• Exercise prescriptions;
  • Patient-provider communication;
  • Arthritis


  1. Top of page
  2. Abstract


To describe how patients and their rheumatologists discuss exercise, and to identify predictors of exercise prescriptions.


Twenty-five rheumatologists and 132 patients with rheumatoid arthritis completed questionnaires and were audiotaped during a subsequent clinic visit. Chi-square and t-tests assessed associations between variables. Principal components analysis identified patterns of talk about exercise. Multivariate logistic regression identified predictors of an exercise prescription.


Seventy of the 132 patients (53%) discussed exercise. Of these, 18 (26%) received an exercise prescription. Principal components analysis identified 3 patterns of talk about exercise. Aerobic exercise discussions contained more information about drawbacks, side effects, pain, and bargaining than did discussions about general exercises, and referral to physical therapy for exercise. Significant predictors of a prescription included rheumatologist-initiated discussion about exercise (odds ratio [OR] 4.6; P = 0.03); talk about exercise in improving function, exercise instructions, opinions about the usefulness of exercise (OR 3.1; P = 0.01); and discussions about non-exercise treatments (OR 1.6; P = 0.01).


Exercise and referral to physical therapy for exercise are discussed differently and are 4 times more likely to occur when the rheumatologist initiates the discussion. These discussions strongly impact on the likelihood a patient receives an exercise prescription.


  1. Top of page
  2. Abstract

By the early 1960s the role of exercise in maintaining and promoting health in the general population was well established (1, 2). In the last 40 years, research has shown that its benefits extend to patients with a variety of chronic diseases. While range-of-motion exercises and physical therapy have long been essential components of arthritis management (3), aerobic and strengthening exercises are still a relatively new addition to arthritis interventions (4). However, by the 1990s there was a significant research base indicating that patients with arthritis can participate in regular, moderate intensity exercise, enabling improvements in cardiac fitness, endurance, muscle strength, flexibility, and function (3, 5–9).

Dissemination of research findings into practice often lag considerably. The purposes of the current study were to describe how patients and rheumatologists discuss exercise during clinical visits and to identify predictors of an exercise prescription. In a previous study (10) we illustrated the impact of patients' and rheumatologists' attitudes and beliefs about exercise and exercise behaviors on the likelihood an exercise discussion occurred during the clinical visit. These findings concurred with earlier research that highlights the importance of attitudes in predicting intention to exercise and exercise behavior (1, 11).

We used the Theory of Planned Behavior (12), along with its precursor the Theory of Reasoned Action (13), and Janis and Mann's Decision-Making Theory (14) as the framework for this study. The Theory of Planned Behavior is based on a body of well-established psychological and behavioral literature that describes the relationship among attitudes, beliefs, intentions, and behaviors, and mediating factors that influence behavior (e.g., discussing exercise in the clinical encounter). Rheumatologist-patient discussions about exercise are influenced by many factors, including beliefs about decision-making roles, social roles, social norms regarding the appropriateness of an exercise discussion during the clinical encounter, beliefs about and attitudes toward the efficacy of exercise in managing rheumatoid arthritis, and individual preferences for treatments (10). Attitudes are strongly influenced by exposure to others' opinions and beliefs, especially when messages are personally relevant and delivered by experts (doctors in the case of the clinical encounter), either through reinforcing the use of exercise or by communicating misgivings (15–18). Therefore, rheumatologists' attitudes and beliefs are not only likely to affect the manner in which exercise is presented to the patient, but also are likely to affect the dynamics of the conversation.

However, how these attitudes and beliefs interact during the clinical encounter and their impact on the quality and context of exercise discussions is unclear.

The Janis and Mann Decision Theory (14) explains how people evaluate information and make decisions in periods of stress (11, 19, 20). A person who thoroughly reviews all the available alternatives, assesses the risks and benefits of each alternative, and develops a plan to cope with potential drawbacks tends to be more committed to the treatment. This skill is particularly important when deciding to use exercise to manage arthritis symptoms. Unlike other conditions, arthritis symptoms may change from day to day; therefore, understanding when exercises are appropriate allows patients to modify their treatment regimen accordingly.

We hypothesized that exercise prescription rates would be directly influenced by the qualities of exercise discussions and by the attitudes and beliefs rheumatologists and patients brought to the discussion. Figure 1 depicts the hypothesized relationship among study variables. In summary, the attitudes and beliefs of patients and rheumatologists were hypothesized to influence their communication about exercise, which in turn influenced the likelihood of an exercise prescription. It was hypothesized that patients would be less likely to receive an exercise prescription if negative attitudes regarding exercise were shared. If positive attitudes and beliefs were shared, a prescription would likely occur.

thumbnail image

Figure 1. Hypothesized relationship among study variables leading to an exercise prescription. PT = physical therapy; Rx = prescription. Adapted from Iversen MD, Fossel AH, Daltroy LH. Rheumatologist-patient communication about exercise and physical therapy in the management of rheumatoid arthritis. Arthritis Care Res 1999;12:180–92. With permission.

Download figure to PowerPoint


  1. Top of page
  2. Abstract


This cross-sectional study described characteristics of conversations between patients and rheumatologists about exercise prescription. Exercise prescription is operationally defined as an exercise prescription by the rheumatologist for general strengthening, range of motion, and aerobic exercise or referral to physical therapy for exercise. The study was approved by the Institutional Review Board of the participating hospital and was embedded in an NIAMS MAMDC grant (A Randomized Controlled Trial to Evaluate an Intervention to Improve Doctor-Patient Decision-Making in Arthritis, Lawren H. Daltroy, DrPH, Principal Investigator, April 1992 to March 1997 [AR-36308 Matthew H. Lian, MD, Center Principal Investigator]) designed to test an intervention to enhance rheumatologist-patient decision making about major treatment changes to manage active rheumatoid arthritis.


Details of the main study are described in a previous publication (10). Briefly, rheumatologists and patients were recruited from the arthritis clinic of a major teaching hospital between June 1994 and February 1996 (10). Forty-two rheumatologists were invited to participate, of whom 25 (59%) provided informed consent and completed baseline questionnaires. Twenty-eight percent of the rheumatologists were female and more than half were considered senior faculty (> 5 years experience). The mean (±SD) age was 43 ± 10.4 years, and 42% reported they exercised regularly. Thirteen of 25 rheumatologists (52%) were primarily bench researchers who had clinical responsibilities in addition to their research.

The rheumatologists identified 338 patients with active disease, who were also approaching a decision milestone (e.g., considering a new medication or surgery). Two hundred ninety-three patients (87%) met the additional eligibility criteria (age > 17 years; met the American College of Rheumatology criteria for rheumatoid arthritis [21]; and were able to complete questionnaires and participate in telephone interviews in English) and were contacted. Of these, 135 (46%) refused and 158 patients (54%) agreed to participate. One hundred thirty-two patients (83%) provided informed consent, completed the baseline questionnaires, and were audiotaped during their clinical encounter. The patients were mostly white (94%) and female (79%), and had a mean (±SD) age of 54 ± 15.4 years. Approximately two-thirds (67%) had been diagnosed with RA for >5 years. This study used the baseline data together with the audiotape data from the first clinical encounters to identify predictors of an exercise prescription.

Data collection and procedure.

The full set of baseline questionnaires included demographic information such as age, sex, race, marital and employment status, educational attainment, income, and health insurance data, and information on current and past medical and non-medical treatments (10). Selected subscales of the Short-Form 36 (22, 23) were used to assess pain, physical function, and mental health. These subscales have internal reliability coefficients of 0.78–0.93 (22, 23). Self-efficacy for arthritis management, an important factor influencing perception of general health, was measured using Lorig's Self Efficacy Other scale (24), which has an internal reliability alpha of 0.91. We assessed expectations about the benefits and risks of exercise at baseline and immediately following the clinic visit using an expectancy-value questionnaire, in which patients indicated the likelihood of each risk and benefit occurring using a 7-point Likert Scale (extremely unlikely to extremely likely). We assessed perceived social support for exercise, attitudes toward regular home exercise and weekly physical therapy, and beliefs about exercise effectiveness in managing rheumatoid arthritis symptoms using 7-point Likert scales (ranging from unfavorable to favorable) based on work by Ajzen and Fishbein (12, 13). Past experiences with exercise, history of adherence to exercise, and current exercise behaviors were also assessed. The patients and rheumatologists judged the patient's functional status and disease severity immediately following the clinical encounter using a short questionnaire and reported whether an exercise discussion occurred (Table 1).

Table 1. Study measures
Patient demographicsAge, sex, race, work status, insurance, and marital status.
Doctor demographicsAge, sex, race, years experience (at baseline).
Medical historyNumber of comorbidities, past medications, duration of illness, collected via chart review, and by patient and doctor report at baseline.
Decision-making about treatmentsBeisecker and Beisecker 13-item Locus of Authority Scale (25) adapted. Internal reliability alpha of 0.73 (at baseline).
Patient expectations of exercise/physical therapyA 7-point scale (extremely unlikely to extremely likely) asking patients to rate the likelihood of attaining specific vigorous and moderate functional activities from their arthritis treatments (at baseline).
Self-efficacy for managing rheumatoid arthritisLorig's summated 9-item questionnaire (24) using a 7-point Likert scale (very uncertain to very certain). Internal reliability alpha of 0.91 (at baseline and immediately following the visit).
General healthSelected subscales of the Short Form-36 (22, 23) with internal reliability alpha of 0.78–0.90 (at start of treatment).
Exercise attitudes, beliefs and perceived social norms7-point Likert scale based on Ajzen and Fishbein's model (17, 18) with an internal consistency alpha of 0.60 (10) (at baseline).
Health at visitRanking patients' disease severity & activity using an 11-point scale. Three questions assessed function with a response set ranging from completely able to severely limited. The internal reliability alpha was 0.83 (10) (immediately following the clinic visit).
Report of exercise discussionSingle item question assessed occurrence of prescription (immediately following the clinic visit).
Non-exercise treatmentsNumber of non-exercise treatments discussed in clinical encounter recorded from transcripts and summed to create a single variable.
Exercise discussionContent and amount of exercise discussion as noted on transcripts.

Rheumatologists' demographic information included sex, age, and years and type of experience (lab versus clinical). Rheumatologists' attitudes and beliefs about home exercise and referral to physical therapy for exercise were assessed using 12-item scale with 5-point Likert response sets (strongly agree to strongly disagree). The questionnaire assessed self-efficacy about ability to teach exercises; beliefs about patient adherence to exercise prescriptions; accessibility of physical therapy; and beliefs about exercise instruction. Beliefs about instruction included judgments as to the time required to teach exercise and beliefs about their relative ability to instruct patients in exercise, compared with physical therapists. The attitude scale had an acceptable but low internal reliability alpha of 0.60. Personal experience with exercise was measured with 2 questions (Table 1). We adapted Beisecker and Beisecker's 13-Item Locus of Authority Scale (25) to measure patients' and rheumatologists' decision-making styles. The scale had an internal reliability alpha of 0.73.

An exercise discussion was considered to have occurred if the visit included talk about flexibility exercises, muscle strengthening (isometric, isotonic, weight training), range of motion exercises, and aerobic exercises (swimming, fast walking, low impact aerobics) or physical therapy for exercise instruction. We defined an exercise prescription as an oral or written instruction that the patient begin a regular regimen of muscle strengthening, range of motion, or aerobic exercise or referral to physical therapy for exercise instruction. General exercises were defined as flexibility exercises, muscle strengthening, and range of motion exercises. Aerobic exercises included swimming, treadmill, fast walking, low impact aerobics, and active sports.

Each patient's first audiotaped visit was analyzed for exercise discussion. In the few instances where technical problems arose, a subsequent taped visit was used. Audiotapes were transcribed and edited twice for completeness and accuracy. Transcripts were then coded for exercise content (Table 2), by coders who were unaware of the patients' and rheumatologists' identity. We coded the content of exercise discussions by the rheumatologist and patient in each transcript for the following: conversation initiator; new or continuing exercise/physical therapy; intervention plan, drawbacks/side effects, pain relief, function, symptom relief, opinions, bargaining, patient efficacy, adherence to exercise, instruction, and statements about past experience with exercise or physical therapy. Each conversational utterance, or smallest phrase to which meaning could be ascribed (26), was categorized using the coding scheme. A 5-point weighting scheme indicated the intent of the conversation (e.g., persuade or discourage use of exercise). If a statement was deemed neutral then a value of 0 was assigned. The interrater reliability for this coding scheme was acceptable (ICC 0.80) (10). We also coded discussion about non-exercise treatments such as medications and home modifications.

Table 2. Transcript coding scheme of conversation during clinical encounters*
Coding categoryDescription
  • *

    Statements weighted based on the perceived intention of the speaker to persuade (1, 2) or dissuade (−1,−2) the person from adopting the behavior or describing the negative or positive attributes of the treatment. Adapted from Iversen MD, Fossel AH, Daltroy LH. Rheumatologist-patient communication about exercise and physical therapy in the management of rheumatoid arthritis. Arthritis Care Res 1999;12:180–92. With permission. ROM = range of motion.

Side effectsMedical symptoms or outcomes related to exercise
DrawbacksCost or inconvenience of physical therapy or exercise
Pain reliefSelf-explanatory
Improvement of symptomsSelf-explanatory
OpinionUnsubstantiated statement
BargainsSuggestions made by patient or doctor to enhance acceptability of exercise programs (i.e., patient agrees to start a walking program but will not bike)
Patient efficacyConfidence in ability to do therapeutic exercises or follow through with physical therapy recommendations
Functional improvementCan do more, relates to independence
Treatment characteristicsRelative advantage of aspects of the exercise prescription (i.e., immediate relief of stiffness with ROM exercises)
Patient adherenceDiscussion of success/failure with exercise
Timing of exerciseWhen and how long to do exercises
Exercise instructionRanked on a 4-point scale indicating the detail of the explanation provided (e.g., 1 = oral instruction, 2 = written, 3 = demonstration, 4 = referral for exercise instruction)
Evaluation of exercise effectivenessSelf-explanatory
Past experience with physical therapy or exercise programsSelf-explanatory


We evaluated the relationships between baseline predictor variables using parametric and non-parametric tests, as appropriate. McNemar's chi-square tests and paired t-tests were used for paired comparisons. We used principal components analysis to develop profiles of exercise discussions (27). The binary outcome, exercise prescription (yes/no), was modeled using logistic regression and the log of the conversation profile score to achieve a more normal distribution. Variables selected as potential predictors included conversation initiator, profiles of talk, number of non-exercise treatments discussed at the clinic visit, and the patients' health status at the day of the visit. One goal was to identify immediate predictors of exercise prescriptions during the clinical encounter once an exercise discussion took place. Our previous study (10) identified demographic characteristics such as patients' past experience with exercise and physical therapy and rheumatologists' and patients' current exercise behavior, as significant predictors of an exercise discussion in the clinical encounter. Therefore, variables considered in this modeling included only variables from the clinic visit such as health status at visit, conversation profiles, initiator of discussion, and number of non-exercise treatments discussed (home modifications and medications). All analyses were performed using the SAS statistical package (28).


  1. Top of page
  2. Abstract

Baseline features.

Seventy patients (53%) discussed exercise with their rheumatologist. Patients who discussed exercise were similar on all baseline characteristics to those who did not discuss exercise (Table 3). Of these, 18 patients (26%) received an exercise prescription. Men and women were equally likely to receive a therapeutic exercise prescription (relative risk 1.04, P = 0.5). Nine prescriptions (50%) were referrals to physical therapy for instruction in exercise, and 9 (50%) were rheumatologist instructions for exercise. Among those who received a prescription, 3 (17%) believed the decision to try exercises belonged to the patient, while 12 (66 %) felt the patient should decide when to be referred to physical therapy for exercise. Of the patients who did not receive an exercise prescription, 6 (12%) reported it was the patients' decision to try exercises, while 8 (15%) believed the patient should decide when to receive a physical therapy referral.

Table 3. Characteristics of patients with rheumatoid arthritis who attended the clinic visit (n = 132)*
Patient characteristicsExercise discussion (n = 70)No exercise discussion (n = 62)P
  • *

    Values are the mean ± SD unless indicated otherwise. Wilcoxon or t-test used unless otherwise indicated. RA = rheumatoid arthritis.

  • χ2 test used

Age, years54 ± 16.454 ± 14.10.9
Caucasian, number (%)68 (97)55 (89)0.07
Female, number (%)54 (77)55 (82)0.1
Diagnosis of RA > 5 years, number (%)27 (38)17 (27)0.2
SF-36 Function Subscale49 ± 2852 ± 260.07
SF-36 Mental Health Subscale70 ± 2271 ± 170.1
SF-36 Pain47 ± 2344 ± 230.08
Expectations of function from treatments4.6 ± 14.5 ± 11.0
Self-efficacy for managing RA (0–10)4.5 ± 14.5 (1)1.0
Number of medications2.4 ± 0.92.4 ± 0.80.1

On average, there was no difference in disease activity, severity, function, past or current experience with physical therapy or exercise between patients who received a referral and those who did not receive a referral. There were marginal but insignificant differences by patient age and social approval for exercise between those who received a prescription and patients who did not (mean 48 years versus 56 years, P = 0.08; mean 6.5 versus 6.1; P = 0.06, respectively) (Table 4).

Table 4. Characteristics of patients who received an exercise prescription compared with those who did not among those who discussed exercise (n = 70)
 Prescription (n = 18)Non-prescription (n = 52)P
  • * Values are mean ± SD unless otherwise indicated. P values based on t-tests.

  • Determined by Wilcoxon ranked sum test.

 Age48 ± 1656 ± 160.08
 Mental status74 ± 1770 ± 230.6
 Physical function50 ± 2748 ± 290.8
 Pain50 ± 1945 ± 250.3
 Self-efficacy4.6 ± 14.5 ± 1.50.8
 Believe exercise useful5.2 ± 25.6 ± 20.7
 Perceive positive social norms for exercise6.5 ± 0.46.4 ± 10.8
 Believe physical therapy is useful5.6 ± 25.4 ± 20.8
 Expectations of function from treatments4.9 ± 1.34.5 ± 1.50.2
 Perceive positive social norms for physical therapy6.5 ± 16.1 ± 10.6
Health status at clinic visit   
 Disease activity5.1 ± 35.6 ± 30.6
 Disease severity5.5 ± 26.3 ± 20.2
 Physical function1.7 ± 0.51.8 ± 0.60.8

Conversation profiles.

We were interested in describing how exercise was discussed in clinical encounters. We treated rheumatologist and patient discussion of the same topic as a unit because we felt the interchanges would be too interdependent to treat separately. This was confirmed by preliminary analysis. We subjected the items in Table 1 to a principal components analysis with orthogonal rotations to identify patterns or profiles of exercise discussions. Three main factors merged (Table 5). The asterisk in Table 5 represents the significant factor loadings that are the correlation coefficients between the study variables (listed in the rows) and the retained variables (listed in the columns). The retained variables retained were then summed to create a factor-derived variable for each profile. These profiles were identified statistically, but were accepted as meaningful because they were characteristic of certain clinical discussions. Profiles, defined as patterns of conversation about exercise, were then used to describe rheumatologist-patient exercise conversations. These profiles could occur more than once in a clinical discussion.

Table 5. Principal components analysis of exercise conversations between rheumatologists and patients with rheumatoid arthritis
 Rotated factor pattern
Profile 1 “Pros and Cons”Profile 2 “How To”Profile 3 “Adherence”
  • *

    Represents the significant factor loadings that are correlation coefficients between study variables and retained variables.

Doctor-patient communication   
Side effects and drawbacks84*38
Pain and symptom relief82*136

Profile 1, “the Pros and Cons profile,” is characterized by conversation pertaining to side effects and drawbacks, pain and symptom relief, and bargaining. We defined bargaining as conversation about altering the frequency, duration, or type of exercise. An example of bargaining is:

Physician (MD): “… and your exercises?”

Patient (PT): “Yes, well when I do them, they are so boring and …”?

MD: “They're boring?”

PT: “Yes, and I'm doing things in the morning, then I have to go out. When I get home, well they just take too much time.”

MD: “Maybe you could do them at night? Before bed?”

PT: “But then I may get cramps when I'm sleeping?”

MD: “How about a walk at lunch?”

The second conversation profile, the “How to” profile, consisted of instructions on how to perform exercises, opinions about the usefulness of exercise and physical therapy instruction, and the impact of these on function. This discussion illustrates a typical “How to” conversation:

PT: “I have been having trouble lately with my leg.”

MD: “Have you been doing your exercises? The leg ones?”

PT: “Sometimes, but it doesn't seem like much.”

MD: “The quad exercises are really helpful. They will strengthen the muscle in the front of your thigh. This will make you more steady on your feet and help you to walk upstairs. You really should do them. Do you remember how?”

PT: “Not exactly.”

MD: “Well, you sit with you legs out straight, then try to push the back of your leg down against the floor—like this…”

Here the rheumatologist demonstrated how to perform the exercise and linked the performance of the exercise to a functional benefit. He also added his opinion about the usefulness of the exercise in enhancing function and his desire for the patient to perform the exercise.

The third profile, known as the “Adherence Profile,” focused on discussions pertaining to patient efficacy and adherence to exercise regimens. An example is provided:

PT: “Before you ask, I want you to know I haven't been exercising.”

MD: “Oh, really?”

PT: “It is not that I can't. I can do them. It is just that I haven't. There's no real reason. I just haven't.”

MD: “Do you plan to exercise now? Maybe you could join a health club? Or walk with a friend. You might be more likely to exercise if you go with someone.”

These profiles were then used as predictors of exercise prescriptions.

During the 70 exercise discussions, the mean number of statements/questions characterized by the “Pros and Cons” profile was similar for rheumatologists and patients, and nearly similar for the “How to” profile. However, patients contributed nearly 5 times more than rheumatologists to discussions about adherence to exercise and self-efficacy (a mean of 1.6 statements/questions for patients compared with 0.3 for doctors) (Figure 2). The proportion of positive and negative statements made by rheumatologists and patients within conversation the “How to” and “Pros and Cons” profiles were also nearly identical. Yet, there was twice as much positive talk compared with negative talk about pain, symptom relief, drawbacks, and side effects. There was also twice as much talk about functional benefits and favorable opinions of exercise than negative opinions about exercise. On average, patients' comments about their self-efficacy for exercise and adherence to an exercise were more positive than rheumatologists' comments about patients' ability to adhere to or perform exercises (Figure 2).

thumbnail image

Figure 2. Total profile talk about exercise and physical therapy by rheumatologists and patients.

Download figure to PowerPoint

Because the amount of exercise discussion was associated with who initiated the discussion, we examined each profile based on whether it was rheumatologist- or patient-initiated. (Figure 3). When the doctor initiated the exercise discussion, there were no significant differences in the mean number of statements/questions in either “Pros and Cons” profile (a mean of 2.8 for doctors compared with a mean of 1.2 for patients) or “Adherence” profile (a mean of 2.0 for doctors and 1.85 for patients). The amount of exercise conversation characterized by the “How to” profile was greater, although not significantly, when the doctor initiated the exercise discussion (mean for doctor-versus patient-initiated discussions 4.5 versus 2.5; P = 0.07).

thumbnail image

Figure 3. Differences in the amount of exercise discussion characterized by each profile depending on who initiated the therapeutic exercise discussion.

Download figure to PowerPoint

Next, we examined whether differences existed between the groups (prescription versus non-prescription) on the amount, type, and differences in profiles of conversations between rheumatologists and patients. Discussions including exercise prescriptions contained more talk (mean of 19.3 statements/questions compared with 9 statements/questions; T score −3.3; P = 0.003) and were characterized by “How to” conversations. These discussions contained 3 times more talk about opinions of exercise and functional benefits and included instructions on how to perform exercises when compared to conversations in which exercises were not prescribed (mean for prescription discussion 7 versus mean without prescriptions discussion 2.4; T score −2.6; P = 0.02). However, these 2 groups were not statistically different in the amount of discussion fitting the “Pros and Cons” profile (T score −0.53; P = 0.6) or the “Adherence ” profile (T score 1.6; P = 0.1) (Table 6). An exercise prescription was 3 times more likely to occur when the doctor initiated the exercise discussion.

Table 6. Differences in the amount and type of exercise discussion between patients who received an exercise prescription and those who did not (n = 70)*
 Prescription (n = 18)Non-prescription (n = 52)
  • *

    Values are mean ± SD

Total talk19.0 ± 12.09.0 ± 10.0
 Profile 1: “Pros and Cons”2.4 ± 3.01.9 ± 4.0
 Profile 2: “How To”7.0 ± 7.32.4 ± 2.4
 Profile 3: “Adherence”1.4 ± 1.42.1 ± 2.2
Profile difference: (rheumatologist-Patient)  
 Difference Profile 1: “Pros and Cons”0.1 ± 0.8−0.3 ± 1.9
 Difference Profile 2: “How To”1.8 ± 3.60.2 ± 1.8
 Difference Profile 3: “Adherence”−0.94 ± 1.1−1.54 ± 1.8

Differences in talk by type of exercise and physical therapy.

Figure 4 demonstrates the differences in conversation profiles according to discussions about referral to physical therapy for exercise and the type of exercise prescribed. Conversations pertaining to aerobic exercise were characterized by twice as much discussion about the “Pros and Cons” of exercise compared with discussions about general range of motion and strengthening exercises. Aerobic exercise discussions had nearly 5 times more talk about side effects, drawbacks, pain, and symptom relief compared with conversations about referrals to physical therapy. Discussions characterized by the “How to” profile were similar across the 3 categories (aerobic exercise, general strengthening and range of motion, and physical therapy). However, discussions about physical therapy referrals for exercise contained only half as much conversation about compliance and efficacy compared to discussions about aerobic or general exercises.

thumbnail image

Figure 4. Differences in conversation profiles according to discussions about types of exercise or physical therapy.

Download figure to PowerPoint

Report and predictors of exercise prescription.

Post-encounter reports of prescriptions were compared with the transcripts to determine if patients recognized whether an exercise prescription was given. Using a pairwise comparison, there were no significant differences in patient and rheumatologist reports of an exercise prescription. In fact, 83% of the time rheumatologists and patients agreed on whether an exercise prescription was made (Xinline image = 50.3; P = 0.001).

Multivariate logistic regression and stepwise methods identified 3 significant predictors of an exercise prescription: rheumatologist initiated discussion about exercise (odds ratio [OR] 4.6; P = 0.03); conversations characterized by “How To” profile (OR 3.1; P = 0.01); and the number of non-exercise treatments discussed such as medications and home modifications (OR 1.6; P = 0.01). This model predicted the correct response (prescription versus no prescription) in 81% of the cases.


  1. Top of page
  2. Abstract

We described clinical discussions of exercise between patients with rheumatoid arthritis and their rheumatologists and identified predictors of an exercise prescription. Three conversation profiles were identified that characterized exercise discussions. The conversation profiles make sense clinically, and are consistent with the literature on the benefits of exercise. For example, although research demonstrates the benefit of moderate aerobic exercise for patients with rheumatoid arthritis (3–5), it is less frequently used, and rheumatologists' attitudes toward aerobic exercise are less positive than their attitudes towards other forms of exercise. In this study, conversations about aerobic exercise reflected patients' and rheumatologists' concerns about potential side effects, pain, and symptom exacerbation. These concerns led to more negotiation about aerobic exercise than strengthening and range-of-motion exercises that are viewed as having fewer potential side effects (4, 6, 7, 10). It is not surprising then that discussions about range-of-motion exercises and physical therapy were primarily characterized by the “How to” profile.

On average, exercise discussions revealed a greater proportion of “How to” talk, evenly distributed between rheumatologists and patients, indicating equal sharing of information. This process of mutual decision making and sharing of opinions enhances adherence to treatments, and results in improved satisfaction with care (29–32). However, rheumatologists and patients differed substantially regarding the amount of talk devoted to exercise adherence and self-efficacy. In fact, there were nearly 3 times more discussion from patients compared with rheumatologists on this topic. Research demonstrates that self-efficacy for a behavior and adherence to that behavior are positively correlated (33–36). Perhaps patients are more knowledgeable about exercise than other therapies used to manage their disease, leading to more conversation on this topic. Another possible explanation for this disparity may be that conversation reflects rheumatologists' beliefs and attitudes toward exercise. Rheumatologists may be less responsive to talk about adherence especially since they reported their patients would be less likely to adhere to exercise prescriptions. Exercise prescriptions were more likely to occur with discussions characterized by the “How to” profile, which included offering opinions about exercise and linking exercise behavior to functional benefits. Establishing the association between a behavior (exercise) and an outcome (improved functional status), while providing instructions on how to perform the behavior (exercise), reflects the characteristic of a typical prescription and has been shown to improve adherence and patient outcomes (37). Rheumatologists may view exercises as one method of reducing symptoms and discuss exercise as they would other treatment opinions. Exercise discussion may then take the form of a lecture to patients, with explicit instructions on how to perform the exercise, as well as reinforcement of the desired behavior through sharing of opinions about exercise. Perhaps talk is responsive to the patient's needs. For example, patients who clearly do not appear to understand what is being described (how to do an exercise) may require more extensive instructions and reinforcing, translating into more “How to” conversations.

Conversation profiles also differed based on the type of exercise prescription. Conversations contained more discussion about adherence when exercise regimens were prescribed than when a referral was made for physical therapy. Perhaps the greater emphasis on adherence discussions reflects the rheumatologists' concern with prescribing and instructing patients in home exercise. The current literature estimates that one-third to two-thirds of patients are non-adherent to exercise prescriptions (33, 38–42) and that long-term adherence to exercise is substantially lower than adherence to short-term exercise (34). Unfortunately, adherence rates among persons with chronic disease are significantly lower than adherence rates among individuals with acute medical illnesses (34, 43). By emphasizing the importance of adherence, offering alternative forms of exercise and addressing patients' self-efficacy regarding undertaking an exercise program, rheumatologists may enhance patients' motivation to follow the prescribed regimen (44).

Discussions about other treatment options (medications, home modifications) were also associated with a greater likelihood of an exercise prescription. There may be 2 explanations for this result. Perhaps, as rheumatologists offer patients more options, allowing patients a more inclusive role in decision making, patients may be more likely to follow an exercise regimen, given that they have positive attitudes about exercise. Perhaps as more treatment options are discussed, some with potentially serious side effects (gold, Plaquenil), patients may view therapeutic exercise as a more tolerable option. Decision-Making Theory (14) supports these explanations. As patients fully examine their various treatment options and weigh the costs and benefits of each, they increase the likelihood they will adhere to the regimen (14, 19, 44).

Finally, this study demonstrates the rheumatologists' influence on the content of exercise discussions and the likelihood of an exercise prescription. Clearly, when rheumatologists initiate exercise discussions there is a dramatic increase in the likelihood of receiving a prescription, even though patients contributed 5 times more to exercise discussions and were twice as likely to make positive statements about exercise. Patients may be more open to exercise as a method for managing their disease as they enter the visit, but may be less likely to receive an exercise prescription even when they initiate exercise discussions, because they do not demonstrate persuasive communication powers. Rheumatologists, viewed as experts, may be more likely to motivate patients to attend to their suggestions and to engage in a discussion about the pros and cons of exercise, influencing patients' attitudes toward exercise (16).

This is the first audiotaped study of rheumatologist-patient discussions about exercise in managing rheumatoid arthritis. It examines the quality of the clinical discussions and the impact of exercise discussions on the likelihood of receiving an exercise prescription. The strengths of this study include the use of audiotapes of clinical discussions; the prospective design of the study; double data entry to reduce data entry errors; the use of a reliable method of coding and quantifying the audiotape data; and the use of well-trained coders who were unaware of patient and provider reports of attitudes and beliefs about exercise and physical therapy instruction for exercise.

Limitations of the study should also be acknowledged. These results may not be generalizable to other patients with arthritis, as the sample consisted primarily of white, educated women, whose conversations may tend to differ from those of other patients with rheumatoid arthritis. Also, the low response rate limits the generalizability of the study results. Due to the cross-sectional nature of the study we do not have information on any prior exercise discussions these patients may have had with other providers. The potential exists for random misclassification of the outcome, exercise prescriptions, because patients and rheumatologists may have conceptual differences in their definition of exercise prescriptions. However, post-encounter reports of exercise prescriptions and actual prescriptions, as noted on the transcript data, were comparable, making misclassification less likely. Although the attitudes and beliefs scale has not been previously validated, the questionnaire was based on the theoretical framework of Ajzen and Fishbein (12, 13) that has been proven reliable and valid in a variety of settings.

The attitudes and beliefs of rheumatologists and patients affect the dynamics of clinical discussions. These beliefs reflect personal experiences as well as shared discussions in the clinical encounters, and have a direct impact on exercise prescription rates. The duration of the rheumatologist-patient relationships may therefore affect patient attitudes, beliefs, behaviors, and clinical discussions surrounding the usefulness of exercise in arthritis management. Exercise and physical therapy are discussed differently during the clinical encounter and are 4 times more likely to occur when the rheumatologist initiates exercise discussion. These discussions strongly impact on the likelihood a patient receives an exercise prescription.


  1. Top of page
  2. Abstract
  • 1
    Curfman GD. The health benefits of exercise: a critical reappraisal. N Engl J Med 1993; 328: 5746.
  • 2
    Boutelle KN, Murray DM, Jeffery RW, Hennrikus DJ, Lando HA. Associations between exercise and healthy behaviors in a community sample of working adults. Prev Med 2000; 3: 21724.
  • 3
    Jensen G, Lorish C. Therapeutic exercise in rheumatoid arthritis. Arthritis Care Res 1990; 7: 1819.
  • 4
    Minor MA. Arthritis and exercise: the times are a-changin'. Arthritis Care Res 1996; 9: 7981.
  • 5
    Ekdahl C, Andersson SI, Moritz U, Svensson B. Dynamic versus static training in patients with rheumatoid arthritis. Scand J Rheum 1990; 19: 1726.
  • 6
    Hicks JE. Exercise in patients with inflammatory arthritis and connective tissue disease. Rheum Dis Clin North Am 1990; 16: 84570.
  • 7
    Minor MA, Hewett JE, Webel RR, Anderson SK, Kay DR. Efficacy of physical conditioning exercise in patients with rheumatoid arthritis and osteoarthritis. Arthritis Rheum 1989: 32, 1396405.
  • 8
    Van den Ende CH, Hazes JM, le Cessie, S, Mudler WJ, Relfor DG, et al. Comparison of high and low intensity training in well controlled rheumatoid arthritis: results of a randomized clinical trail. Ann Rheum Dis 1996; 55: 798805.
  • 9
    Daltroy LH, Robb-Nicholson C, Iversen MD, Wright EA, Liang MH. Effectiveness of minimally supervised home aerobic training in patients with systemic rheumatic disease. Br J Rheumatol 1995; 34: 10649.
  • 10
    Iversen MD, Fossel AH, Daltroy LH. Rheumatologist-patient communication about exercise and physical therapy in the management of rheumatoid arthritis. Arthritis Care Res 1999; 12: 18092.
  • 11
    Godin G, Shephard R. Importance of the type of attitude to the study of exercise behavior. Psychol Rep 1986; 58: 9911000.
  • 12
    Ajzen I. From intentions to actions: a theory of planned behavior. In: Kuhl J, Beckman J, editors. Action control: from cognition to behavior. Heidelberg: Springer-Verlag; 1983.
  • 13
    Ajzen I, Fishbein M. Understanding attitudes and predicting social behavior. Englewood Cliffs (NJ): Prentice Hall; 1980.
  • 14
    Janis I, Mann L. Decision-making: a psychological analysis of conflict, choice, and commitment. New York: The Free Press; 1977.
  • 15
    Cialdini RB, Petty RE, Cacioppo JT. Attitude and attitude change. Annu Rev Psychol 1981; 32: 357404.
  • 16
    O'Keefe DJ. Persuasion. Newbury Park (CA): Sage; 1990.
  • 17
    Iverson DC, Fielding JE, Crow RS, Christenson GM. The promotion of physical activity in the United States population: the status of progression medical, worksite, community, and school settings. Public Health Rep 1985; 100: 21224.
  • 18
    Nader PR, Taras HL, Sallis JF, Patterson TL. Adult heart disease prevention in childhood: a national survey of pediatricians' practices and attitudes. Pediatrics 1987; 79: 84350.
  • 19
    Wankel LM. Decision-making and social support strategies for increasing exercise involvement. J Card Rehabil 1984; 4: 12435.
  • 20
    Cegala DJ, Marinelli T, Post D. The effects of patient communication skills training on compliance. Arch Fam Med 2000; 9: 5764.
  • 21
    Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for classification of rheumatoid arthritis. Arthritis Rheum 1988; 31: 31524.
  • 22
    Ware JE. SF-36 health survey: manual and interpretation guide. Boston: The New Health Institute, New England Medical Center; 1993.
  • 23
    Ware JE, Sherbourne CD. The MOS 36-item short form health status survey (SF-36): conceptual framework and item selection. Med Care 1992; 30: 47383.
  • 24
    Lorig K, Chastain R, Ung E, Shoor S, Holman H. Development and evaluation of a scale to measure perceived self-efficacy in people with arthritis. Arthritis Rheum 1989; 32: 3744.
  • 25
    Beisecker AE, Beisecker TD. Patient information seeking behaviors when communicating with physicians. Med Care 1990; 28: 1928.
  • 26
    Roter DL. The Roter method of interaction process analysis. (RIAS Manual); 1991.
  • 27
    Rosner B. Fundamentals of Biostatistics. Boston: Duxbury Press; 1995.
  • 28
    SAS Institute Incorporated. SAS/STAT user's guide, release release 6.03 edition. Cary (NC): SAS Institute; 1988.
  • 29
    Buller MK, Buller DB. Physicians' communication style and patient satisfaction. J Health Soc Behav 1987; 28: 3705.
  • 30
    Eisenberg, JM. Sociologic influences in decision making by physicians. Ann Intern Med 1979; 90: 95764.
  • 31
    Ong LML, Dettaes JCJM, Hoos AM, Lammes FB. Doctor-patient communication: a review of the literature. Soc Sci Med 1995; 40: 90318.
  • 32
    Roter DL, Hall JA. Doctors talking with patients/patients talking with doctors: improving communication in medical visits. Westport (CT): Auburn House; 1992.
  • 33
    Ice R. Long-term compliance. Phys Ther 1985; 65: 18329.
  • 34
    Sluijs EM, Kok GJ, Van der Zee J. Correlates of exercise compliance in physical therapy. Phys Ther Rev 1993; 73: 77182.
  • 35
    Vidmar PM, Rubinson L. The relationship between self-efficacy and exercise compliance in a cardiac population. J Cardiopulm Rehabil 1994; 14: 24654.
  • 36
    Wichowski HC, Kubsch SM. The relationship of self-perception of illness and compliance with health care regimens. J Adv Nurs 1997; 25: 54853.
  • 37
    Kaplan SH, Greenfield S, Ware JE. Assessing the effects of physician-patient interactions on the outcomes of chronic disease. Med Care 1989; 27 Suppl: S110.
  • 38
    Dishman RK. Compliance/adherence in health-related exercise. Health Psychol 1982; 1: 23767.
  • 39
    Dubbert PM, Martin JE, Epstein LH. Exercise. In: KA Holroyd, TL Greer, editors. Self-management of chronic disease handbook of clinical interventions and research. Orlando (FL): Academic Press; 1986.
  • 40
    Harkpaa K, Jarvikoski A, Mellin G, Hurri H. A controlled study on the outcome of inpatient and outpatient treatment of low back pain. Scand J Rehabil Med 1989; 21: 819.
  • 41
    Jette AM. Improving patient cooperation with arthritis treatment regimens. Arthritis Rheum 1982; 25: 44753.
  • 42
    Martin JE, Dubbert PM. Exercise applications and promotion in behavioral medicine: Current status and future directions. J Consult Clin Psychol 1982; 50: 100417.
  • 43
    Price JL. Compliance: convincing patients it's not a matter of choice. J Am Acad Physicians Asst 1996; 6: 5161.
  • 44
    London F. Improving compliance, what you can do. RN 1998; 61: 436.