Journal ofGerontologv: PSYCHOLOGICAL SCIENCES
1996, Vol. 5IB. No. l", P24-P29
Copyright 1996 by The Geromological Society of America
Self-Efficacy and Pain in Disability
With Osteoarthritis of the Knee
W. Jack Rejeski,1 Tim Craven,2 Walter H. Ettinger, Jr.,2
Mary McFarlane,2 and Sally Shumaker2
'Department of Health and Sport Science, Wake Forest University.
Public Health Sciences, Bowman Gray School of Medicine of Wake Forest University.
2
This study examined the relationship between self-efficacy beliefs and pain during the performance of stair climbing
and lifting!carrying tasks on speed of movement, ratings of task difficulty, and perceived task ability in a group of
patients with osteoarthritis (OA) of the knee. Seventy-nine patients with knee OA completed the tasks in a controlled
laboratory setting. Before completing each task, patients' self-efficacy was assessed; following task performance they
rated (a) the most intense knee pain experienced, (b) the difficulty of the task, and (c) their perceived ability as they
performed each task. Results demonstrated that, even after controlling for physical function, self-efficacy, and knee
pain during performance, each contributed significantly to understanding either speed of movement or self-reported
ratings of task difficulty and perceived ability.
M
OST geriatricians agree that functional impairment and
disability are as critical to medical care as is the
identification of underlying pathology (Minaire, 1992; Nagi,
1965). For example, Guralnik, Branch, Cummings, and
Curb (1989) suggest that "By understanding the functional
capacities of patients, caregivers are better able to judge
disease severity, the impact of multiple morbidity (which is
common in older individuals), and the need for rehabilitation
and support services" (p. M141). Another well-accepted
fact is that physical disability adversely affects the quality of
life of participants (Rejeski & Shumaker, 1994; Schulz &
Williamson, 1993) and increases their risk for morbidity and
mortality (Hofman, Grobbee, De Jong, & Van Den Ouweland, 1991; Jette & Branch, 1985).
In recent years, increased attention has been given to the
study and conceptualization of physical disability (Dekker,
Boot, van der Woude, & Bijlsma, 1992; Guralnik et al.,
1989; Pope & Tarlov, 1991), and a significant advance has
been made in the development of multidimensional measures that include psychosocial responses to chronic diseases
(cf. Guccione & Jette, 1990). However, some authors have
argued that little attention is devoted to the direct role that
psychological processes play in activities of daily living
(ADLs; cf. Feinstein, Josephy, & Wells, 1986). One exception to this oversight is represented in the work by Mary
Tinetti and her group, who have demonstrated that selfefficacy for falls is an independent predictor of self-reported
activity restriction, after controlling for history of recent
falls and injury (Tinetti, Mendes de Leon, Doucette, &
Baker, 1994; Tinetti & Powell, 1993).
"Perceived self-efficacy refers to beliefs in one's capacities to mobilize the motivation, cognitive resources, and
P24
course of action needed to meet given situational demands"
(Bandura, 1991, p. 229). There is a growing body of
empirical data demonstrating that self-efficacy is related to
health behaviors as well as health status (cf. Bandura, 1986,
1991; Strecher, DeVellis, Becker, & Rosenstock, 1986). In
addition, Kaplan, Ries, Prewitt, and Eakin (1994) reported
that self-efficacy for walking was a significant predictor for
survival among participants with chronic obstructive pulmonary disease.
In the present study, we elected to focus on the role of
self-efficacy in activity restrictions with individuals who had
knee osteoarthritis (OA). Rather than limiting ourselves to
studying self-reported function (Tinetti et al., 1994), we also
evaluated the ability of self-efficacy to predict performance
during simulated activities of daily living (Rejeski, Ettinger,
Shumaker, James, Burns, & Elam, 1995). A factor that
complicates the seemingly direct relationship between selfefficacy levels, declining physical function, and limited
ADL performance in knee OA is the salience and aversive
nature of pain. Bandura (1991) has argued that self-efficacy
is related to persistence with activities in the face of obstacles. Along these lines, O'Leary, Shoor, Lorig, and Holman
(1988), in studying rheumatoid arthritis patients, demonstrated that enhancing self-efficacy beliefs through cognitive
behavior therapy increased intentions to pursue painful activities. Buescher et al. (1991) reached a similar conclusion
in studying rheumatoid arthritis patients, and there are now
studies in the osteoarthritis literature which demonstrate that
self-efficacy intervention programs can reduce pain and
increase self-reported levels of physical activity (Lorig &
Holman, 1993). In general, these latter studies are consistent
with a larger body of evidence linking control beliefs to both
P25
SELF-EFFICACY AND PAIN IN DISABILITY
a reduction in pain and lower levels of physical disability
(Jensen, Turner & Romano, 1991, 1994; Jensen, Turner,
Romano, & Karoly, 1991).
METHODS
Participants
This investigation involved a series of analyses performed
on baseline data that were collected on 79 participants who
are part of an ongoing clinical trial involving the efficacy of
exercise as a nonpharmacological intervention for treating
OA of the knee — The Fitness and Arthritis in Seniors Trial
(FAST). FAST is a two-center clinical trial being conducted
at Bowman Gray School of Medicine of Wake Forest University (WFU) and the University of Tennessee at Memphis
(UTM). Participants were community-based adults with
knee OA who were recruited through local advertisements
and mass mailings. The sample used in the present study
represents a random subset of participants from the WFU
cohort. The mean (SD) age of this sample was 68.8 years
(6.4 yrs). Eligibility criteria were: (1) radiographic evidence
of knee OA on standing AP X-ray of the knees, (2) selfreport of pain in the knee(s) for most days of the month, (3)
difficulties in activities of ambulation and transfer due to
knee OA, and (4) age >60 years. Radiographs were read by
a radiologist according to the San Francisco grading criteria
(Altman et al., 1987). Demographic data are presented in
Table 1.
Instrumentation
Vo2 peak. — Maximal oxygen consumption was determined during a graded exercise test with a modified
Naughton treadmill protocol. Gas exchange was measured
continuously during the test using a Medical Graphics CPX
system. This system was calibrated according to the manufacturer's instructions. Prior to testing, oxygen consumption
values obtained with this system were compared to those
using a separate set of analyzers and a Tissot tank. There was
no significant difference between the values obtained from
these two methods (r = .98).
Knee strength. — Knee extension muscular strength was
tested through a joint range of motion 90° to 30° (note 0° is
full extension) at an angular velocity of 30 deg.sec 1 using a
Kin Com 125E isokinetic dynamometer. A practice session
was provided using a standardized testing protocol to habituate the patient to the testing environment. Participants'
torsos and the involved thigh were strapped to the testing
chair with the participants' arms folded across their chests.
The input axis of the dynamometer was aligned with the
knee joint. Both the left and right knees were tested with the
least involved knee (fewer symptoms) tested first. Force and
torque output were corrected for gravity based on the weight
of the limb measured at a 45° angle. Two maximal effort
trials that were similar in pattern and magnitude were averaged from a maximum of six trials to yield representative
values. The first and last 10° of motion were deleted from
each torque reading to adjust for any possible acceleration
and deceleration at the beginning or end of the range of
Table 1. Descriptive Statistics on Study Sample
Variable
Percentage
Gender
Male
Female
30.4%
69.6
Race
White
African American
82.3
17.7
Income
< 12th grade
12th grade
> 12th grade
24.1
17.7
58.2
Comorbidities
Heart disease
High blood pressure
Arthritis other areas
Kidney disease
Diabetes
Cancer
Lung disease
Peripheral artery disease
Obesity (BMI>30)
7.6
45.6
72.2
2.5
7.6
6.3
12.7
27.8
50.5
Body Mass Index (%)
Vo2 peak
Knee strength (mean torque)
Mean
(SD)
31.43 (5.20)
17.61 (5.40)
106.05 (39.50)
Stair climb time (sec)
9.07 (3.35)
Lifting/Carrying time (sec)
9.30 (2.16)
Stair climb efficacy
57.11 (26.87)
Lifting/Carrying efficacy
65.14(27.89)
Stair climb pain
Lifting/Carrying pain
2.83 (2.12)
2.33 (1.92)
Stair climb difficulty
2.60 (2.05)
Lifting/Carrying difficulty
2.50 (2.18)
Stair climb physical ability
16.86 (3.14)
Lifting/Carrying physical ability
17.19 (3.20)
motion. The variable analyzed was the average torque between 80° and 40°. This variable provided information
concerning the performance of the knee extensors throughout the range of motion. Data from both legs were averaged
to yield a single strength score.
Stair climb and lift/carry tasks. — The stair climb and
lift/carry tasks are part of a performance test battery that
was developed according to standard psychometric protocol
to assess physical functioning in participants with OA of the
knee (Rejeski et al., 1995). The stair climb involved ascending and descending a set of five stairs that had a
handrail, whereas the lift-and-carry task involved picking
up a 22 kg object from a shelf at knee level, walking a short
distance, and returning the weight to a shelf that was fixed
at each patient's shoulder height. Both tasks have excellent
2-week test-retest reliabilities (>.85). Performances on the
stair climb and lift/carry correlate in expected directions
with Vo2 peak (r = -.37 and -.38, respectively), knee
strength (-.58 in both cases), and self-reported dysfunction
P26
REJESKIETAL.
(r range from .30 to .38); all p values <.01 (see Rejeski et
al., 1995).
Self-efficacy. —Task-specific efficacy beliefs were evaluated following a standardized measurement protocol (Bandura, 1977). Specifically, we presented participants with a
confidence ladder that had 10 steps ranging from 0 (completely uncertain) to 10 (completely certain). For each task,
participants were asked to rate the level of certainty that they
could complete it (a) 2 times, (b) 4 times, (c) 6 times, (d)
8 times, and (e) 10 times without stopping. Self-efficacy
scores for each task were calculated by summing across the
five levels of difficulty for each task and multiplying this
result by 2. Thus, self-efficacy scores ranged from 0 to 100
for each task.
Pain, task difficulty, and perceptions of physical ability.
— Immediately after participants performed both the stair
climb and lift/carry, we evaluated the most intense pain that
was experienced during each task, the perceived difficulty of
each task, and participants' perceptions of their physical
ability. Knee pain was rated on a 10-point ladder with the
phrases "no pain at all" and "pain as bad as it could be"
anchoring the ends of the scale, whereas task difficulty was
rated on a 10-point ladder with "easy" and "extremely
difficult" anchoring the extremes and the word "average"
placed between the fifth and sixth step. Performance-related
perceptions were assessed by asking participants to rate
themselves on the following 5-point, bipolar scales: uncomfortable (l)-comfortable (5), uncoordinated (l)-coordinated
(5), weak (l)-strong (5), and unstable (l)-stable (5). Principal components analysis of the bipolar adjectives for each
task revealed that the five items could be summed to create a
single score for perceptions of physical ability. The items
had very high loadings on each of the principal component
analyses (>.70), with alpha internal consistency reliabilities
of .88 and .91 for the stair climb and lift/carry, respectively.
Procedures
After completing informed consents, participants' physical function was evaluated by having them perform tests that
assessed Vo2 peak, knee strength, and performance on two
tasks that mimic the demands for many ADLs and IADLs:
stair climbing and lifting/carrying. These three assessments
were conducted on different days within a 2-week period.
On the day of the performance testing, subjects completed
pretask self-efficacy scales prior to each task, performed the
task, then completed posttask measures for most intense
pain, task difficulty, and perceptions of physical ability.
The stair climb involved ascending and descending a set
of stairs that had been built for controlled experimental
studies. The stair climb had five steps with a rise of 17.78 cm
and a run of 30.48 cm. There was a 83.82 cm x 121.92 cm
platform at the top. Participants began the task by standing
on a line that was 27.24 cm from the first step with their hand
placed on the handrail. When told to begin, they climbed to
the top of the steps with their left hand on the rail and
immediately turned around and climbed down using the
same handrail. The task was scored as the total time to go up
and down the stairs.
The lift-and-carry task consisted of two movable shelves
affixed to brackets on a wall, a starting line which was 272.4
cm from the wall, and a cone positioned 435.84 cm from the
wall directly behind the starting line. Prior to the start of the
task, the shelves were adjusted so that the lower shelf was
aligned with the center of the patella, whereas the top shelf
was aligned with the acromion process of the scapula. At the
command " G O , " participants walked to the shelves as
quickly as they could, picked up a 22 kg weight that was
located on the bottom shelf with both hands, turned, and
carried the weight around the cone back to the wall, where
they then placed the weight on the top shelf. This was also a
timed task that was terminated when the weight struck the
top shelf.
Statistical Analyses
After normalizing the performance data (i.e., time to
complete stair climb and lifting/carrying tasks) by employing log transformations, we first computed zero-order Pearson-Product Moment correlations between the variables of
interest. Subsequently, stepwise linear regression models
were employed to test for the hierarchical effects of (1) Vo2
peak and strength, (2) most intense pain, and (3) selfefficacy on both performance and self-reported indices of
physical disability (i.e., ratings of task difficulty and perceived ability). Performance scores included time on the
stair climb and lift/carry task, whereas self-reported dysfunction was assessed by ratings of task difficulty and
perceptions of physical ability. In these latter analyses, we
examined two different hierarchical models. After controlling for Vo2 peak and strength, the first model stepped-in
pain and then self-efficacy. In the second model, we once
again controlled for physiologic function and entered pain
after the forced entry of self-efficacy.
RESULTS
Zero Order Correlations
Bivariate correlations between the predictor and outcome
variables are presented in Table 2. A striking feature of these
data is the strength of the relationships between self-efficacy
as well as most intense pain and multiple indices of physical
disability. For example, self-efficacy shared substantial
common variance with performance on the stair climb and
lifting/carrying tasks (35% and 23%, respectively). The
magnitude of these effects was similar for posttask selfreported difficulty rating and perceptions of physical ability.
Additionally, the relationships between self-efficacy and
performance were similar in magnitude to physiological
markers of dysfunction (Vo2 peak and knee strength).
On the other hand, examination of the correlation matrices
among predictor and outcome variables (Table 3) illustrates
that, although self-report and performance-related measures
of disability share common variance, they also possess
substantial amounts of unique variance. For example, task
difficulty ratings following the stair climb shared 40% in
common with timed performance, whereas perceptions of
physical ability had only 23% in common with time to
complete the stair climb. These latter data suggest that some
participants perceive themselves to be very competent in
SELF-EFFICACY AND PAIN IN DISABILITY
P27
Table 2. Zero Order Correlations Between Outcome
Predictors
Outcome Measures
Vo2 peak"
Strength"
Self-Efficacy"
Knee pain*
-.55
-.33
.32
-.51
-.28
.21
-.53
-.49
.35
-.54
-.45
.36
-.59
-.61
.58
-.48
-.41
.50
.42
.58
-.52
.36
.52
-.36
1
Stair climb time (sec)
Stair climb difficulty rating1
Stair climb physical ability"
Lift/Carry time (sec)1
Lift/Carry difficulty rating1
Lift/Carry physical ability6
•For measures with this superscript, higher values indicate greater dysfunction.
"For measures with this superscript, lower values indicate greater dysfunction.
All p-values <.01, except Vo2 peak with Lift/Carry Physical Ability scores wherep < .05.
Table 3. Correlation Matrices for Outcome and Predictor Variables
Stair climb time
Stair climb difficulty
Stair climb physical ability
Lift/Carry time
Lift/Carry difficulty
Lift/Carry physical ability
Vo2 peak
Knee strength
Self-efficacy
Pain
Stair Climb
Time
Stair Climb
Difficulty
1
-.63
1
Stair Climb
Physical
Ability
-.48
-.77
1
Lift/Carry
Time
Lift/Carry
Difficulty
.80
.51
-.41
1
.40
.40
-.44
.51
1
Lift/Carry
Physical
Ability
-.34
-.51
.63
-.35
-.56
1
Vo2 peak
Knee Strength
Self-Efficacy
Pain
1
.61
1
.54
.40
1
-.18
-.27
-.35
1
All p-values < . 0 1 , with the exception of pain with Vo2 peak, which was nonsignificant.
stair climbing, an attribute that is not captured by their speed
of movement. Perhaps perceptions of physical ability correlate more highly with outcomes such as gait, which permit
one to evaluate "quality of movement."
Hierarchical Modeling With Multiple Regression
Since Bandura (1977, 1986) suggests that self-efficacy
beliefs are due, in part, to physical capabilities and symptoms experienced during performance, our next strategy was
to conduct hierarchical regression models for each outcome
variable. These analyses provided a conservative test of the
relative importance of self-efficacy beliefs on disability; that
is, how much of the variance in self-reported and performance-related disability can be explained by self-efficacy
beliefs over and above physiologic function (i.e., Vo2 peak
and knee strength) and most intense knee pain experienced
during performance of the task. Subsequent to testing the
main effect of self-efficacy, we also entered interaction
terms to ascertain whether efficacy beliefs interacted with
Vo2 peak or knee strength in predicting the variables of
interest.
Panel I of Table 4 presents the incremental change in R2
for the outcome measures in which self-efficacy was entered
after the sequential entry of the physiologic measures and
most intense pain. Inspection of these data reveals that selfefficacy explained a significant portion of the variance in
both performance and the self-report of physical disability.
The added percentage of explained variance ranged from a
low of 3 % to a high of 12.5 %; p-values ranged from < . 05 to
<.001). None of the interactions between self-efficacy and
the physiologic variables added in any meaningful way to the
explanation of either task performance or subjective assessments following performance of the tasks.
These analyses underscore the fact that most intense knee
pain during activity is important in understanding speed of
movement (8% and 5% of the variance in stair climb and the
lift/carry tasks, respectively) as well as posttask ratings of
task difficulty and perceptions of physical ability (all pvalues <.001). In fact, when entered in step 2, knee pain
accounts for from 22% to 43% of the variance in these
subjective evaluations after controlling for Vo2 peak and
knee strength. Although conceptually one might argue that
the influence of knee pain on performance is mediated
through pretask self-efficacy (Bandura, 1977), pain may
have an independent, direct effect on disability. To evaluate
this latter proposition, we conducted hierarchical models
where most intense pain was entered after Vo2 peak, knee
strength, and pretask self-efficacy. Addition of knee pain
accounted for substantial increases in variance explained for
models to predict each outcome measure. Increases in R1
ranged from 3% to 12%, p-values ranged in significance
from <.05 to <.001 (see panel II of Table 4). Parenthetically, in each of the hierarchical models discussed previously, we also examined whether it was meaningful to
REJESKIETAL.
P28
Table 4. Results of Hierarchical Regression Analyses
PANEL I: Physiological Variables, Pain, Efficacy
R2 in %
•
A m%
(% Change with Each Step)
Step 1
Vo2 Peak &
Step 2
Step 3
Outcome Variables
Knee Strength
Pain
Self-Efficacy
Stair Climb
Time
Task difficulty
Physical ability
37***
24***
14***
45 (8)**
46 (22)***
33 (20)***
52 (7)**
58(12)***
46(12)***
Lift/Carry
Time
Task difficulty
Physical ability
35***
20***
13***
40 (5)**
63(43)***
41 (28)***
43 (3)**
69 (6)***
48 (7)***
PANEL II: Physiological Variables, Efficacy, Pain
R2 in %
(% Change with Each Step)
Outcome Variables
R2 \n%
Step 1
Vo2 Peak &
Knee Strength
Stair Climb
Time
Task difficulty
Physical ability
37***
24***
14***
48(11)***
47(13)***
36 (22)***
52 (4)*
58(11)***
46(10)***
Lift/Carry
Time
Task difficulty
Physical ability
35***
20***
13***
40 (5)**
35(15)***
27(14)***
43 (3)*
69 (34)***
48(21)***
Step 2
Self-Efficacy
Step 3
Pain
*p < .05; **p < .01; ***p < .001.
include race and gender as covariates. To our surprise, none
of the resultant solutions were influenced in any significant
way by the addition of either variable.
DISCUSSION
The results of this study reinforce the position that although performance-related and self-reported function share
common variance, one method of assessment cannot be used
as a surrogate for the other (Myers, Holliday, Harvey, &
Hutchinson, 1993; Rejeski et al., 1995). Moreover, pretask self-efficacy beliefs and knee pain experienced during the performance of physical tasks influence speed of
movement, posttask difficulty ratings, and perceptions of
physical ability. These data extend both the research by
Tinetti and her colleagues (1993; 1994), who found that selfefficacy for falls was an independent predictor of selfreported activity restriction, and research in the osteoarthritis literature which has shown that knee pain is related to
disability (Davis, Ettinger, Neuhaus, & Mallon, 1992;
McAlindon, Cooper, Kirwan, & Dieppe, 1992). Moreover,
the present data corroborate the important role of control
beliefs in disability that is exacerbated by clinical pain
(Jensen et al., 1991) and reinforce the wisdom of including
self-efficacy beliefs in arthritis management training programs (see Lorig & Holman, 1993).
As suggested by the bivariate correlations and the hierarchical regression analyses, self-efficacy beliefs covary, in
part, with the pain that participants experience during the
performance of activities of daily living such as stair climb-
ing and lifting/carrying. This finding is not surprising. For
example, McAuley and Courneya (1992), studying older
adults, found that individuals with higher preexisting selfefficacy toward cycle ergometry work gave lower ratings of
exertion and felt better during exercise than those with low
self-efficacy. Also, Bozoian, Rejeski, and McAuley (1994)
have shown that women with higher exercise self-efficacy
report large increases in revitalization and positive engagement following acute bouts of work in contrast to their low
efficacy counterparts. Thus, while pain undoubtedly contributes to the formation of self-efficacy beliefs, participants
who have low efficacy beliefs place themselves at risk for
negative experiences both during and following involvement
in physical activity (Bandura, 1991). Interestingly Taylor,
Bandura, Ewart, Miller, and DeBusk (1985), investigating
male heart attack victims, found that improvement in cardiovascular function across a 6-month time interval was greatest
when participants and their spouses had high efficacy beliefs
toward participants' cardiac capacities. Initial treadmill performance did not predict level of function at the time of
follow-up testing when efficacy beliefs were partialled out of
the analyses; however, self-efficacy did predict improvement in function over and above initial levels of treadmill
performance. O'Leary et al. (1988) conducted an intervention with rheumatoid arthritis participants and found that
cognitive behavioral therapy increased efficacy beliefs such
that participants reported less pain and were more inclined to
pursue potentially painful activities. Similarly, Lorig and
Holman (1993), in a series of investigations, found that
patients with osteoarthritis who participated in a selfefficacy oriented program for arthritis management reported
less pain and indicated that they had increased time spent
walking.
What is interesting from the second set of regression
analyses is that pain appears to have a direct relationship to
physical disability, independent of its role through selfefficacy. This finding is consistent with conceptual work by
Rotter (1954) and others (e.g., Bandura, 1986; Maddux &
Stanley, 1986), who argue that outcome value is an important determinant of motivation. In other words, prior to
engaging in physical activity, participants may have a reasonable level of confidence in what they can achieve. However, knee pain experienced during behavior is an aversive
stimulus that can sabotage even the most optimistic point of
view. If the experience and perception of pain modify the
value of performing ADLs, the incentive to complete those
activities is diminished although an individual's confidence
(self-efficacy) remains relatively high for carrying out
ADLs. Bandura (1986) stresses that without adequate incentives self-efficacy beliefs are not sufficiently motivating for
behavior to occur. Thus, the examination of outcome value
of ADLs performed with pain requires study.
It would also be interesting to examine how participants
react to intermittent episodes of acute discomfort. For example, it is reasonable to hypothesize that some participants
make optimal use of days that are relatively free of pain/
discomfort by increasing their level of activity, whereas
others get caught in a cycle of chronic inactivity. It would be
important for both researchers and clinicians to prospectively observe the reciprocal relationships between the value
SELF-EFFICACY AND PAIN IN DISABILITY
of performing ADLs for knee OA patients in painful versus
pain-free conditions and their self-efficacy beliefs in coping
with performance under these conditions. Further, Lorig,
Chastain, Ung, Shoor, & Holman (1989) have developed an
arthritis self-efficacy measure that has subscales for control
over (1) pain, (2) function, and (3) symptoms such as fatigue
and negative affect. Future research should explore the
relationships between, and independence of, these subscales
to performance-related disability. Intervention programs
may well benefit from focusing on multiple dimensions of
the self-efficacy construct. Clearly, learning how to cope
with pain and to manage other symptoms is crucial to any
program that is designed either to prevent or counter the rate
of decline found with a degenerative disease such as knee
OA (Keefe et al., 1990a, 1990b; Rejeski & Shumaker,
1994).
ACKNOWLEDGMENTS
This project was supported by National Institute on Aging Grant AG
10484-01.
Address correspondence and requests for reprints to Dr. W. Jack Rejeski, Department of Health and Sport Science, Wake Forest University,
Reynolda Campus, Box 7234, Winston-Salem, NC 27109.
REFERENCES
Altman, R. D., Fries, J. F., Bloch, D. A., Carstenf, J., Cooke, T. D.,
Genant, H., Gofton, P., Groth, H., McShane, D. J., Murphy, W. A.,
Sharp, J. T., Spitz, P., Williams, C. A., & Wolfe, F. (1987). Radiographic assessment of progression in osteoarthritis. Arthritis and Rheumatism, 30, 1214-1225.
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavior
change. Psychological Review, 84, 191-215.
Bandura, A. (1986). Social foundations of thought and action: A social
cognitive theory. Englewood Cliffs, NJ: Prentice-Hall.
Bandura, A. (1991). Self-efficacy mechanism in physiological activation
and health-promoting behavior. In J. Madden (ed.), Neurobiology of
learning, emotion, and affect (pp. 229-269). New York: Raven Press.
Bozonian, S., Rejeski, W. J., & McAuley, E. (1994). Self-efficacy
influences feeling states associated with acute exercise. Journal of Sport
and Exercise Psychology, 16, 326-333.
Buescher, K. L., Johnston, J. A., Parker, J. C , Smarr, K. L., Buckelew,
S. P., Anderson, S. K., & Walker, S. E. (1991). Relationship of selfefficacy to pain behavior. The Journal of Rheumatology, 18, 968-972.
Davis, M. A., Ettinger, W. H., Neuhaus, J. M., & Mallon, K. P. (1992).
Knee osteoarthritis and physical functioning: Evidence from the
NHANES I epidemiologic follow-up study. Journal of Rheumatology,
18, 591-598.
Dekker, J., Boot, B., vanderWoude, L. H. V., &Bijlsma, J. W. J. (1992).
Pain and disability in osteoarthritis: A review of biobehavioral mechanisms. Journal of Behavioral Medicine, 15, 189-214.
Feinstein, A. F., Josephy, B. R., & Wells, C. K. (1986). Scientific and
clinical problems in indexes of functional disability. Annals of Internal
Medicine, 705,413-420.
Guccione, A. A., & Jette, A. M. (1990). Multidimensional assessment of
functional limitations in participants with arthritis. Arthritis Care Research, 3, 44-52.
Guralnik, J. M., Branch, L. G., Cummings, S. R., & Curb, J. D. (1989).
Physical performance measures in aging research. Journal of Gerontology: Medical Sciences, 44, M141-M146.
Hofman, A., Grobbee, D. E., De Jong, P. T. V. M., & Van Den
Ouweland, F. A. (1991). Determinants of disease and disability in the
elderly. European Journal of Epidemiology, 7, 403-422.
Jensen, M. P., Turner, J. A., & Romano, J. M. (1991). Self-efficacy and
outcome expectancies: Relationship to chronic pain, coping strategies
and adjustment. Pain, 44, 263-269.
Jensen, M. P., Turner, J. A., & Romano, J. M. (1994). Correlates of
improvement in multidisciplinary treatment of chronic pain. Journal of
Consulting and Clinical Psychology, 62, 172-179.
P29
Jensen, M. P.,Turner, J. A., Romano, J. M., & Karoly, P. (1991). Coping
with chronic pain: A critical review of the literature. Pain, 47, 249283.
Jette, A. M., & Branch, L. G. (1985). Impairment and disability in the
aged. Journal of Chronic Diseases, 38, 59-64.
Kaplan, R. M., Ries, A. L., Prewitt, L. M., & Eakin, E. (1994). Selfefficacy expectations predict survival for participants with chronic
obstructive pulmonary disease. Health Psychology, 13, 366-368.
Keefe, F. J., Caldwell, D. S., Williams,"D. A., Gil, K. M., Mitchell, D.,
Robertson, C , Martinez, S., Nunley, J., Beckman, J. C , Crisson, J.
E., & Helms, M. (1990a). Pain coping skills training in the management of osteoarthritic knee pain: A comparative study. Behavior Therapy, 21, 49-62.
Keefe, F. J., Caldwell, D. S., Williams, D. A., Gil, K. M., Mitchell, D.,
Robertson, C , Martinez, S., Nunley, J., Beckman, J. C , & Helms, M.
(1990b). Pain coping skills training in the management of osteoarthritic
knee pain-II: Follow-up results. Behavior Therapy, 21, 435-447.
Lorig, K., Chastain, R. L., Ung, E., Shoor, S., & Holman, H. R. (1989).
Development and evaluation of a scale to measure perceived selfefficacy in people with arthritis. Arthritis and Rheumatism, 32, 37-44.
Lorig, K., & Holman, H. (1993). Arthritis self-management studies: A
twelve-year review. Health Education Quarterly, 20, 17-28.
Maddux, J. E., & Stanley, M. A. (1986). Self-efficacy theory in contemporary psychology: An overview. Journal of Social and Clinical Psychology, 4, 249-255.
McAuley, E., & Coumeya, K. S. (1992). Self-efficacy relationships with
affective and exertion responses to exercise. Journal of Applied Social
Psychology, 22, 312-326.
McAlindon, T. E., Cooper, C , Kirwan, J. R., & Dieppe, P. A. (1992).
Knee pain and disability in the community. British Journal of Rheumatology, 31, 189-192
Minaire, P. (1992). Disease, illness and health: Theoretical models of the
disablement process. Bulletin of the World Health Organization, 70,
373-379.
Myers, A. M., Holliday, P. J., Harvey, K. A., & Hutchinson, K. S.
(1993). Functional performance measures: Are they superior to selfassessments? Medical Science, 48, M196-M206.
Nagi, S. Z. (1965). Some conceptual issues in disability and rehabilitation.
In M. B. Sussman (Ed.), Sociology and rehabilitation. Washington,
DC: American Sociological Association.
O'Leary, A., Shoor, S., Lorig, K., Holman, H. R. (1988). A cognitivebehavioral treatment for rheumatoid arthritis. Health Psychology, 7,
527-544.
Pope, A. M., & Tarlov, A. R. (1991). Disability in America: Toward a
national agenda for prevention. Washington DC: National Academy
Press.
Rejeski, W. J., Ettinger, W. H., Shumaker, S., James, P., Burns, R., &
Elam, J. T. (1995). Assessing performance-related disability in participants with knee osteoarthritis. Osteoarthritis and Cartilage, 3, 157-167.
Rejeski, W. J., & Shumaker, S. (1994). Knee osteoarthritis and healthrelated quality of life. Medicine and Science in Sports and Exercise, 26,
1441-1445.
Rotter, J. B. Social learning theory and clinical psychology. Englewood
Cliffs, NJ: Prentice-Hall, 1954.
Schulz, R., & Williamson, G. M. (1993). Psychosocial and behavioral
dimensions of physical frailty. Journal of Gerontology: Psychological
Sciences, 48, P39-P43.
Strecher, V. J., DeVellis, B. E., Becker, M. H., & Rosenstock, I. M.
(1986). The role of self-efficacy in achieving health behavior change.
Health Education Quarterly, 13, 73-91.
Taylor, C. B., Bandura, A., Ewart, C. K., Miller, N. H., & DeBusk, R. F.
(1985). Exercise testing to enhance wives' confidence in their husbands'
cardiac capabilities soon after clinically uncomplicated acute myocardial infarction. American Journal of Cardiology, 55, 635-638.
Tinetti, M. E., & Powell, L. (1993). Fear of fall ing and low self-efficacy: A
cause of dependence in elderly persons. Journal of Gerontology: Medical Sciences, 48, M35-M38.
Tinetti, M. E., Mendes de Leon, C. F., Doucette, J. T., & Baker, D. I.
(1994). Fear of falling and fall-related efficacy in relationship to
functioning among community-living elders. Journal of Gerontology:
Medical Sciences, 49, M140-M147.
Received March 8, 1995
Accepted August 2, 1995