Secondary Logo

Journal Logo

Original Investigation

Facilitated Health Coaching Improves Activity Level and Chronic Low back Pain Symptoms

Ellingson, Laura D.1; Lansing, Jeni E.2; Perez, Maria L.2; DeShaw, Kathryn J.3; Meyer, Jacob D.2; Welk, Gregory J.2

Author Information
Translational Journal of the ACSM: Spring 2022 - Volume 7 - Issue 2 - e000192
doi: 10.1249/TJX.0000000000000192
  • Free



Low back pain is the leading cause of disability worldwide (1). Approximately 20% of these cases become chronic and are accompanied by varying levels of fatigue, cognitive impairment, sleep problems, and feelings of depression and anxiety (2–4). Pharmaceutical and surgical treatments are commonly used to manage symptoms of chronic low back pain (CLBP). However, evidence suggests these methods are largely ineffective and costly, and can have significant harmful side effects (5,6). Consequently, behavioral treatments are recommended for pain management, with research showing they have the potential to be beneficial and cost-effective (5).

Exercise, a common behavioral treatment, can be effective for both prevention and treatment of CLBP (7,8). However, similar to pain-free adults, adherence to regular exercise regimens is poor, with nonadherence rates ranging from 50% to 70% (9). There is also evidence that CLBP patients may benefit from increasing less structured physical activity (PA) like walking and/or reducing sedentary behavior (SB), and that these behaviors may be perceived as more feasible and thus lead to better adoption and adherence than higher intensity and/or gym-based exercise programs (10). For example, several reviews suggest walking may be effective for managing CLBP, although the quality of this evidence was low to moderate and walking was not universally superior to usual care (11,12). Limited evidence also shows that high self-reported sitting time may be a risk factor for development of CLBP (13) and that reducing time in SB may be effective for improving disability in this population (14). While promising, evidence regarding the effectiveness of targeting changes in these behaviors for improving CLBP remains sparse.

When promoting PA-related behavior change, interventions utilizing health coaching techniques like motivational interviewing (MI) are beneficial (15). Health coaching based in MI has also been shown to positively impact SB (16). Our group has previously demonstrated the feasibility and utility of facilitated health coaching applications that integrate the use of activity monitors to promote behavior change (17). More recent work from our group documented the utility of our training model for health coaches (18) and the potential of MI to enhance habit formation related to use of wearable technology (19). Our previous work in pain-free adults suggested that those who were less active at baseline may benefit the most from the intervention with larger improvements in both PA and SB (19). Controlled trials are needed to determine if individuals with CLBP, who often have low activity levels, could benefit from this type of intervention.

To date, use of MI has been minimally examined in CLBP, with equivocal results. For example, a randomized controlled trial found that the addition of MI improved self-efficacy and compliance with physical therapy to a greater extent than physical therapy alone (20). Conversely, in a larger trial, MI was not more effective in increasing self-efficacy or PA compared with simply providing PA guidelines in low back pain patients (21). More recently, the combination of MI and mHealth was examined in a smaller pilot study of CLBP (n = 68) with promising results showing that the intervention was feasible and may reduce care seeking (22). Thus, additional studies are needed to specifically examine the utility of MI for activity-related behavior change in this population.

The purpose of this randomized controlled, pilot study was to examine the effectiveness of targeting increases in PA and decreases in SB for improving CLBP symptoms. Specifically, this study examined how providing a wearable activity monitor alone or in combination with a low dose of health coaching based on MI influenced PA and SB, CLBP-related symptoms, and Patient Global Impression of Change (PGIC) in adults with CLBP. A secondary purpose was to identify behaviors associated with symptom changes to inform targets for future interventions. Hypotheses were 1) participants who receive an activity monitor in combination with health coaching would have greater improvements in symptoms, PGIC, mood, quality of life (QOL), PA, and SB compared with those receiving a tracker alone; 2) both intervention groups would have greater improvements compared with a control condition; and 3) improvements in pain symptoms and PGIC would be related to PA and SB changes across the intervention.



Participants were recruited from the campus community using electronic mailing lists. Inclusion criteria were endorsement of the statement “I have had pain in my lower back for at least 3 months,” 24–65 yr of age, regular access to a computer or smartphone, and willing to wear an activity monitor for 3 months. Exclusion criteria were current use of an activity monitor, presence of health conditions that prevented safe engagement in PA, taking blood pressure medication, or current participation in a structured exercise program.


The present study was a three-arm randomized controlled trial of the effectiveness of a 12-wk intervention targeting PA and SB for symptom improvement in adults with CLBP. All procedures were approved by the local institutional review board (Approval No. 15-748). The protocol consisted of three study visits and two phone calls over a 3-month intervention period. Data were collected between February 2016 and April 2017.

During the first visit, participants completed informed consent and questionnaires including an eligibility questionnaire, demographic questionnaire, the PA Readiness Questionnaire (23), and the Chronic Low Back Pain Minimal Dataset (24). Next, height (stadiometer), weight and body composition (bioelectrical impedance; InBody 720; InBody, Cerritos, CA), and blood pressure and heart rate (Omron HEM712C; Omron Healthcare, Inc., Hoffman Estates, IL) were measured. Questionnaires regarding mood state, pain symptoms, and health-related QOL were then administered electronically. Afterward, participants were given ActiGraph and activPAL activity monitors to wear for 1 wk to assess baseline PA and SB. Participants were encouraged to behave as usual during this baseline assessment, and monitoring was rescheduled if they indicated the subsequent week would be atypical (e.g., travel).

One week later, participants returned and were randomly assigned to receive a wearable activity monitor alone (WAM), or in combination with health coaching based on principles of MI (WAM + HC), or to a wait-list control (WLC) condition. Those assigned to WAM and WAM + HC received a Fitbit Charge™ (Fitbit Inc., San Francisco, CA) wrist-worn activity monitor and were instructed to use it to help them move more and sit less for the duration of the intervention. Participants were assisted with monitor and account setup, and features were demonstrated (e.g., logging food, dashboard displays). Fitbit use was monitored throughout the 3-month intervention via Fitabase Software (Small Steps, LLC; San Diego, CA) as a measure of implementation, including metrics for Fitbit wear time and steps per day.

In addition, participants assigned to the WAM + HC group met with a trained health coach for approximately 30 min to discuss their self-identified goals regarding increasing PA and decreasing SB. Health coaches received 6 wk of training in MI before the start of the study and demonstrated proficiency through recorded test conversations. The coaching sessions employed established principles of MI (25) and were aimed at helping the participant establish feasible goals.

During weeks 4 and 8, participants in WAM and WAM + HC were contacted via phone for a check-in. Calls for the WAM group were aimed at encouraging ongoing use of the monitor and answering questions as needed, whereas calls for those in WAM + HC included additional health coaching. During these calls, coaches revisited each participant’s self-determined goals for changing PA and SB and promoted reflection about their motivation(s) for making these changes; coaches had access to participants’ Fitbit data to inform these communications. Calls for WAM + HC lasted ~15 min longer calls for WAM to allow time to review, reflect, and revise participant goals. Previous research by our team revealed no differences in outcomes between in-person and telephonic health coaching (17). Participants in the WLC condition received no contact from study personnel during the intervention period and had the opportunity to receive a Fitbit and coaching after follow-up data had been collected.

During week 12, participants were again issued research-grade PA monitors to assess behavior change. Participants then returned to the laboratory for their final study visit, repeating all baseline assessments and completing the PGIC.

Primary Outcomes

Pain symptoms

Consistent with IMMPACT recommendations for interpreting clinical importance of treatment outcomes, two separate measures were used to examine primary outcomes, a measure specific to pain symptoms and a measure of overall improvement (26). Pain symptoms were assessed with the Short Form McGill Pain Questionnaire (SF-MPQ). This is a psychometrically strong instrument (α = 0.76–0.78) that is commonly used to assess the intensity of pain symptoms (27,28). For each of 15 descriptors, participants rated their pain over the past week from 0 (no pain) to 3 (severe). A sum of scores for all descriptors was then calculated for a total pain symptom score (MPQ-Total), with higher scores indicating greater pain-related symptoms.

The PGIC was used to examine overall improvement since beginning the intervention (29). This single-item measure asks respondents to integrate different aspects of their responses to the intervention including changes in activity limitations, symptoms, emotions, and overall QOL. It uses a 7-point rating scale with anchors of “a great deal better, and a considerable improvement that has made all the difference,” and “no change (or the condition has got worse).” Higher scores indicate greater perceived improvement.

PA and SB

Before and after intervention, PA and SB were assessed objectively using ActiGraph GT3X+ (ActiGraph, LLC, Fort Walton Beach, FL) and activPAL3 (PA Technologies, Glasgow, United Kingdom) activity monitors. Participants wore the ActiGraph on their hip using an elastic belt and attached the activPAL to the midline of the thigh using hypoallergenic tape. Use of two monitors allowed for integration of postural information and acceleration at the hip to more accurately assess the full range of human movement using the method described briefly below. Participants were instructed to wear both monitors during all waking hours, except when bathing or swimming. They were asked to remove the ActiGraph during sleep and given the option wear the activPAL overnight, as it can aggravate skin when taken on and off repeatedly. Participants were also given a log sheet and asked to record time on/off for each monitor and sleep and wake times. Minimal inclusion criteria for these data were 4 d of at least 10 h of wear, including one weekend day. Wear time was calculated as time with both monitors on simultaneously. Monitor wear time averaged 14.48 ± 1.25 h for WAM + HC, 14.10 ± 1.56 h for WAM, and 14.26 ± 1.52 h for WLC.

Initial processing of monitor data was completed using proprietary software (ActiLife 6.13.3 and PAL Analysis 7.2.32) with outputs at the 1-s epoch. These data were then integrated using their timestamps and further analyzed using the validated Sojourns Including Posture method (30), which utilizes postural information from the activPAL to inform a decision tree in combination with a neural network approach to assign MET values to bouts of activity and inactivity. Sojourns Including Posture output was used to calculate behavior-related outcomes including daily average minutes of light PA (LPA), moderate-to-vigorous PA (MVPA), and SB (total and in 60-min bouts).

Secondary Outcomes

To examine the influence of the intervention on mood and overall QOL, the short form of the Profile of Mood States (POMS-SF) and the SF-36v2 were administered before and after intervention. The POMS-SF assesses mood over the past week, and a total mood disturbance score was computed with higher values indicating worse mood. The SF-36v2 assesses QOL over the past 4 wk. From this questionnaire, physical and mental health composite scores were calculated using standard scoring procedures, with higher scores indicating better QOL (31). A score for overall bodily pain was also calculated, with higher scores indicative of better outcomes (i.e., lower pain and less pain interference with normal activities).

As a measure of intervention implementation, use of Fitbits was objectively monitored through the Fitabase software noted previously. Data from Fitabase were downloaded in 1-min epochs and summarized to determine frequency for wearing the Fitbit as well as accumulation of steps per day. A valid monitor wear day was defined as accumulating steps during at least 10 h·d−1.

Statistical Analyses

All analyses were performed using SPSS 27.0 (IBM, Armonk, NY). Descriptive statistics were calculated for demographic factors, and baseline values for pain symptoms, mood, QOL, and the PA and SB variables. Data from Fitabase were also examined descriptively and compared numerically across time and between groups to assess use of Fitbits during the study. For the primary analyses, an intention-to-treat approach was taken including all those with baseline data. Specifically, baseline observations were carried forward when follow-up data were missing. For the secondary analyses, a completers-only analysis approach was used, including only those with both baseline and 12-wk data. To determine the influence of the intervention on primary outcomes (symptoms, PGIC, MVPA, and SB) and secondary outcomes (mood and QOL), within- and between-groups changes were descriptively compared using effect size calculations (Hedges’ g) (32). To address our first hypothesis, a series of multiple linear regression analyses were performed for the primary outcomes with continuous dependent variables (SF-MPQ, minutes of LPA, minutes of MVPA, and total and prolonged SB). Independent variables were group, age, sex, body mass index (BMI), and baseline value of the selected outcome variable. To examine our second hypothesis regarding influence of the intervention on PGIC (an ordinal outcome variable), the PLUM procedure (33) was employed using the same independent variables with the exception of the baseline value of PGIC (no baseline value for this measure). Multicollinearity was examined with all variance inflation factors ≤3, so all independent variables were retained.

Next, to examine behavioral predictors of improvements in pain symptoms and PCIG, two multiple regression analyses were conducted with group, baseline pain symptoms, changes in total and prolonged (in bouts of 60+ min) SB and changes in LPA, and MVPA as independent variables and change in pain symptoms (MPQ-Total Post–Pre) and PGIC as the dependent variables. Again, for PGIC, the PLUM procedure was employed. Multicollinearity was similarly examined, and all independent variables were retained.

All metrics for PA and SB variables were average minutes per day. Effect sizes of 0.20, 0.50, and 0.80 were interpreted as indicating small, medium, and large effects, respectively (32). Because the primary purpose of this pilot study was to inform future trials, the α level was set to 0.05 for all analyses.


Interested individuals were screened for eligibility via phone (n = 142). Those eligible (n = 71) were randomized into the study. Sixty-four of these completed baseline assessments and 56 participants completed both baseline and follow-up (Fig. 1). Participants were primarily White, college-educated, and overweight (Table 1).

Figure 1:
CONSORT diagram.
TABLE 1 - Participant Characteristics.
Full Sample (n = 64) WAM + HC (n = 23) WAM (n = 19) WLC (n = 22) Group Differences (P Value)
Age (yr) 44.5 (10.2) 46.8 (10.5) 41.4 (9.6) 44.8 (10.1) P > 0.05
Sex (% male) 47.7 43.5 55 45.5 P > 0.05
Education (% with college degree) 93.8 91.3 95 95.5 P > 0.05
Marital status (% married) 72.3 73.9 75 68.2 P > 0.05
Income (% >$50,000/yr) 78.5 78.3 80 77.3 P > 0.05
Race/ethnicity (% White) 81.5 78.3 90 77.3 P > 0.05
Employment status (% employed full-time) 92.3 82.6 95 100 P > 0.05
Heart rate (bpm) 70.13 (8.53) 68.56 (8.74) 73.63 (9.54) 68.60 (6.52) P > 0.05
Systolic blood pressure (mm Hg) 122.29 (14.34) 122.74 (14.11) 119.85 (13.34) 124.05 (15.76) P > 0.05
Diastolic blood pressure (mm Hg) 74.92 (10.48) 74.65 (11.13) 73.50 (7.52) 76.50 (12.22) P > 0.05
Body mass index, kg·m−2 29.50 (5.65) 29.63 (6.38) 30.36 (5.00) 28.60 (5.50) P > 0.05
Values are mean (SD) for continuous variables or n (%) for categorical variables.

With respect to intervention delivery, 81% of all intervention calls were completed, with 87% of participants in WAM + HC receiving both intervention calls. Fitbit wear was similar between intervention groups with WAM + HC averaging 5.4 ± 1.6 d of valid wear and WAM averaging 5.3 ± 1.7 d across the 12 wk. Both groups started slightly higher averaging >6 d of wear during week 1 and declining to 4.6 ± 2.9 d in WAM + HC and 3.6 ± 3.1 for WAM by week 12. Across the intervention, steps per day averaged 9990 ± 4500 for WAM + HC, whereas WAM averaged slightly less at 9212 ± 3137 (g = 0.20). Furthermore, participants in WAM + HC increased from an average of 8844 ± 3786 during week 1 to a peak of 10,975 ± 4767 during week 6 and maintained approximately that level ±1000 steps per day for the remaining 6 wk. In contrast to this, participants in WAM averaged 8763 ± 2956 during week 1, peaked at an average of 9764 ± 3554 during week 9, and returned to a level slightly below baseline by week 12 at 8515 ± 2460.

Average changes in pain, mood, QOL, and the behavioral variables of LPA, MVPA, and SB by group, along with Hedges’ g comparisons within and between groups are presented in Table 2. In general, there were improvements ranging from moderate to large for pain symptoms for WAM + HC and WAM (g = 0.59–0.80) and minimal changes in WLC (g = <0.001–0.05). Improvements in mood were small and similar across groups (g = 0.20–0.39). Physical health–related QOL showed a small improvement in WAM + HC (g = 0.32) and a moderate improvement in WAM (g = 0.56), whereas those in WLC showed a slight worsening (g = 0.17). Mental health–related QOL improved similarly in all groups over the 12 wk (WAM + HC, g = 1.45; WAM, g = 0.99; WLC, g = 1.42). Lastly, PGIC after intervention showed larger values in both WAM + HC and WAM (g = 1.06 and 1.33, respectively) when directly compared with WLC.

TABLE 2 - Baseline and Postintervention Values as well as Average Changes in Pain, Mood, QOL, and the Behavioral Variables of LPA, MVPA, and SB by Group, Along with Hedges’ g Comparisons within and between Groups.
WAM + HC WAM WLC Within-Group ES Between-Group ES
Baseline 12 wk Change Baseline 12 wk Change Baseline 12 wk Change WAM + HC WAM WLC WAM + HC vs WLC WAM vs WLC
MPQ-Total 22.55 (5.68) 19.73 (3.88) −2.81 (5.21) 22.60 (4.54) 19.50 (6.02) −3.10 (4.89) 23.29 (6.51) 23.23 (8.19) 0.09 (4.08) 0.59 0.59 0.00 0.62 0.71
POMS–TMD 115.68 (11.93) 111.57 (12.44) −3.91 (9.44) 119.90 (13.11) 114.00 (16.44) −5.90 (7.99) 114.10 (14.47) 111.09 (14.90) −4.47 (10.93) 0.33 0.39 0.20 0.05 0.15
SF-36 Physical Health 47.09 (6.11) 49.00 (5.60) 1.61 (6.29) 45.04 (6.65) 48.79 (6.71) 3.75 (5.12) 47.59 (6.83) 46.33 (8.09) −0.86 (7.49) 0.32 0.56 −0.17 0.36 0.73
SF-36 Mental Health 37.01 (4.91) 46.89 (8.74) 9.36 (9.45) 38.13 (5.41) 47.26 (12.96) 9.13 (10.02) 39.21 (4.22) 49.37 (10.03) 10.56 (11.54) 1.45 0.99 1.42 0.11 0.13
SF-36 Bodily Pain 59.78 (13.78) 69.66 (13.85) 9.17 (17.98) 55.50 (16.7) 70.25 (19.99) 14.75 (15.34) 63.69 (18.33) 64.66 (15.95) 2.02 (17.93) 0.71 0.80 0.05 0.39 0.76
PGIC NA 4.56 (1.52) NA NA 4.33 (1.88) NA NA 2.55 (1.47) NA NA NA NA 1.40 1.06
MVPA/day 74.01 (38.24) 82.72 (41.14) 8.63 (18.33) 68.49 (36.92) 76.52 (34.96) 8.03 (25.86) v73.69 (28.96) 69.88 (26.45) −5.69 (20.30) 0.22 0.22 −0.13 0.74 0.59
LPA/day 253.63 (92.31) 267.03 (102.29) 13.41 (83.29) 241.42 (102.67) 242.31 (114.43) 0.88 (57.59) 261.25 (88.45) 226.34 (80.64) −35.45 (80.72) 0.14 0.01 −0.41 0.60 0.52
Sedentary time/day 539.15 (85.79) 524.79 (86.67) −14.37 (76.35) 536.26 (126.32) 529.20 (142.92) −7.06 (92.62) 513.72 (96.06) 562.81 (79.14) 49.08 (98.17) 0.17 0.05 −0.56 0.73 0.59
Sedentary time in 60+ min bouts/day 152.94 (94.07) 140.92 (82.05) −12.02 (61.67) 138.50 (102.41) 133.75 (92.52) −4.76 (67.56) 135.28 (60.72) 157.35 (69.27) 22.07 (70.86) 0.14 0.05 −0.34 0.51 0.39
All values are mean (SD). Change scores are 12-wk minus baseline scores. Effect sizes (ES; Hedges g) compare within-group changes and also changes across the intervention period for WAM + HC or WAM compared with WLC, as indicated with negative cores indicating a worsening of the measure.

With respect to average daily PA and SB, WAM + HC showed small improvements across all metrics with increases in PA and decreases in both total and prolonged SB (g = 0.14–0.22). WAM showed a small improvement MVPA (g = 0.22) and minimal change in LPA (g = 0.01) or SB (g = 0.05). Conversely, WLC had, on average, small decreases in MVPA and LPA (g = 0.13–0.41) and moderate increases in both total and prolonged SB (g = 0.51–0.73).

For the primary regression examining associations with pain symptoms, group (WAM + HC, P = 0.02; WAM, P = 0.02) and baseline MPQ-Total (P < 0.001) were significantly associated with follow-up MPQ-Total. For the PGIC, the odds of WAM + HC reporting a greater overall change were 10.18 (95% confidence interval [CI], 2.76–37.4) times the odds of WLC (P < 0.001). Similarly, the odds of WAM reporting a greater overall change were 7.35 (95% CI, 1.98–27.24) times the odds of WLC (P = 0.003). BMI was also significantly and positively associated with PGIC (P = 0.004). Full model results are presented in Table 3.

TABLE 3 - Results from the Regressions Examining Predictors of Pain-Related Symptoms, PGIC, and Active Behavior and SB Follow-up.
Independent Variables Dependent Variables, All Measured at 12 wk
MPQ-Total PGIC (n = 56) LPA MVPA Total Sedentary Time Prolonged Sedentary Time
ß P ORs 95% CI P ß P ß P ß P ß P
Group WAM + HC −0.25* 0.02 10.175* 2.76–37.4 <0.001 0.23* 0.04 0.21* 0.03 −0.24* 0.03 −0.17 0.13
WAM −0.28* 0.01 7.35* 1.98–27.24 0.003 0.13 0.24 0.17 0.08 −0.15 0.19 −0.11 0.31
Age −0.09 0.40 0.980 0.92–1.026 0.33 −0.09 0.39 −0.08 0.35 0.25* 0.02 0.17 0.10
Sex 0.17 0.09 0.86 0.31–1.2 0.77 0.16 0.11 0.03 0.69 −0.24* 0.02 −0.09 0.34
BMI 0.09 0.34 1.15* 1.046–1.265 0.004 0.08 0.38 −0.06 0.49 −0.21* 0.04 −0.03 0.77
Baseline level of DV 0.64* <0.001 NA NA NA 0.70* <0.001 0.83* <0.001 0.63* <0.001 0.68* <0.001
Overall model R 2 = 0.50 <0.001 R 2 = 24.58 <0.001 R 2 = 0.49 <0.001 R 2 = 0.64 <0.001 R 2 = 0.47 <0.001 R 2 = 0.45 <0.001
Reference groups are WLC for group and female for sex. With the exception of PGIC, all analyses included the full intent-to-treat sample (n = 64).
*Significant, P < 0.05.
ß, standardized beta coefficient; NA, not applicable; OR, odds ratio.

For the regressions examining associations with PA and SB at follow-up, baseline values of the dependent variable were significantly associated with postvalues in each model (all P < 0.001). Furthermore, receiving the combined intervention (WAM + HC) was significantly associated with LPA (P = 0.04), MVPA (P = 0.04), and total SB (P = 0.03) at follow-up. Age (P = 0.02), sex (P = 0.02), and BMI (P = 0.04) were also significantly associated with total SB at follow-up. Full model results are presented in Table 3.

For the secondary regressions examining whether behavior changes were associated with pain symptom changes, significant associations were found for group (WAM + HC, P = 0.04; WAM, P = 0.02), baseline MPQ-Total (P = 0.04), and changes in prolonged SB (P = 0.01). For PGIC, the odds of participants in WAM + HC reporting a greater overall change were 8.11 (95% CI, 2.07–31.84) times that of the WLC group. Similarly, the odds of participants in WAM reporting a greater overall change were 7.97 (95% CI, 2.09–30.29) times that of the WLC group. Full model results for are shown in Table 4.

TABLE 4 - Results from the Regressions Examining Predictors of Change in Pain Symptoms and PGIC over the Intervention.
Independent Variables Dependent Variables
Change in SF-MPQ PGIC
Standard ß P Odds Ratio 95% CI (Lower, Upper) P
 WAM + HC vs WLC −0.32* 0.04 8.11* 2.07, 31.84 0.003
 WAM vs WLC −0.36* 0.02 7.97* 2.09, 30.29 0.002
Baseline SF-MPQ −0.26* 0.04 NA NA NA
Change in total sedentary time −0.02 0.91 1.001 0.99, 1.01 0.91
Change in prolonged sedentary time 0.48* 0.01 0.99 0.99, 1.01 0.28
Change in LPA 0.32 0.09 1.01 0.99, 1.01 0.42
Change in MVPA −0.04 0.79 0.98 0.96, 1.01 0.14
Overall model R 2 = 0.44* P = 0.007 χ 2 = 18.99* P = 0.006
These analyses included only participants who completed follow-up assessments (n = 56).
*Significant, P < 0.05.
ß, standardized beta coefficient; NA, not applicable.


This pilot study was intended to examine how targeting changes in activity-related behaviors affected pain symptoms and perception of condition improvement in individuals with CLBP. In line with our hypotheses, results demonstrated that providing a wearable activity monitor in combination with a minimal dose of health coaching based on MI was effective for significantly increasing MVPA, LPA, and total SB as well as reducing pain symptoms. Perhaps more importantly, PGIC scores demonstrated that the intervention was perceived as making a noticeable difference in participants’ lives (a score of at least 5 on the 0–7 scale) for more than 63% of participants in this group with zero participants reporting no change and/or worsening.

Those receiving an activity monitor alone also showed a significant improvement in pain symptoms and perceived positive changes across the intervention with over 58% of the sample reporting a noticeable difference in how they felt. However, on average, there was only a small increase in MVPA and virtually no changes in either LPA or SB. This suggests that use of activity monitors alone likely has benefits for individuals with CLBP, but the addition of HC may lead to broader improvements in both active behavior and SB that could further influence health outcomes associated with CLBP, as previous research suggests that changes in SB are independently associated with symptoms of chronic pain (14). This hypothesized effect (supporting a combined intervention approach) aligns with the findings of a recent meta-analysis, in which SB decreased significantly when HC was used with self-monitoring and that this combined approach was more effective than other approaches, like education (34). Further research with longer-term follow-ups will be needed to determine whether this is the case.

Although the effects on LPA and SBs were less pronounced for WAM than for those also receiving HC, it is notable that pain outcomes for individuals who received an activity monitor alone were superior to those who received no intervention over the 12-wk intervention period. For WLC, pain symptoms remained at baseline levels, and both PA and SB worsened. For example, minutes of MVPA decreased by an average of 5 min·d−1, LPA decreased by an average of more than 35 min·d−1, and total SB increased by nearly 50 min·d−1. Thus, without intervention, individuals with CLBP may reduce PA levels and/or increase SB, which could ultimately negatively influence their condition. Simply prescribing a wearable activity monitor may be a cost-effective and acceptable strategy that could help blunt the negative progression of CLBP, at least in the short term.

A secondary purpose of this project was to identify behaviors that were associated with symptom changes to inform behavioral targets for future intervention trials. Many different activity behaviors (e.g., increasing MVPA, reducing total SB, breaking up prolonged SB) have been found helpful in managing CLBP (7,8,11,12,14), although few studies have explored which type(s) result in better pain outcomes. Interestingly, our results showed that change in time spent in prolonged SB was significantly associated with pain symptom change, whereas changes in total SB, MVPA, and LPA were not. This finding is in line with a small body of existing evidence showing similar relationships between prolonged SB and back pain. For example, Thorp and colleagues (35) examined changes in fatigue, musculoskeletal discomfort, and work productivity after reducing SB in individuals with back pain and reported a 31.8% reduction in low back pain in individuals engaging in intermittent breaks from prolonged SB compared with a standardized sitting control group. In addition, a 6-month randomized trial using a multicomponent intervention to break up prolonged SB in adults with CLBP found significant reductions of SB per day and perceptions of disability compared with a control group (14). Thus, reducing prolonged SB may be an effective approach for reducing chronic pain symptoms, although additional research is needed to further elucidate this hypothesis.

There were several notable strengths and limitations of this study. Strengths include the design (i.e., randomized control trial), the conservative, intention-to-treat approach taken for the primary analyses, use of both control and comparator groups (i.e., WLC and WAM, respectively), multiple pain symptom outcomes (i.e., MPQ-Total, PGIC), use of a low dose of health coaching with potential for translatability, and inclusion of theoretical constructs relevant to behavior change (e.g., MI, self-monitoring). The primary limitation of this study is the small sample size, potentially resulting in nonsignificant findings due to a lack of power. Additional limitations are no assessment of MI fidelity or changes in motivation, and a moderately high attrition rate, although dropout was similar across groups. Furthermore, our finding that changes in MVPA were not associated with pain improvements was unexpected and may have resulted from relatively high baseline activity levels or low power. The accumulation of additional minutes of MVPA per day (~10 min) may not have made a meaningful difference in this sample, although it may in those with lower baseline activity levels. Nonetheless, this pilot study provides important data to guide the development of a larger trial, and recruiting individuals more representative of the CLBP patient profile (e.g., less active), assessing fidelity and/or changes in theoretical constructs (e.g., motivation), and employing additional retention strategies would strengthen the study.


Results from this pilot randomized controlled trial demonstrate that use of wearable activity monitors in combination with a minimal dose of HC may be an effective strategy for improving PA behaviors and pain symptoms in patients with CLBP. Reductions in prolonged SB seem to be associated with symptom improvements and thus may be an effective target for behavior change efforts to reduce chronic pain symptoms. Future research including larger sample sizes, targeting prolonged SB, and examining outcomes related to pain and associated symptoms will be important for identifying optimal translational treatment strategies for CLBP.

The results of the present study do not constitute endorsement by the American College of Sports Medicine.

There are no conflicts of interest or funding associated with this project.


1. GBD 2016 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 328 diseases and injuries for 195 countries, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet. 2017;390(10100):1211–59.
2. Elliott AM, Smith BH, Penny KI, et al. The epidemiology of chronic pain in the community. Lancet. 1999;354(9186):1248–52.
3. Kelly GA, Blake C, Power CK, et al. The association between chronic low back pain and sleep: a systematic review. Clin J Pain. 2011;27(2):169–81.
4. Sagheer MA, Khan MF, Sharif S. Association between chronic low back pain, anxiety and depression in patients at a tertiary care centre. J Pak Med Assoc. 2013;63(6):688–90.
5. Foster NE, Anema JR, Cherkin D, et al. Prevention and treatment of low back pain: evidence, challenges, and promising directions. Lancet. 2018;391(10137):2368–83.
6. Qaseem A, Wilt TJ, McLean RM, et al. Noninvasive treatments for acute, subacute, and chronic low Back pain: a clinical practice guideline from the American College of Physicians. Ann Intern Med. 2017;166(7):514–30.
7. Steffens D, Maher CG, Pereira LSM, et al. Prevention of low back pain: a systematic review and meta-analysis. JAMA Intern Med. 2016;176(2):199–208.
8. Miyamoto GC, Lin C-WC, Cabral CMN, et al. Cost-effectiveness of exercise therapy in the treatment of non-specific neck pain and low back pain: a systematic review with meta-analysis. Br J Sports Med. 2019;53(3):172–81.
9. Beinart NA, Goodchild CE, Weinman JA, et al. Individual and intervention-related factors associated with adherence to home exercise in chronic low back pain: a systematic review. Spine J. 2013;13(12):1940–50.
10. Lansing JE, Ellingson LD, Goode K, Meyer JD. Comparison of self-efficacy for reducing sedentary time to self-efficacy for increasing physical activity. Ann Behav Med. 2021;55(S1):S113.
11. Lawford BJ, Walters J, Ferrar K. Does walking improve disability status, function, or quality of life in adults with chronic low back pain? A systematic review. Clin Rehabil. 2016;30(6):523–36.
12. Sitthipornvorakul E, Klinsophon T, Sihawong R, Janwantanakul P. The effects of walking intervention in patients with chronic low back pain: a meta-analysis of randomized controlled trials. Musculoskelet Sci Pract. 2018;34:38–46.
13. Park S-M, Kim H-J, Jeong H, et al. Longer sitting time and low physical activity are closely associated with chronic low back pain in population over 50 years of age: a cross-sectional study using the sixth Korea National Health and Nutrition Examination Survey. Spine J. 2018;18(11):2051–8.
14. Barone Gibbs B, Hergenroeder AL, Perdomo SJ, et al. Reducing sedentary behaviour to decrease chronic low back pain: the stand back randomised trial. Occup Environ Med. 2018;75(5):321–7.
15. Kunstler BE, Cook JL, Freene N, et al. Physiotherapist-led physical activity interventions are efficacious at increasing physical activity levels: a systematic review and meta-analysis. Clin J Sport Med. 2018;28(3):304–15.
16. Alothman S, Alenazi AM, Alshehri MM, et al. Sedentary behavior counseling intervention in aging people with type 2 diabetes: a feasibility study. Clin Med Insights Endocrinol Diabetes. 2021;14:11795514211040540.
17. Bus K, Peyer KL, Bai Y, et al. Comparison of in-person and online motivational interviewing-based health coaching. Health Promot Pract. 2018;19:513–21.
18. DeShaw K. Methods and evaluation of a health coach training practicum experience for healthy lifestyle behavior change. Grad Theses Diss [Internet]. 2019. Available from:
19. Ellingson LD, Lansing JE, DeShaw KJ, et al. Evaluating motivational interviewing and habit formation to enhance the effect of activity trackers on healthy adults’ activity levels: randomized intervention. JMIR Mhealth Uhealth. 2019;7(2):e10988.
20. Vong SK, Cheing GL, Chan F, et al. Motivational enhancement therapy in addition to physical therapy improves motivational factors and treatment outcomes in people with low back pain: a randomized controlled trial. Arch Phys Med Rehabil. 2011;92(2):176–83.
21. Leonhardt C, Keller S, Chenot JF, et al. TTM-based motivational counselling does not increase physical activity of low back pain patients in a primary care setting—a cluster-randomized controlled trial. Patient Educ Couns. 2008;70(1):50–60.
22. Amorim AB, Pappas E, Simic M, et al. Integrating Mobile-health, health coaching, and physical activity to reduce the burden of chronic low back pain trial (IMPACT): a pilot randomised controlled trial. BMC Musculoskelet Disord. 2019;20(1):71.
23. Adams R. Revised Physical Activity Readiness Questionnaire. Can Fam Physician. 1999;45:992–1005.
24. Deyo RA, Dworkin SF, Amtmann D, et al. Report of the NIH Task Force on Research Standards for Chronic Low Back Pain. Phys Ther. 2015;95(2):e1–18.
25. Rollnick S, Miller WR. What is motivational interviewing?Behav Cogn Psychother. 1995;23(4):325–34.
26. Dworkin RH, Turk DC, Wyrwich KW, et al. Interpreting the clinical importance of treatment outcomes in chronic pain clinical trials: IMMPACT recommendations. J Pain. 2008;9(2):105–21.
27. Melzack R. The short-form McGill Pain Questionnaire. Pain. 1987;30(2):191–7.
28. Wright KD, Asmundson GJ, McCreary DR. Factorial validity of the Short-Form McGill Pain Questionnaire (SF-MPQ). Eur J Pain. 2001;5(3):279–84.
29. Hurst H, Bolton J. Assessing the clinical significance of change scores recorded on subjective outcome measures. J Manipulative Physiol Ther. 2004;27(1):26–35.
30. Ellingson LD, Schwabacher IJ, Kim Y, et al. Validity of an integrative method for processing physical activity data. Med Sci Sports Exerc. 2016;48(8):1629–38.
31. Ware JE, Kosinski M, Bjorner JB, et al. Users Manual for the SF-36v2 Health Survey. 2nd ed. Lincoln (RI): QualityMetric Incorporated; 2007.
32. Hedges LV. Distribution theory for Glass’s estimator of effect size and related estimators. J Educ Stat. 1981;6(2):107–28.
33. Decarlo LT. Using the PLUM procedure of SPSS to fit unequal variance and generalized signal detection models. Behav Res Methods Instrum Comput. 2003;35(1):49–56.
34. Nieste I, Franssen WMA, Spaas J, et al. Lifestyle interventions to reduce sedentary behaviour in clinical populations: a systematic review and meta-analysis of different strategies and effects on cardiometabolic health. Prev Med. 2021;148:106593.
35. Thorp AA, Kingwell BA, Owen N, Dunstan DW. Breaking up workplace sitting time with intermittent standing bouts improves fatigue and musculoskeletal discomfort in overweight/obese office workers. Occup Environ Med. 2014;71(11):765–71.
Copyright © 2022 by the American College of Sports Medicine