• Self-management;
  • Low back pain;
  • Primary care patients;
  • Intervention design;
  • Recruitment


  1. Top of page
  2. Abstract
  8. Appendix A


To describe the rationale and design of a self-management program for low-income, urban, primary care patients with acute low back pain. Issues related to recruitment and protocol delivery, and attendance patterns and predictors of program attendance are described.


Two hundred eleven adult patients (73% female; 60% African American) were recruited from primary care neighborhood health centers. Focus groups were conducted for program development, and participants then completed a baseline interview and were randomized into groups receiving either usual care or a self-management intervention.


Twenty-nine percent of the intervention group attended the self-management class. Significant predictors of attendance included being older, reporting less income, and not working for pay. Attendees did not differ from nonattendees on back pain severity, symptoms, health-related quality of life, self-management processes, or satisfaction with care.


Effective minimal-contact behavioral interventions are needed to reach larger portions of the patient population.


  1. Top of page
  2. Abstract
  8. Appendix A

Self-management programs promote patient responsibility for managing daily health in conjunction with traditional health care (1). By building behavioral skills that promote self efficacy and self management, these programs may ultimately improve patients' health status. The Arthritis Self-Management Program (ASMP) epitomizes a successful strategy that has consistently improved health outcomes and reduced health care utilization among patients with various rheumatic conditions (2–5). Given the prevalence and costs associated with acute low back pain in the US (6), particularly among socioeconomically vulnerable persons (7), we sought to tailor the ASMP to low-income, inner-city patients with acute low back pain (ALBP). We designed our program to be sensitive to potential recruitment difficulties, protocol delivery, attendance, and adherence goals.

Interventions seeking to instill behavioral change typically recruit highly motivated subjects using strategies such as newspaper advertisements and organizational mailing lists (8, 9). Although these strategies may increase attendance and adherence to program goals, using volunteers introduces selection bias (e.g., those who are interested in exercise are more likely to join and follow a physical activity program). These methods also tend to attract well-educated, middle-upper class whites (10, 11). As a consequence, programs using such strategies would generalize to highly selected volunteers, rather than the majority of patients with medical conditions.

Protocol delivery strategies represent another critical consideration in designing the program. ASMP uses group sessions for 6 weeks or longer (4). However, our focus groups suggested that having 6 group sessions would not be pragmatic for socioeconomically vulnerable populations. Other programs for this population have reported barriers of lack of time, transportation, and flexibility with work schedules or childcare (12). Thus, efficacious programs may be ineffective among vulnerable populations with such barriers.

This article describes how we implemented and revised strategies for patient recruitment, protocol delivery, and attendance patterns for a randomized, controlled trial of a self-management intervention among low-income, inner-city primary care patients with ALBP. In addition, we identified baseline patient predictors of program attendance to elucidate correlates of program acceptance among this population.


  1. Top of page
  2. Abstract
  8. Appendix A


We conducted a longitudinal, randomized, controlled trial of an ALBP self-management program (ALBPSMP). Upon receiving approval from our local institutional review board to conduct this study, we randomized eligible patients who completed informed consent and a baseline interview to 2 groups: 1 group to receive usual care and 1 to receive ALBPSMP. We followed them for 1 year. Persons blinded to group assignment collected outcome data at 4 and 12 months. Our primary outcome measure was class attendance. Baseline predictors of attendance included health-related quality of life (HRQOL), back pain symptoms, patient satisfaction, arthritis self efficacy, fear of injury/avoidance of activity, social support, and self-management practices.


We recruited patients receiving care at university-affiliated neighborhood health centers, located throughout the inner city of Indianapolis and from the emergency department of an inner-city, public teaching hospital. Inclusion criteria included the following: 1) a primary visit diagnosis reflected ALBP; 2) the current ALBP episode had lasted less than 3 months as reported by either physician or patient and was not due to severe trauma; 3) the patient was deemed eligible for study by the primary care physician; 4) patient age was ≥18; and 5) the patient had access to a working telephone. We required access to a telephone because data collection was initiated by telephone contact.

Patients were excluded from the study if they met any of the following criteria: 1) chronic back pain (≥90 consecutive days of back pain); 2) prior surgery for back pain; 3) either receiving disability insurance payments or in the process of applying for back pain disability; 4) residing in an institution; 5) being incompetent for interview; 6) severely impaired in vision, hearing, or speech; 7) unable to understand and speak English; 8) being pregnant, having a terminal illness with a life expectancy of less than 1 year; or 9) having a severe comorbid condition deemed by themselves or their primary care physician as limiting their functional ability. We designed these criteria to exclude adults for whom the proposed interventions were inappropriate or for whom there may have existed disincentives for improvement. The Regenstrief medical record system (RMRS) contained most of these data, and computerized RMRS audits identified potentially eligible patients (13). Physicians, manual chart audits, and initial patient telephone contact were used to apply the remaining eligibility criteria.


Patients were enrolled as quickly as possible after their initial visit for ALBP. Because all physicians at our primary care and emergency department study sites use the RMRS microcomputer workstations to enter all orders (13), the staff programmed the RMRS to notify physicians about potential patients using computerized alerts generated when orders were entered. When the primary visit diagnosis suggested the patient had ALBP, the alerts briefly described the study and its eligibility criteria, then asked the physicians whether the patient was eligible for our study and competent for interviews. The physicians' responses (“yes” or “no”) were automatically E-mailed to our project coordinator who then screened for study eligibility among those patients marked “yes” by their treating physicians.

We mailed a recruitment letter to potentially eligible patients that described the study and contained their physician's recommendation for participation. The letter was followed within a few days by a telephone call to confirm eligibility and determine the patient's interest in participating. For patients agreeing to participate, an interviewer blinded to our hypotheses obtained informed consent and conducted the baseline interview at either the primary care center, the patient's home, or by telephone. An attempt was made to schedule all baseline interviews within 7 days of confirming eligibility. We paid patients $10 after completing each interview. Immediately after completing the baseline interviews, the patients were randomized to the self-management intervention or the usual care group.

Intervention group

Patients randomized to the ALBPSMP were invited to attend 3 weekly self-management sessions held at the closest primary care center. The sessions were based on social-cognitive theory (14) and focused on increasing self efficacy and social support to self manage low back pain. Participants also received 3 followup telephone calls (1 weekly and 2 biweekly) and 10 monthly followup calls from the project coordinator to discuss behavioral contracts and problem solve to sustain behavioral change until the end of the study.

Usual care group

All participants in the study received usual care. Based on the physician's judgment, this could include referral to occupational therapy, physical therapy, or a neurological center; nonnarcotic or narcotic analgesics; and back exercise sheets. Half of the participants were randomized to receive usual care only. The other half received usual care plus the intervention.

Program structure

To maximize the relevance of the program for our population, we conducted focus groups to develop, pilot test, and finalize the intervention program and educational materials. Primary care patients with a visit to the primary clinic within the past 2 months for ALBP were identified by the RMRS and screened by staff for eligibility. A total of 144 patients visited the primary care clinic for back pain during the 2 months preceding the focus groups. Of these 144 patients, 10 met the eligibility criteria. Of these 10 eligible patients, 8 attended 2 focus groups (4 attended each group). Topics that were discussed included: patients' understanding of their ALBP; perceptions regarding the effectiveness of treatments (both those prescribed and not prescribed by physicians); suggestions regarding the structure and content of the intervention program; preferences regarding the program leader (e.g., physical therapists, nurses, lay leaders, community leaders); and perceived barriers to participation and retention. We found that patients expressed frustration with the recurrence of ALBP, disappointment with inadequate treatments, anxiety regarding progression of pain and functional limitations, and a desire to prevent future back pain episodes through exercise. The patients did not have a preference for group leaders (1 patient preferred a physician, 1 patient preferred a physical therapist, and 2 patients preferred anyone who had experienced back pain to lead the groups) and indicated that they would attend a maximum of 4 classes. Thus, the protocol, which initially called for 6 weekly classes, was reduced to 4 classes with 2 followup telephone calls.

A separate focus group to evaluate all educational materials found that patients accepted and understood the content. All 4 patients reported that they understood the content and evaluated the format, print, colors, pictures, and content of the educational materials. The pilot test of the intervention revealed 2 issues: patients preferred to attend intervention classes at their neighborhood health centers rather than in church or community centers; and patients were not willing to attend 4 classes. Therefore, we changed the venue of the intervention classes to the community rooms of the neighborhood health centers, and reduced the intervention to 3 classes and 3 followup telephone calls. The content of the final intervention program is outlined in Table 1. The content of the telephone calls focused on problem solving and behavioral contracting.

Table 1. Content of acute low back pain self-management program
Class I
 1. Basic understanding of back pain
  a. Sources of back pain
  b. Defining “serious” back pain
  c. Distinguishing “hurt” from “harm”
  d. Red flag symptoms
  e. Back pain vocabulary (i.e., words they may hear, with simple explanations)
 2. Factors influencing pain
 3. Activities for fun and fitness
  a. Flexibility and movement
  b. Strengthening
  c. Aerobic-type activities
  d. Monitoring activities
 4. Introduction to goal setting
 5. Behavioral contracting
Class II
 1. Feedback/problem solving
 2. Activities for fun and fitness
 3. Relaxation—deep breathing
 4. Overcoming fears
 5. Dealing with negative emotions
 6. Role of positive thinking
 7. Behavioral contracting
Class III
 1. Feedback/problem solving
 2. Activities for fun and fitness
 3. Handling back pain flares
 4. Working with health care providers and employers
 5. Reviewing and sharing accomplishments
 6. Identifying residual fears/planning ahead
 7. Behavioral contracting

Members of our research team led intervention sessions using a standardized, written protocol. To promote self efficacy, the program began with patient education on the natural history of ALBP, emphasizing the intermittent nature of the condition. This was designed to enhance patients' optimism for a rapid recovery, but provide realistic expectations regarding the likelihood of recurrence. We reviewed common diagnostic and treatment strategies for ALBP. Because patients often expected physicians to order diagnostic tests, the educational component also described the rationale for ordering (or not ordering) diagnostic tests. We encouraged patients to discuss both traditional and nontraditional therapies that they had heard of or used for treating ALBP. Session leaders promoted behaviors based on current evidence-based treatment recommendations found to have at least moderate support in the medical literature. In particular, the instructor encouraged patients to minimize bed rest and return to normal activities as soon as possible. In addition, patient education materials showed recommended exercises, including walking. Patients were encouraged to adhere to their physicians' treatment recommendations, as well as to reduce their body weight (when appropriate).

To promote behavior change, the program included goal setting and problem solving. Barriers to engaging in these behaviors were discussed during the group sessions and followup telephone calls. Participants chose behaviors to modify, and group leaders encouraged them to solve their individual problems rather than receive prescribed programs. To promote perceived self efficacy, participants received individualized feedback about their progress from their group leaders, other participants when possible, and their personal physicians (by correspondence prepared by staff). In addition, patients were encouraged to ask questions of their providers and to use positive talk and relaxation, two strategies that have been successful in the ASMP to manage negative affect (2, 3).

Patients recorded their behaviors and progress on a standardized form. We made 3 copies of this progress report and gave 1 to the patient, 1 to the patient's primary care physician, and kept the third in the research office.

Social support was promoted in 2 ways. First, patients were encouraged to bring their spouse or significant other to all sessions of the program. Second, based on the progress reports, we sent letters to intervention patients that were printed with the scanned signature of the primary care physician, generated within 2 days after each session, and tailored to the content of each session. These letters encouraged patients' further participation in the program. In addition, patients received monthly followup telephone calls to check on progress and assist with problem solving.


Our main study outcome was class participation in the ALBPSMP. To predict class participation, we utilized baseline data that included patient demographics and our primary and secondary outcomes of the randomized controlled trial.

We obtained patients' baseline data from the RMRS and patient interviews. Our primary outcome measures of the intervention trial included HRQOL (i.e., functional status, back pain symptoms) and patient satisfaction. Secondary outcomes of the trial comprised fear of movement/activity, self efficacy, self-management behaviors, role function, and social support. We also collected baseline demographic information. The number of items per measure, measure definitions, and response codings are presented in Appendix A. Demographic information included sex, age, race, marital status, education, living arrangement, income (15), and employment status. Specific back pain questions included total years of having back pain, frequency of medication use during the past month, use of remedies for low back pain other than medication or surgery, and severity of low back pain during the past 4 weeks. We also assessed HRQOL using the Arthritis Impact Measurement Scales (AIMS2) (16–18) and supplemented the AIMS2 with the low back pain-specific Roland Disability Questionnaire (19, 20). To assess the process of self management, we included self efficacy to self manage (21), a fear of movement/(re)injury (22), and minutes spent during last week walking for exercise (21). In addition, we included a specific patient satisfaction measure of low back pain care received (23).


We recorded patient recruitment, characteristics, and ALBPSMP attendance throughout the randomized controlled trial. Furthermore, we conducted univariate analyses of baseline predictors of ALBPSMP attendance as defined by attendance of 1 or more classes. For normally distributed continuous predictors, t-tests were used. Chi-square tests (or Fisher's exact test where appropriate) were used for categorical predictors and Wilcoxon rank sum tests were used for ordinal predictors and continuous predictors failing to meet the normality and/or equal variance assumptions of t-tests.


  1. Top of page
  2. Abstract
  8. Appendix A


A total of 5,669 patients were identified by the RMRS who visited the urban primary care neighborhood health center and had a back pain diagnosis during the 2-year recruitment period. Of those, 5,129 were ineligible for the following reasons: physician's clinical judgment that the patient did not meet our criteria (38%), back pain was chronic and/or caused by severe trauma (18%), incompetence (e.g., substance abuse), resided in an institution (18%), lacked telephone access or valid address (16%), severe comorbidity (e.g., oxygen dependent; 4%), prior surgery for back pain (2%), current visit was for a problem other than ALBP (2%), the patient currently received disability payments for back pain or had submitted disability claim and under review (1%), or pregnancy (1%). The remaining 540 patients (see Figure 1) fulfilled eligibility criteria for participation in our study. Of this total, 281 (52%) were willing to participate, 211 of whom were randomized to 1 of the 2 groups. The remaining 70 patients either became unreachable or ineligible before baseline assessment (i.e., no telephone access).

thumbnail image

Figure 1. Summary of recruitment and randomization.

Download figure to PowerPoint

Sample characteristics

The average age of the patients randomized in this study was 46 years (range 19–83 years). The sample was 73% female, 60% African American, 1% Hispanic, 31% married, and 18% living alone. Approximately 19% reported that they had sufficient income to meet their needs, 62% reported that their income was enough to “just make ends meet,” and 15% indicated that they were “unable to make ends meet.” The remaining 4% did not know or refused to answer. In addition, ∼62% completed high school.

Program attendance

Table 2 presents program attendance by delivery mode during the randomized controlled trial for all 3 classes. During the initial program, participants randomized to the ALBPSMP were invited to attend scheduled class sessions. After the first 3 months of program implementation, class attendance was less than expected. Subsequently, 3 adjustments were made to the protocol to increase intervention exposure. Table 2 under “initial program” displays attendance when the protocol was implemented as a traditional self-management program. That is, we invited patients to attend group sessions, but provided no alternative mode of delivery for those who did not attend. Patients reported that they were unable to attend because of schedule conflicts and transportation problems. Therefore, we made the first adjustment and mailed class materials to absent patients and followed up with telephone calls to review the material. Table 2 illustrates participant attendance after “adjustment one” of the protocol delivery during the randomized controlled trial. One-third of the enrolled patients attended the class sessions. The remainder received the class materials by mail, but only 2 were available to receive telephone followup.

Table 2. Acute low back pain self-management program attendance*
Intervention delivery modeWeekly classes
Class 1 n (%)Class 2 n (%)Class 3 n (%)
  • *

    During the initial program, participants randomized to the acute low back pain self-management program were invited to attend scheduled class sessions. During adjustment one, participants who did not attend class sessions were mailed class materials and telephoned. During adjustment two, participants who did not attend class sessions were mailed audiotapes of the classes, a hand-held cassette player, and class materials. Followup telephone calls were attempted. No intervention during this time period denotes that mailed materials/audiotapes were returned because patient moved without a forwarding address. During adjustment three, participants received individual sessions with staff at the neighborhood health centers to meet their schedules when they could not attend group sessions.

Initial program (n = 11)   
 Attended class4 (36)3 (27)1 (9)
 Mailed information/no telephone call001 (9)
 No intervention7 (64)8 (73)9 (82)
Adjustment one (n = 21)   
 Attended class7 (33)6 (29)6 (29)
 Mailed information/no telephone call12 (57)13 (62)13 (62)
 Mailed information/telephone call2 (10)2 (10)2 (10)
Adjustment two (n = 34)   
 Attended class7 (21)6 (18)5 (15)
 Audiotapes/no telephone call13 (38)20 (59)22 (65)
 Audiotapes/telephone call11 (32)5 (15)4 (12)
 No intervention3 (9)3 (9)3 (9)
Adjustment three (n = 40)   
 Attended class12 (30)8 (20)7 (18)
 Audiotapes/no telephone call19 (48)21 (53)23 (58)
 Audiotapes/telephone call9 (23)5 (13)3 (8)
 No intervention01 (3)1 (3)

To provide social models to patients not attending the group sessions, we made a second adjustment and audiotaped the 3 class sessions. We provided these tapes, hand-held audiocassette players, and batteries to all patients for the remainder of the study. “Adjustment two” in Table 2 shows intervention exposure for participants after we provided audiotapes of the classes. This modification dramatically increased the intervention exposure for the treatment group members. Classes 1, 2, and 3 were attended by 19%, 18%, and 15% of the treatment patients respectively, and an additional 32%, 15%, and 12% of patients received the intervention by audiotape and telephone followup for the 3 classes, respectively. During telephone followup, patients were asked if they had listened to the tapes. If they had not, patients were asked to listen to the tapes and another telephone followup call was scheduled. Finally, a third adjustment was made (Table 2, “adjustment three”) and individual sessions were scheduled at the centers to accommodate patients' schedules.

In summary, of the 106 intervention patients who were randomized to attend the ALBPSMP, 30 (28.3%) attended at least 1 class, 66 (62.3%) received the intervention by mail/telephone/audiotape, and 10 (9.4%) received no intervention. The size of the neighborhood health center classes ranged from 1 to 3 patients.

Baseline predictors of program attendance

Univariate analyses of baseline patient characteristics indicated that only demographic variables significantly predicted attending at least 1 class session (Table 3). Specifically, patients who attended at least 1 intervention class were more likely to be older, unable to make ends meet financially, and less likely to work for pay. Patients that attended the intervention classes did not differ from nonattendees in back pain, HRQOL, self-management process, or patient satisfaction for back pain care.

Table 3. Baseline predictors of acute low back pain self-management program attendance: univariate analyses*
VariableAttended class one (n = 30)Did not attend any class (n = 76)P
  • *

    Means and percentages, where noted, are presented in the columns. Standard deviations are presented in parentheses. For severity of low back pain and health-related quality of life (except low back pain specific functioning), a lower score indicates better health. Scale ranges are included with the variables.

  • Indicates a significant P value.

 Women, %66.773.70.47
 Age, years51.4 (14.3)43.2 (13.4)0.01
 African American, %66.751.40.15
 Married, %
 Education <12 years, %37.940.00.85
 Lives alone, %23.322.40.91
 Income (unable to make ends meet), %
 Works for pay, %33.357.90.02
Back Pain   
 Years of back pain9.8 (9.4)7.1 (7.5)0.13
 Frequency of medication use, 1–52.3 (1.2)2.6 (1.3)0.34
 Remedies used other than medication/surgery, %
 Severity of low back pain, 0–108.1 (2.5)7.9 (2.4)0.85
Health-related quality of life   
 Physical function, 0–102.2 (1.5)2.7 (1.7)0.25
 Affect, i.e., mood, 0–104.1 (2.0)4.7 (2.1)0.20
 Back symptoms, 0–105.9 (2.1)6.3 (2.3)0.40
 Social interaction, 0–103.7 (2.0)3.7 (1.9)0.88
 Role function, 0–103.4 (3.0)3.6 (2.6)0.84
 Low back pain specific functioning, 1–2415.4 (7.3)15.3 (6.3)0.78
 Self-efficacy to self-manage, 1–105.7 (2.1)6.0 (1.9)0.47
 Fear/avoidance of activity, 1–4025.4 (7.5)24.9 (6.8)0.64
 Walk for exercise, minutes/week65.0 (66.7)48.6 (62.2)0.31
Patient satisfaction   
 Satisfaction with low back pain care received, 10–3022.5 (3.9)22.3 (4.3)0.92


  1. Top of page
  2. Abstract
  8. Appendix A

The traditional self-management programs (e.g., ASMP) usually include at least 6 weekly group meetings that are typically held in the community and are facilitated by volunteer lay leaders. We presented a modified self-management program for inner-city primary care patients with ALBP. Unlike the ASMP, the patients in our program presented with acute pain of the low back rather than chronic arthritis pain, attended 3 group meetings compared with the usual 6, attended in smaller groups, and the groups were facilitated by staff rather than volunteer lay leaders. In addition, we actively recruited patients from primary care practices rather than by passive methods in the community.

Recruiting from a defined, clinical population enabled us to determine our eligibility and participation rates. After screening, 39% of all eligible patients were randomized into our study, a rate almost 4 times greater than that reported in a study of a low back pain self-management program conducted in a suburban, middle-class white population (5). However, the group session participation rate in that study was much higher than our study. Thus, although we reached a greater portion of the eligible patients, the group session mode did not meet the needs of all interested patients. Future analyses will provide some indication of the extent to which the different modes yield different patient outcomes. The only other self-management program that targeted alternative populations was the Spanish Arthritis Self-Management Program (24). However, this randomized trial recruited volunteers from the community, therefore eligibility and comparison rates of class attendance are not available.

Another benefit of our ALBPSMP is the involvement of patients' primary care physicians in the process of patient recruitment. Implementing a computer alert program to notify physicians of potential patients greatly improved the efficiency of the patient screening process. It also enabled the physician to screen patients as required by the institutional review board. Future work may investigate the importance of physician referral to patient involvement.

We faced several challenges in implementing a self-management program in a low-income population. First, patients often expressed difficulty with transportation, a concern unique to a socioeconomically vulnerable population. Therefore, it became difficult to hold large group sessions in central locations. Patients expressed a preference for attending sessions at their neighborhood health center rather than other centers, even when the distance was <10 miles. Despite the transportation difficulties, however, our data suggest that a significant portion of low-income patients were interested in self-management in conjunction with traditional medical care.

To identify those patients attracted to our program, we analyzed baseline patient characteristics associated with program attendance. We reported that older, poorer, and unemployed patients were more likely to attend at least 1 intervention class. Thus, our program appeared to reach patients with little financial resources and more free time. These patients were more likely to take advantage of the extra program offered free of charge for their back pain. It is possible that younger, employed patients with greater financial resources may have greater access to other resources for their back pain (i.e., over-the-counter products, coworkers' experiences) and may perceive our program as less essential. Employed patients may also have trouble getting time off to attend the classes, or the classes may compete with their time after work. The “working poor” may be particularly vulnerable, having neither the time for interventions such as ours nor the financial resources for more traditional health care interventions. Given that this population has often been underserved regarding health behavioral programs (25), additional research is needed to understand program acceptance.

The second, not unique, challenge is motivating patients to attend nonpharmaceutical treatment programs. With only one-third of our intervention group attending class sessions, alternative methods were required to deliver the intervention. We chose to audiotape class sessions and provide hand-held cassette players with batteries as an inexpensive alternative to group sessions. Although this method increased the intervention exposure to an additional 30% of the self-management group, approximately 40% of the self-management group was unavailable for telephone followup after receiving the tapes. Exposure to the first class also increased when we offered individual sessions at the convenience of the patient's schedule and clinic preference. However, individual sessions limited the group dynamics (i.e., social modeling) that may assist in behavioral change.

Offering different delivery modes of educational programs may be important for increasing patient participation in behavioral interventions, and the mode may depend on the symptom severity (acute versus chronic). Patients experiencing greater pain or disability from their medical condition may be more motivated to attend behavioral interventions.

We recognize the limitations of alternative intervention delivery methods; however, traditional self-management efforts reached only one-third of the eligible patients. As our next step, we plan to evaluate the effectiveness of our intervention. Because we are not powered to evaluate 3 separate intervention groups, we will evaluate group differences and adjust our analyses based on those differences. The results presented in this article provide a foundation that can be used to build alternative intervention programs while avoiding ineffective methods for a low-income population. Developing effective, minimal-contact behavioral interventions remains an objective for future research.


  1. Top of page
  2. Abstract
  8. Appendix A

Appendix A

  1. Top of page
  2. Abstract
  8. Appendix A
Appendix A: Measures of predictors of attending the acute low back pain self-management program*
MeasureNo. of itemsDefinitionCoded
  • *

    meds medications; LBP low back pain; HRQOL health-related quality of life; AIMS2 Arthritis Impact Measurement Scales.

 Women1Female sex0 = no 1 = yes
 Age1Age in yearsContinuous
 African American1Being African American0 = no 1 = yes
 Married1Currently married0 = no 1 = yes
 Education1<12 years of education0 = no 1 = yes
 Lives alone1Lives alone0 = no 1 = yes
 Income1Unable to make ends meet0 = no 1 = yes
 Works for pay1Employed with pay0 = no 1 = yes
Back pain   
 Years of back pain1Total years had LBPContinuous
 Frequency of medications used1Frequency of meds for LBP during past month1 = no days 5 = all days
 Other remedies used1Remedies used for LBP other than medication/surgery0 = none 1 = used other remedies
 Severity of low back pain1Severity of LBP during past 4 weeks0 = no pain 10 = severe pain
 Physical function18AIMS2—mobility, walking, bending, household tasks, self-care0 = better health 5 = poorer health
 Affect (mood)10AIMS2—level of tension, mood0–5
 Back symptoms4AIMS2—low back pain0–5
 Social interaction9AIMS2—support from others0–5
 Role function5AIMS2—work0–5
 Low back pain specific functioning24Low back pain disability1 = Describe me 0 = Does not describe me
 Self efficacy6Confidence to self manage1 = no confidence 10 = very confident
 Fear/avoidance of activity10Degree avoid activity for fear of (re)injury1 = strongly agree 4 = strongly disagree
 Walk for exercise1Minutes/weekContinuous
Patient satisfaction   
 Satisfaction with low back pain care10Satisfaction with care received1 = dissatisfied 2 = uncertain 3 = satisfied