
the survey. first layer of ECSS questioning). In the latter case,
The tool was administered six weeks after the there is always a chance of implementing an item
participants completed the intervention, as some that is not at all effective in managing fatigue.
of the strategies required modification of the All the data for the study were analysed using
physical environment and/or involvement of family version 11.0 of the Statistical Package for the
members. Some of the strategies addressed by the
Social Sciences (SPSS version 11.0).15 The analyses
instrument include: change in work heights; alter-
for the study were completed in a series of defined
ing ergonomic positioning of the body; elimination
of the steps of a single task and communication stages.
with family or friends concerning unmet needs.
1) Internal consistency : Internal consistency of
the ECSS was analysed using Cronbach’s a. A
Analyses
Cronbach’s a of 0.8 or greater is considered an
Using the response from the second layer of
indication of good internal consistency.16 In
ECSS, the outcome variable ‘mean effectiveness ’
was conceptualized. This outcome variable was addition, the item-to-total correlation coeffi-
derived by calculating the average effectiveness cient of B/0.2 is an indication that an indivi-
rating across the strategies a participant identified dual item is not consistent with the overall
as having tried (refer back to Figure 1). In the scale.6
second level of ECSS questioning, participants 2) Rank ordering of items : Based on the propor-
who identified ‘yes’ to the first level were asked tion of ‘yes’ responses for a given item, an
to identify how effective a strategy had been for analysis on all the items of ECSS at six weeks
them, using a rating question from 1 to 10 where 1 postintervention and seven weeks postinter-
means not effective and 10 means highly effective vention was carried out. Spearman’s correla-
(refer back to Figure 1). It is calculated by tion was carried out to determine if the rank
summing the effectiveness scores across all those ordering of the items was stable over time.
Evaluation of Energy Conservation Strategies Survey 541

3) Test /retest reliability : Since the data for the Table 2 Cronbach’s a and Item-to-total correlations of items
on the Energy Conservation Strategies Survey (ECSS)
variable ‘mean effectiveness ’ were normally
distributed, test /retest reliability was evalu- Item no. (yes/no) Item-to-total
ated using paired t-tests and intraclass corre- correlations
lation coefficients (ICC). Paired t-tests were
Time 1 Time 2
carried out and the scores from six weeks
postintervention were compared with the 1. Change in work heights 0.59 0.59
scores from seven weeks postintervention 2. Change in the placement of 0.63 0.74
using p B/0.05. In addition, ICCs were also equipment and tools
3. Operation of adaptive tools and 0.59 0.62
carried out on ‘mean effectiveness ’ at p B/0.05 equipment
level of significance. The formula used for 4. Ergonomic positioning of the body 0.41 0.21
computing ICC is given by Winer’s three-step 5. Elimination of the steps of a single task 0.69 0.41
formula.16 6. Assignment of a task or parts of a task 0.79 0.65
to others
7. Communication for unmet needs 0.68 0.72
Using these strategies, stability over time exists if 8. Modification of an activity 0.78 0.69
the paired t-test is a statistically insignificant p - 9. Setting the priorities 0.64 0.75
value and if the ICC has a value of 0.75 or greater. 10. Simplification of tasks 0.75 0.64
11. Balance work and rest 0.59 0.51
12. Change of time for an activity 0.58 0.50
13. Rests in a day 0.64 0.66
14. Rests during tiring tasks 0.69 0.61
Results Cronbach’s a (reliability coefficient) 0.92 0.89

Consistent with the analysis plan, the results are as 3) Test /retest reliability : Paired t-test at p B/0.05
follows: level of significance revealed that mean effec-
tiveness of the participants across the imple-
1) Internal consistency: Cronbach’s a at six mented items was not significantly different at
weeks is 0.92 while that at seven weeks is six weeks and seven weeks with a t-value/
0.89. All the item-to-total correlations for 1.687, p -value /0.099, with df/45. ECSS is,
both the time points are presented in Table therefore, stable over time as there is appar-
2. The table clearly shows that all 14 items ently no change in the mean effectiveness
represent a single construct. In both cases, reported by participants between the two
item no. 4 (ergonomic positioning of the time points. ICC coefficient was calculated
body) was the least consistent at both six and the coefficient of 0.79 supports good
and seven weeks. stability over time.
2) Rank ordering of items : The analysis of the
rank ordering of items (Table 3) revealed that
item no. 4 (ergonomic positioning of the
body) appeared to be the most commonly Discussion
employed strategy by the participants at both
time periods (i.e., high endorsement). Item no. Findings of these analyses revealed that the inter-
3 (operation of adaptive tools and equipment) nal consistency of the instrument is good and that
was the most difficult strategy to implement all items represent the construct of behaviours that
by the participants (i.e., low endorsement). people are supposed to change as a function of the
Spearman’s correlation between the rank training. Item no. 4 (‘Whether you have changed
ordering of items from six weeks and seven the way you position your body to do an activity’9)
weeks revealed a strong correlation. This was the least consistent item with an item /total
suggests that the items have been implemented correlation of 0.21 at seven weeks post intervention
in the same manner and also suggests that the (lowest item /total correlation).
level of implementation of items is stable over There are two potential explanations for the
time. lower item-to-total correlation for this item. First,
542 PS Mallik et al.

Table 3 Rank ordering of items

Strategy no. Items % ‘yes’ time 1 % ‘yes’ time 2

1 Change in work heights 49.1 58.5


2 Change in the placement of equipments and tools 66.0 71.7
3 Operation of adaptive tools and equipments 43.4 52.8
4 Ergonomic positioning of the body 75.5 86.8
5 Elimination of the steps of a single task 58.5 64.2
6 Assignment of tasks to others 58.5 71.7
7 Communication for unmet needs 64.2 67.9
8 Modification of an activity 64.2 69.8
9 Setting the priorities 66.0 79.2
10 Simplification of tasks 69.8 69.8
11 Balance work and rest 71.7 75.5
12 Change of time for an activity 50.9 56.6
13 Rests in a day 69.8 75.5
14 Rests during tiring tasks 71.7 77.4
Correlation between rank ordering of items for time 1 and time 2 0.87
p B/0.001

this finding may reflect poor item wording, since sistent with what is taught in the energy conserva-
currently it may be capturing some other construct tion course. It is interesting to note that item no. 4
that is loosely tied to behaviours that people are (ergonomic positioning of the body) and item no. 3
supposed to change as a function of the training. (operation of adaptive tools and equipment) are
Rephrasing the item might make it fit better with both strategies that relate to ergonomics, yet item
the overall items. Some of the suggestions for no. 4 (ergonomic positioning of the body) was the
rephrasing are: (a) ‘Whether you have followed most easily implemented item for participants,
proper body mechanics to do an activity’ or (b) while item no. 3 (operation of adaptive tools and
‘Whether you have consistently maintained correct equipment) was the most difficult item to imple-
body postures to do an activity’. ment.
A second explanation may be statistical in The striking difference across these two items is
nature. Since most of the participants implemented consistent with the overall findings that behaviour
the item, there is very little variability across the changes requiring personal effort are relatively
sample for this item. Inadequate variability can easy for participants to implement whereas the
result in low correlation coefficients and this items that involve others or require equipment or
statistical artefact may also be playing a role in physical environment changes are more difficult to
the low item-to-total correlation.17 implement. This is also consistent with the findings
Overall, findings of the study suggest that the of the study by Vanage et al .13 during which this
participants are making behaviour changes con- same implementation pattern was also observed
upon administering the Measuring Change Assess-
ment,14 which records adoption of behaviour
Clinical messages changes after the ‘Managing Fatigue’ course.14
This assessment is the only other tool that was
. ECSS can be used as an outcome tool to located that specifically addresses the issues of
measure behaviour change as a result of behavior change in response to energy conserva-
participating in the six-week energy conser- tion education.
vation course designed by Packer et al .14 Another explanation for the low implementation
. The use of the ECSS would help therapists rate of item no. 3 (operation of adaptive tools and
to identify the energy conservation strategies equipment) may be that the participants did not
that clients are having difficulty implement- have an opportunity to try out equipment during
ing. the sessions. The course manual coached the
instructors to distribute catalogues of different
Evaluation of Energy Conservation Strategies Survey 543

adaptive equipment instead of displaying and persons with multiple sclerosis. Arch Phys Med
demonstrating a few to the participants. The Rehabil 2001 ; 82: 449 /56.
participants might have been more keen to fol- 4 Welham L. Occupational therapy for fatigue in
low-up with this strategy had they been exposed to patients with multiple sclerosis. Br J Occup Ther
1995; 58: 507 /509.
the beneficial effects of using some of those
5 Freal JE, Kraft GH, Coryell JK. Symptomatic
adaptive equipments. fatigue in multiple sclerosis. Arch Phys Med
Findings support that the ECSS is stable over Rehabil 1984; 65: 135 /38.
time. Paired t-tests also revealed that there is no 6 Carmines EG, Zeller RA. Reliability and validity
significant change in the way the participants rate assessment . Beverly Hills: Sage Publications, 1979.
the overall effectiveness of the implemented stra- 7 LaBan MM, Martin T, Pechur J, Sarnacki S.
tegies across the one-week time-frame. This de- Physical and occupational therapy in the treatment
monstrates that the instrument ECSS has good of patients with multiple sclerosis. Phys Med
test/retest reliability, which is one of the basic Rehabil Clin N Am 1998; 9: 603 /14.
features of a psychometrically sound instrument. 8 Rosenblum D, Saffir M. Therapeutic and
Future research may want to explore the use of the symptomatic treatment of multiple sclerosis. Phys
Med Rehabil Clin N Am 1998; 3: 587 /601.
energy conservation education and the ECSS with 9 Schapiro RT. Symptom management in multiple
other populations, for example, people with sclerosis, third edition. New York: Demos, 1998.
chronic fatigue syndrome, rheumatoid arthritis 10 Stuifbergen AK, Rogers S. The experience of fatigue
and fibromyalgia. It would also be valuable to and strategies of self-care among persons with
explore whether the ECSS has utility for other multiple sclerosis. Appl Nurs Res 1997; 10: 2 /10.
energy conservation education courses, beyond the 11 Pedretti W, Zoltan B. Occupational therapy practice
one developed by Packer et al .14 skills for physical dysfunction , third edition. St.
Note : A copy of the Energy Conservation Louis: C.V. Mosby, 2004.
Strategies Survey can be obtained by contacting 12 Bowcher H, May M. Occupational therapy for the
Dr Mathiowetz at mathi003@umn.edu management of fatigue in multiple sclerosis. Br J
Occup Ther 1998; 61: 488 /92.
13 Vanage SM, Gilbertson KK, Mathiowetz V. Effects
of an energy conservation course on fatigue impact
References for persons with progressive multiple sclerosis. Am J
Occup Ther 2003; 57: 315 /23.
1 Krupp LB. Mechanisms, measurements and 14 Packer TL, Brink N, Sauriol A. Managing fatigue: a
management of fatigue in multiple sclerosis. six-week course for energy conservation . Tucson,
Multiple sclerosis: clinical challenges and AZ: Therapy Skill Builders, 1995.
controversies. London: Martin Dunitz, 1997: 15 Maxim P. Quantitative research in social sciences.
283 /94. New York: Oxford University Press, 1999.
2 Krupp LB, Alvarez LA, LaRocca NG, Scheinberg 16 Winer BJ. Statistical principles in experimental
LC. Fatigue in multiple sclerosis. Arch Neurol 1988; design . New York: McGraw-Hill, 1962.
45: 435 /37. 17 Munro BH. Statistical methods for health care
3 Mathiowetz V, Matuska KM, Murphy ME. research , fourth edition. Philadelphia: Lippincott,
Efficacy of an energy conservation course for 2001.

You might also like