Article Text

Download PDFPDF

Combined education and patient-led goal setting intervention reduced chronic low back pain disability and intensity at 12 months: a randomised controlled trial
  1. Tania Gardner1,
  2. Kathryn Refshauge2,
  3. James McAuley3,4,
  4. Markus Hübscher3,
  5. Stephen Goodall5,
  6. Lorraine Smith1
  1. 1 Department of Pharmacy, University of Sydney, Camperdown, New South Wales, Australia
  2. 2 Department of Health Sciences, University of Sydney, Sydney, New South Wales, Australia
  3. 3 Neuroscience Research Australia, Sydney, New South Wales, Australia
  4. 4 University of New South Wales, Sydney, Australia
  5. 5 Centre for Health Economics Research & Evaluation, University of Technology, Sydney, New South Wales, Australia
  1. Correspondence to Dr Tania Gardner, PharmacyUniversity of Sydney, Camperdown, NSW 2006, Australia; tania.gardner{at}


Background One model of care that has not been tested for chronic low back pain (LBP) is patient-led goal setting. We aimed to compare the clinical effectiveness and healthcare use of a patient-led goal setting approach (intervention) with simple advice to exercise (control) over 12 months.

Methods An assessor-blinded randomised controlled trial. Intervention was education combined with patient-led goal setting compared with a control group receiving a standardised exercise programme. The primary outcomes were back pain disability and pain intensity. Secondary outcomes were quality of life, kinesiophobia, self-efficacy, depression, anxiety and stress. Outcomes and healthcare use were assessed immediately post-treatment (2 months) and after 4 and 12 months. Analysis was by intention to treat.

Results Seventy-five patients were randomly assigned to either the intervention (n=37) or the control (n=38) group. Using linear mixed model analyses, adjusted mean changes in primary outcomes of disability and pain intensity were greater in the intervention group than in the control group (disability post-treatment: p<0.05). These differences were clinically meaningful. Mean differences in all secondary measures were greater in the intervention group than in the control group (p<0.05). There was no difference in healthcare use between groups over 12 months.

Conclusion A patient-led goal setting intervention was significantly more effective than advice to exercise for improving outcomes in disability, pain intensity, quality of life, self-efficacy and kinesiophobia in chronic LBP. These improvements were maintained at 12 months. Smaller effects were seen in measures of depression, anxiety and stress.

Trial registration number ACTRN12614000830695.

  • chronic
  • lower back
  • rehabilitation

Statistics from


The impact of chronic low back pain (CLBP) worldwide is substantial in population-level prevalence, costs, individual quality of life and disability.1 Treatment for CLBP often fails to improve patient relevant outcomes such as disability or pain intensity. Many studies show a small to moderate effect at best, minimal sustainability and poor patient adherence to treatment regimens.2 Reasons for this include lack of uptake by clinicians of a biopsychosocial framework to treatment3 and inattention to the everyday activities and needs of patients.4 Although clinical guidelines endorse a biopsychosocial and patient centred approach in CLBP, healthcare practitioners tend not to follow these guidelines.3 Healthcare professionals lack the confidence and skills to address the complex psychosocial issues associated with chronic pain5 and deliver a clinician-centred model of patient care.6

CLBP is characterised by a complex interaction between pain, function and biopsychosocial factors such as patient motivation and confidence.7 A treatment approach aimed at self-management support should be appropriate for CLBP.8 Most self-management programmes include a goal setting component,4–6 motivating a person to change behaviour to achieve a specific outcome. Two concepts are central to goal setting theory and practice: self-determination9 and self-efficacy.10 According to self-determination theory, an individual’s motivation is enhanced if the goal meets the intrinsic needs of the individual.9 Motivation to engage in the process is increased when an individual works on goals that are directly important to them, with a greater likelihood of behaviour change. According to the theory of self-efficacy, a person’s perception of their ability to perform a task or change behaviour determines whether they will initiate that task and how motivated they will be to maintain that task or new behaviour, irrespective of obstacles.10

The effectiveness of goal setting for CLBP has principally been investigated with goal setting embedded within an intervention, with inconsistent findings.11–13 One reason for the inconsistent findings may be that studies used different goal setting processes. Furthermore, the extent of participants’ actual involvement in setting the goals is rarely specified, making it difficult to ascertain whether the clinician or patient sets the goals. Patient’s goals do not correlate well with clinicians’ goals.14 When a healthcare professional directs the goal setting, there is a risk that he or she overlooks the problems that matter most to the patient. Despite guidelines advocating patient-centred self-management support, this is poorly executed by clinicians.4

Patient-led goal setting has been little investigated in CLBP. Our pilot study (n=20) showed that a patient-led goal setting approach is feasible for CLBP and improved the outcomes of disability, pain, quality of life, pain self-efficacy and fear avoidance beliefs and that the magnitude of the changes exceeded minimally clinically important differences (MCID).15 The MCID was determined for disability (Quebec Back Pain Disability  Scale [QBPDS]) >20 and for pain intensity (numerical rating scale [NRS]) >2.15 16 These data provided the rationale for a randomised controlled trial to test the effectiveness of a patient-led goal setting approach in CLBP versus usual care.

We defined a patient-led goal setting approach as an intervention that consists of: (1) education about the neuroscience and psychology of pain; (2) patient-led goal setting, where the patient identifies issues or problems that matter most to them, sets the goals and develops strategies to achieve these goals with guidance from the healthcare professional; and (3) individual feedback with collaborative discussion on barriers and progression of the strategies.15 This approach is supported by the healthcare professional; however, the patient drives the process and is encouraged to undertake the steps towards goal achievement. This approach encourages personal investment in goal setting and optimises the patient’s sense of mastery and ownership in the skills of goal setting and strategy development.

We aimed to establish the effectiveness and healthcare use of a patient-led goal setting intervention compared with standardised advice to exercise (control) for participants with CLBP.

The primary objective was to determine:

  1. Whether goal setting led to decreased disability and pain intensity compared with a control intervention in patients with CLBP.

The secondary objectives were to determine whether for patients with CLBP:

  1. Goal setting led to greater improvements in health-related quality of life, reduced pain-related fear avoidance, symptoms of depression, anxiety and stress and improved pain self-efficacy than a control intervention.

  2. There is a difference in healthcare use between groups.

We hypothesised that a patient-led goal setting approach (intervention) would be more effective in improving clinical outcomes compared with simple advice to exercise (control) over 12 months in CLBP.


Study design

We conducted an assessor-blinded randomised controlled trial with two arms: (1) patient-led goal setting and (2) standardised advice to exercise, with an equal allocation to both arms. Participants were recruited from primary care clinics and community settings within the metropolitan region of Sydney, Australia. The trial was prospectively registered with the Australia and New Zealand Clinical Trials Register:


Inclusion criteria: eligible participants were aged between 18 years and 65 years, with a history of non-specific LBP of a minimum duration of 3 months3, reported pain intensity of at least 4 on an 11-point NRS and disability of at least 20 points on the QBPDS. These levels were chosen to ensure an MCID was able to be detected.16

Exclusion criteria: potential participants were excluded if they: were unable to comprehend written English; reported lumbar spine surgery in the last 12 months; or reported signs and symptoms indicative of serious pathology, such as bowel or bladder dysfunction, change in sensation in the perineal region, worsening pain at night, recent fevers/illness, prolonged morning stiffness or recent worsening of neurological symptoms in the lower legs.

All participants gave written, informed consent.

Randomisation and masking

Eligible participants who consented to enter the study were randomly allocated by coin toss, by an independent research assistant, to either the patient-led goal setting group (intervention) or standardised exercise group (control). This allocation was not concealed to the independent research assistant. Because of the nature of the interventions, it was not possible to blind participants or the treating physiotherapist. A research assistant blinded to group allocation administered all outcome measures. The statistician was unaware of group allocation prior to completion of data collection and data cleaning.


Patient-led goal setting intervention (table 1): the intervention procedure duration was 2 months and followed the protocol previously described and shown to have positive feasibility in the publication of the pilot study results.15 The procedure consisted of five face-to-face sessions (initial session: 1-hour duration, four sessions: 15–30 min duration), conducted at 2 week intervals, followed by two follow-up sessions (30 min duration) spaced 1 month apart. A session was scheduled 12 months from baseline in order for outcome measures to be collected. Outcome measures were collected by self-report questionnaires completed by the participants and returned in opaque envelopes at the 2 month, 4 month and 12 month face-to-face follow-up sessions.

Table 1

Components of the intervention for each group

At the initial session participants were given a ‘Participant Handbook’ (see online supplementary file 1) used in the pilot study, comprising education about neuroscience of pain, background information on the chronic pain model,17 tips for self-management of CLBP, information on setting goals and guidelines following the SMART model.18 The SMART model involves a goal setting process, which is Specific, Measurable, Achievable, Relevant with a Timeframe specified.

Supplemental material

The therapist, an experienced senior physiotherapist (TG), trained in patient-led goal setting, took a history of the participant’s back pain and discussed any problems experienced as a result of the back pain. The participant was asked to prioritise these problems according to what they wanted to focus on. Strategies based on evidence-based guidelines were then discussed, and the participant set specific goals and strategies to work on independently between sessions. Participants recorded in their workbook their identified goals, progress towards achieving these goals, any barriers and agreed strategies to be undertaken towards goal achievement. If goal strategies involved consultation with a healthcare professional, the participant was encouraged to seek this out independently, and this was recorded in the field notes. Nil other explicit instructions were given in regards to their usual care.

Standardised advice to exercise (control): participants assigned to the control group were offered three face-to-face sessions (at baseline: 1-hour duration; 2 months: 30 min duration; and 4 months: 30 min duration) with the treating physiotherapist to discuss their CLBP history and receive advice on a standardised exercise programme. This was a pragmatic study design and the number of sessions in the control group was chosen to assist with revision of exercises and retention of participants. The advice consisted of guideline endorsed exercises commonly prescribed in clinical practice.19 Exercises were prescribed by the therapist and could include flexibility, stability and strength exercises depending on the participant’s individual presentation (see online supplementary file 2 for sample of exercises). The exercise programme was reviewed with each participant at the face-to-face session at 2 months and 4 months. Participants in the control group were asked to continue with their normal routine of back care for the duration of the study. Participants were not provided with education about the neuroscience and psychology of pain and did not discuss treatment goals.

Supplemental material

All interventions were carried out at the participants preferred location, rooms situated at either the University of Sydney campus or a major metropolitan hospital.

Treating physiotherapist: a single treating physiotherapist, who was also the principal investigator, provided both interventions. The treating physiotherapist (TG) was a senior clinician with over 20 years of clinical experience and was trained in the implementation of the patient-led goal setting approach.



Demographic characteristics such as age, sex, level of education and employment status were collected at baseline.

Primary outcomes

Disability: the QBPDS (minimum score=20, maximum score=100) was used to assess disability. The QBPDS is a validated and widely used measure of disability in CLBP research.20

Pain intensity: the NRS was used to measure pain intensity. The anchors were 0 (no pain) and 10 (worst pain imaginable). The NRS is a validated and well-established measurement tool and is responsive to changes in pain.21

Secondary outcomes

Quality of life was assessed using the SF-36 (minimum score=0, total maximum score=100, with a higher score indicating better quality of life) with physical and mental component scores also calculated.22 The SF-36 is a validated multipurpose, short-form health survey with 36 questions and is the most widely used measure of general health.

Symptoms of depression, anxiety and stress were assessed using the Depression Anxiety Stress Scale (DASS).23 The DASS is a validated 21-item questionnaire that yields three subscale scores: depression (range 0–28), anxiety (range 0–20) and stress (range 0–34). A higher score indicates greater severity.

Self-efficacy was assessed using the Pain Self-Efficacy Questionnaire (PSEQ)24 (ranging from 0 to 60). The PSEQ is a validated self-report, 10-item questionnaire and a higher score reflects greater self-efficacy.

Fear of movement/(re)injury was assessed using the Tampa Scale for Kinesiophobia (TSK)25 (minimum score=17, maximum score=68). The TSK is a validated self-report 17 item questionnaire, and a higher score reflects greater fear of reinjury due to movement.

Healthcare use

A healthcare survey was administered to capture healthcare use in the previous month at baseline, 4 months and 12 months (See online supplementary file 3). These time frames were chosen to allow treatment effect to stabilise before we measured differences in resource use. Participants were asked to report on several areas of healthcare use such as number of general practitioner visits, allied health visits, hospital admissions and medication use.

Supplemental material

Statistical analysis

Means and SD were calculated using SPSS Inc V.22. Analysis was by intention to treat. Statistical analysis was carried out by a University of Sydney statistician blinded to the intervention delivered to each allocated group.

Clinical measures

A minimal clinically worthwhile effect was considered to be a change of 20 points on the Quebec Back Pain Disability Scale (QBPDS)16 and a change of 2 points on the NRS.16

The endpoints were established and included disability, pain, quality of life, depression, anxiety, stress, pain-related self-efficacy and fear of movement and were measured at each time point, that is, baseline, post-treatment (2 months), and at 4 and 12 months after conclusion of the intervention. Healthcare use was collected at baseline, 4 months and 12 months.

To estimate whether differences between time points and between groups were statistically significant, a linear mixed model with an unstructured covariance matrix was used for each outcome measure. The interaction between time and group was included to allow the groups to change differently over time and to obtain least significant difference post hoc comparisons of group within time. The mean differences and 95% CIs reported are predicted mean values from the models. This approach allowed data for all subjects to be included in the analysis regardless of whether they completed all study time points. No imputation of missing values was required.

The difference between the mean changes was determined to be statistically significant if p<0.05. Effect sizes were calculated for between group changes in mean differences from baseline to the post-treatment time point for all outcome measures. Cohen’s d was used for changes over time, where a Cohen’s d=0.2 represents a ‘small’ effect size, 0.5 represents a ‘medium’ effect size and 0.8 a ‘large’ effect size. Glass’s delta was used for between group effects.

Healthcare use

The Cochran-Mantel-Haenszel test26 for trend was used to assess if the trends in medication use, visits to allied health practitioners, other health professionals or facilities and days lost at work were different between groups. The difference was determined to be significant if p<0.05.

Sample size

Based on our pilot data, the SD for pain intensity on the NRS was assumed to be 2.0 points.15 To have 95% power to detect a minimum difference of 2.0 points14 between the two groups with alpha=0.025, a minimum sample size of 62 is required. To adjust for 20% drop outs, a minimum sample size of 75 was set.


Seventy-five participants were recruited between January 2014 and February 2015. Figure 1 shows the trial profile as per the Consolidated Standards of Reporting Trials protocol.27 For participants who commenced the study, 100% (intervention and control) follow-up was achieved at post-treatment assessment, 95% (intervention and control) at 4 months, (5.4% dropout rate for intervention, 5.3% dropout rate for control) and 81% (intervention) and 92% (control) at 12 months follow-up (13.5% dropout rate for intervention, 7.9% dropout rate for control). Loss to follow-up at 4 months and 12 months was due to loss of contact with the participant.

Figure 1

CONSORT flow diagram. CONSORT, Consolidated Standards of Reporting Trials.


The mean age of participants was 44.3 (SD 13.2) years, 43% were male, and the mean duration of LBP was 10.1 (SD 9.8) years. Of the 75 participants, 63% had tertiary level education, 59% were employed full-time, 21% were employed part time/student, and 12% were unemployed (table 2). Demographic characteristics were comparable between groups at baseline.

Table 2

Demographic data for participants in each group at baseline (n=38 control, n=37 intervention)

Clinical outcomes

Mean scores, SD and range for primary and secondary outcome measures at all time points, mean between group differences for all time points and mean between group differences from baseline effect size are shown in table 3.

Table 3

Mean and SD for primary and secondary outcome measures and mean (95% CI) between group differences for primary and secondary measures at baseline, post-treatment, 4 months and 12 months

Primary outcomes:  mean between group differences in improvement were significantly greater in the intervention group than the control group (P<0.05) for both disability and pain at post-treatment (disability=11.1; pain=2.3), 4 months (disability=12.9; pain=2.3) and 12 months (disability=11.6; pain=2.1) (table 3). Figure 2 shows the mean (SE) difference from baseline to all time points for the primary measures of disability and pain intensity.

Figure 2

Mean and SE for change from baseline for primary outcomes of disability and pain intensity. NRS, numerical rating scale; QBPDS, Quebec Back Pain Disability Scale.

Moderate-high effect sizes (Cohen’s d: 0.7–1.45) in mean between group differences from baseline to post-treatment were found for disability, pain, quality of life, self-efficacy and fear avoidance (table 3).

Secondary outcomes: mean between group differences were significantly greater (p<0.05) for quality of life, self-efficacy, kinesiophobia and measures of stress at post0treatment, 4 months and 12 months in the intervention group compared with controls (table 3). Significant between group mean differences were greater in the intervention group compared with controls (p<0.05) for depression from baseline to 4 months and from baseline to 12-month follow-up. Significant between group mean differences were greater in the intervention group than in the control group (p<0.05) for anxiety from baseline to 4 months only. Small to moderate effect sizes (Cohen’s d: 0.1–0.38) in mean between group differences form baseline to post-treatment were found for depression, anxiety and stress.

Negative changes favour outcomes of disability, pain intensity, kinesiophobia, depression, anxiety and stress. Positive changes favour outcomes of quality of life and self-efficacy.

Healthcare use

There were no significant between group differences in mean changes for healthcare use across time points when adjusted for baseline differences (p>0.05). A trend was observed for both groups in reduced GP visits, physiotherapy visits and reduced use of prescribed medication for pain from baseline to 4 months that continued to 12 months, without further reduction (table 3).


We found that patient-led goal setting was a more effective approach than standardised advice to exercise for managing CLBP with significantly greater improvement and moderate to high effects on disability, pain, quality of life, self-efficacy and fear of movement. These improvements were clinically meaningful (QBDS >20 points, NRS >2 points),16 achieved at 2 months post-treatment and maintained at 12-month follow-up. There were small to moderate beneficial effects for depression, anxiety and stress.

The high effect on pain intensity with the intervention group at post-treatment was maintained long term. In chronic pain conditions, pain is not expected to improve long term, particularly with an approach that aims to change behaviour.28 29 Systematic reviews have consistently concluded that interventions involving a combination or single intervention of education, goal setting and cognitive–behavioural therapy all have a significant but small short-term effect on pain intensity.11 30 31 A Cochrane review has shown that multidisciplinary interventions, which commonly combine education, goal setting and cognitive therapies, have a small to moderate effect on pain for CLBP in the short term (0.55 [95% CI 0.83 to 0.28]), which reduces to a small effect in the long term (0.21 [95% CI 0.37 to 0.04]).30 Similarly, cognitive therapies have been shown to have small to moderate effect on pain in the short term only; operant therapy (0.43 [95% CI 0.11 to 0.75]) and cognitive–behavioural therapy (0.60 [95% CI 0.22 to 0.97]).11

There is limited evidence that education reduces pain and disability in CLBP long term.31 Smaller effects are found on disability measures in CLBP with multidisciplinary interventions showing small effect both short and long term (short term: 0.41 [95% CI 0.19 to 0.62],30 long term: 0.23 [95% CI 0.06 to 0.940]).30 In contrast to existing practices of goal setting,32 our model incorporates an effective patient-led approach and leads to significant moderate to large effects in pain and disability. This study’s model of goal setting, which allows the patient rather than the clinician to lead the development of goals, has been shown to be effective in facilitating self-management in several chronic conditions such as asthma,33 diabetes,34 heart disease35 and obesity.36 However, the literature is silent on the role and effectiveness of a patient-led goal setting intervention in CLBP.

Self-efficacy and fear avoidance are both postulated to be contributing factors to long-term disability in CLBP.37 We have shown that a patient-led goal setting intervention has moderate to high effect sizes with long-term improvement on both self-efficacy and fear avoidance scores. These improvements exceed those found with education alone, about the neuroscience and the psychology of pain.38 Neuroscience education is a necessary component of our intervention as it underpins other active cognitive and behavioural strategies that enable a person with CLBP to shift their beliefs and re-engage with activity.38 We hypothesise that the addition of the patient-led goal setting component engages the patient with personal and valued goals. This provides an opportunity for the patient to master the skills necessary to change behaviour long term, which is central to the development of strong self-efficacy.10 To tease out the unique effects of goal setting, future research could test and compare a variety of models such as education plus patient-led goal setting versus education plus usual care, education combined with other interventions (eg, cognitive–behavioural therapy) and goal setting alone.

Direct health costs for the management of CLBP are high, with CLBP being one of the most common reasons to seek treatment with a medical practitioner, as well as seeking both conventional and alternative medical approaches a popular practice in patients with CLBP.7 In our study, seven consultations were sufficient to facilitate behaviour change that was maintained up to a 12-month follow-up.


This study has some limitations. The study used a single therapist and face-to-face mode of delivery to conduct the intervention. Although this allowed for consistent intervention during the study, this may have led to preference bias. Future research could explore whether an alternative mode of delivery (eg, telephone, online and digital) or frequency or duration of consultations achieves equally effective outcomes. As participants were randomised at baseline, sex differences were not explored. The small effect on DASS scores was not explored in this study but could be explained by the relatively low to normal scores at baseline; therefore, we would not expect to see large changes. Only direct healthcare costs were calculated; future studies would benefit from a broader economic analysis including data on items such as productivity loss. Although healthcare use was lower in the intervention group, this was offset by seven consultations included in the programme for the intervention group. The time spent with participants in each group varied: patient-led goal setting consisted of seven sessions and the standardised advice (control) consisted of three sessions. We cannot reject the possibility that this difference in therapeutic consultation time may have affected the outcomes. Data on patient adherence to the advice to exercise were not collected, which does not allow us to directly compare the effect of exercise that participants were engaged in. It is unknown whether our findings would generalise to the broader clinical population and whether a general CLBP population would be willing to participate in patient-led goal setting, particularly those who present to tertiary care pain clinics who have higher depression, anxiety and stress levels.


In patients with CLPB, patient-led goal setting improved disability, pain, quality of life, self-efficacy and fear of movement. This is the first evidence that an intervention that encourages patients to set their own treatment goals is a valuable approach to the management of CLBP.

Summary box

  • There is no prior evidence for patient-led goal setting to improve outcomes for chronic low back pain (CLBP). Research into goal setting remains unclear in terms of the processes practised and the level of patient involvement.

What are the findings?

  • First study to investigate the effectiveness of a patient-led goal setting approach for CLBP.

  • A patient-led goal setting approach effectively improved disability, pain, quality of life, self-efficacy and fear avoidance in CLBP, and these improvements are maintained long term.

How might it impact on clinical practice in the future?

  • Patient-led goal setting may be a valuable intervention for CLBP.

  • As effect sizes were moderate to high, this intervention may be as good as or even better than other interventions for CLBP in which behaviour change is the aim. This will need formal head to head testing.

  • Placing patients in the principal role and allowing them to set goals that are intrinsically central to their life may facilitate improved treatment results.

  • Patient-led goal setting involved more contact than the advice to exercise group.

  • Future research could address the mode of delivery and practice of goal setting.


We thank Dr Jenny Peat for her advice and assistance in the statistical analysis.



  • Contributors TG, KR, JM, SG and LS conceived the study and designed the study in collaboration. TG analysed the data and wrote the first draft of the manuscript. All authors interpreted the data and contributed to subsequent drafts of the manuscript, and all authors have seen and approved the final version.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Ethics approval Ethics approvals were obtained from the University of Sydney (HREC: 2014/572) and St Vincent’s Hospital, Sydney (HREC: 14/SVH/337) Human Research and Ethics Committees.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement All of the individual participant data collected during the trial will be shared, after deidentification. Study protocol, statistical analysis plan and informed consent form will be made available. Data will be available beginning 3 months and ending 5 years following article publication. Data will be shared with researchers who provide a methodologically sound proposal to achieve the aims in the approved proposal. Proposals should be directed to; to gain access, data requestors will need to sign a data access agreement.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.