If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
School of Medicine and Public Health, University of Newcastle, University Drive, Callaghan NSW, 2308, AustraliaHunter New England Population Health, Hunter New England Local Health District, Locked Bag 10, Wallsend NSW, 2287, Australia
School of Health Sciences, University of Sydney, PO Box M179, Missenden Rd, Camperdown NSW, 2050, AustraliaNepean Blue Mountains Local Health District, Penrith NSW, 2750, Australia
School of Medicine and Public Health, University of Newcastle, University Drive, Callaghan NSW, 2308, AustraliaHunter New England Population Health, Hunter New England Local Health District, Locked Bag 10, Wallsend NSW, 2287, Australia
School of Medicine and Public Health, University of Newcastle, University Drive, Callaghan NSW, 2308, AustraliaHunter New England Population Health, Hunter New England Local Health District, Locked Bag 10, Wallsend NSW, 2287, Australia
School of Medicine and Public Health, University of Newcastle, University Drive, Callaghan NSW, 2308, AustraliaHunter New England Population Health, Hunter New England Local Health District, Locked Bag 10, Wallsend NSW, 2287, Australia
School of Medicine and Public Health, University of Newcastle, University Drive, Callaghan NSW, 2308, AustraliaHunter New England Population Health, Hunter New England Local Health District, Locked Bag 10, Wallsend NSW, 2287, Australia
To assess the reporting quality of exercise interventions from clinical trials of low back pain (LBP).
Study design and setting
We conducted a systematic review to assess the reporting quality of randomised controlled trials (RCTs) that investigated the effectiveness of exercise interventions for patients with LBP. Five online databases and Clinical Trial Registries were searched (October 2018). We included RCTs that reported interventions for LBP, containing at least 50% exercise. The Template for Intervention Description and Replication (TIDieR) and the Consensus on Exercise Reporting Template (CERT) reporting checklists were then used to assess quality of reporting.
Results
582 trials were eligible for inclusion. Due to the large number of eligible studies, 100 studies were randomly selected for data extraction and coding with the TIDieR and CERT checklists. The random sample was representative of the 582 eligible trials. The overall completeness of reporting (median (IQR)) of TIDieR items was 59.2% (45.5%–72.7%) and CERT was 33.3% (22.2%–52.6%).
Conclusions
We found poor overall reporting with both checklists, which has not improved over time or since the introduction of the checklists. More dedicated work is required to address poor reporting of exercise interventions in clinical trials.
Poor reporting of research is one barrier to the use of evidence-based treatments.
•
Exercise is widely recommended as a treatment for low back pain.
What are the new findings?
•
The reporting of exercise interventions for low back pain is poor.
•
Exercise descriptions do not provide sufficient details to inform use in clinical practice.
•
Reporting quality has not improved over time.
•
Reporting quality has not improved since the introduction of reporting guidelines.
1. Introduction
Low back pain (LBP) is a highly prevalent condition that can cause significant disability and suffering. In 2016 LBP was responsible for approximately 150 million years lived with disability; the leading cause of years lived with disability globally [
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.
]. Low back pain also has a significant economic burden. For example, annual total costs are upward of AUD$9.2 billion in Australia, £12.3 billion in the United Kingdom and US $90.6 billion in the United States of America (USA) [
Clinical practice guidelines for the noninvasive management of low back pain: A systematic review by the Ontario Protocol for Traffic Injury Management (OPTIMa) Collaboration.
Clinical practice guidelines for the noninvasive management of low back pain: A systematic review by the Ontario Protocol for Traffic Injury Management (OPTIMa) Collaboration.
]. Despite these recommendations and evidence to support exercise as an effective treatment, audits of routine practice show that often less than 50% of LBP patients receive exercise as part of their management [
. In particular, accurate description of interventions tested in clinical trials is needed so that treatment protocols can be reproduced for patient care [
]. Since exercise interventions in clinical settings are often complex in nature, due to the range of variables or components that can be modified, they require detailed descriptions to be reproducible in practice [
To improve reporting quality of research, reporting checklists have been created, which highlight important details that should be reported to support interpretation and replication of research [
]. The Enhancing the QUAlity and Transparency Of health Research (EQUATOR) network currently maintains a searchable database of 428 reporting checklists [
. The Template for Intervention Description and Replication (TIDieR) was developed with the aim of improving the completeness of reporting of healthcare interventions [
Interventions to increase adherence to therapeutic exercise in older adults with low back pain and/or hip/knee osteoarthritis: a systematic review and meta-analysis.
, no studies have assessed the reporting quality of exercise interventions for LBP. Without knowledge of how well exercise interventions are reported, it is difficult to determine if reporting quality is a barrier to translation of evidence-based exercise interventions into practice [
To assess the reporting quality of exercise interventions from clinical trials of low back pain.
3. Methods
We conducted a systematic review to assess the reporting quality of randomised controlled trials (RCTs) that investigate the effectiveness of exercise interventions for patients with LBP. We used the TIDieR and CERT reporting checklists to assess quality of reporting. The review was prospectively registered with PROSPERO (registration number CRD42018112421).
3.1 Search strategy and inclusion criteria
We searched the Cochrane Central Register of Controlled Trials, MEDLINE, EMBASE, PsychInfo and CINAHL databases from inception to October 20, 2018. A search strategy was developed with the assistance of a clinical librarian (Appendix 1). We included RCTs published in peer-reviewed journals that recruited people of any age with LBP of any duration to assess the effect of exercise-based interventions. Exercise-based interventions were defined as any intervention that compromised >50% exercise-based therapy; exercise being “a series of specific movements with the aim of training or developing the body by a routine practice or as physical training to promote good physical health” [
]. We had no restrictions on comparison groups or study outcomes. Reported information (number of sessions, timing of components, program description etc) was used to assess if an intervention was >50% exercise. We excluded trials of patients with specific causes of low back pain (including pregnancy-related LBP, fractures, spinal stenosis, radiculopathy/ sciatica, ankylosing spondylitis), trials of cervical or thoracic back pain, trials that did not report data for low back pain separately, papers not written in English, and trials of rehabilitation or physiotherapy interventions with no specific reference to exercise therapy.
Two reviewers assessed titles and abstracts independently, followed by full text articles to identify relevant trials (Figure 1). Any disagreements were resolved by discussion or consultation with a third reviewer. We identified 582 trials eligible for inclusion. Due to the large number of eligible trials, we decided, for reasons of feasibility, to select a random sample of 100 studies for data extraction and synthesis. Prior to random selection, all exercise interventions, from intervention and control arms were coded according to exercise type. Coding was conducted in accordance with the main exercise groups used by the American College of Sports Medicine for testing and prescription [
]. Random sampling was completed by selecting the first 100 studies following random ordering of trials using the random number generator function in Microsoft Excel (Microsoft Corp., Redmond, WA). We compared exercise-type codes of the 100 selected articles to the overall sample of 582 trials to ensure representativeness of our sample.
Two authors independently extracted study data. Discrepancies were resolved through discussion between reviewers or a third reviewer if arbitration was necessary. Data for TIDieR and CERT items were categorized as ‘Yes’, ‘No’, ‘Unsure’, or ‘Not Applicable (N/A)’. An item was assigned ‘unsure’ if the study did not fully report information specified in explanation documents for the TIDieR [
] checklists. An item was classified as ‘N/A’ if it was not relevant to the intervention being reported. For example, CERT Item 9 (home program) was scored as N/A if a home program was not specified as part of the intervention. We did not assess risk of bias, as the outcome of interest was reporting quality.
We calculated summary statistics of trial details. Results of TIDieR and CERT were reported by item and we calculated completeness scores for each trial. We summarised completeness scores for periods before and after release of each reporting checklist (TIDieR- 2014, CERT- 2016) and each of the last three decades. For analysis, ‘unsure’ was considered as ‘no’.
4. Results
The initial search retrieved 13,921 articles. Removal of duplicates via processes in Endnote (Clarivate, Philadelphia, PA) [
] and Covidence (Covidence, Melbourne, VIC) left 8,811 articles for screening. We identified 582 trials for inclusion. The 100 randomly selected trials for data extraction were published between 1990 and 2018 across 50 journals. The proportion of studies by exercise type was similar across the 582 eligible studies and the randomly selected 100 studies (Table 1).
Other’ refers to exercise types that did not specifically fit into any of the other exercise type definitions (e.g. core stability exercises, McKenzie flexion or extension exercises).
Not Specified’ refers to those groups where the intervention or therapy was not described.
27 (2.0%)
1 (0.4%)
Total
1381 (100%)
232 (100%)
N.B. Totals do not equal the number of RCTs as both intervention and control groups were included. Counts reflect trial groups/arms, not the individual studies.
a Combination’ refers to groups that combined 1 or more of the main exercise types (aerobic, flexibility, strength).
b Other’ refers to exercise types that did not specifically fit into any of the other exercise type definitions (e.g. core stability exercises, McKenzie flexion or extension exercises).
c Not Exercise’ refers to those groups that did not include any exercise or included <50% exercise.
d Not Specified’ refers to those groups where the intervention or therapy was not described.
The 100 assessed trials had a median sample size of 60.5 (IQR, 34.3–136.3), and median intervention duration of 6 weeks (IQR, 4-10 weeks), of which 70% were delivered face to face. The majority of assessed trials (67%) investigated chronic LBP populations. The main outcomes reported were pain intensity (79% of trials) and disability (74% of trials). See Appendix 2 for a reference list of the 100 included studies and Appendix 3 for further details regarding the 100 trials.
We extracted data for 180 exercise intervention groups (intervention and control arms) across the 100 included trials. The most commonly reported TIDieR item (Table 2) was item 1 (Brief Name of intervention – 100% of interventions), followed by item 2 (Why: Underlying rationale, theory or goal of essential elements – 92.8%) and item 8 (When and How Much (number, schedule, duration, intensity, dose of intervention – 83.9%). While the least commonly reported items were item 10 (Modifications - 0%), item 9 (Tailoring – 23.2%), and item 11 (How well (Planned) – 35.7%).
Table 2Results for TIDieR items (N = 180 interventions)
For the CERT (Table 3), item 4 (Supervised or unsupervised and how delivered - 68.3%) was the most commonly reported, followed by item 10 (Description of non-exercise components - 64.7%) and item 13 (Description of the exercises used - 58.7%) and item 14a (Generic (one size fits all) or Tailored - 58.7%). The least commonly reported items were item 15 (Description of decision rule for starting level – 8.8%), item 7a (Description of decision rule(s) for progression – 16.9%), and item 6 (Description of motivation strategies used – 19.9%).
Table 3Results for CERT items (N = 180 interventions)
Yes (n)
Unsure (n)
No (n)
N/A (n)
CERT 1- Type of exercise equipment
% from valid responses (n = 165)
26.1% (n = 43)
10.9% (n = 18)
63.0% (n = 104)
- (n = 15)
CERT 2- Qualifications and/or training
% from valid responses (n = 180)
30.0% (n = 54)
37.8% (n = 68)
32.2% (n = 58)
- (n = 0)
CERT 3- Individual or Group
% from valid responses (n = 180)
51.7% (n = 93)
30.6% (n = 55)
17.8% (n = 32)
- (n = 0)
CERT 4- Supervision/ Unsupervised and how delivered
The overall completeness of reporting (median (IQR)) of TIDieR items was 59.2% (45.5%–72.7%) and CERT was 33.3% (22.2%–52.6%) (Table 4). The overall completeness was similar pre and post introduction of TIDieR and CERT checklists (Table 4). Analysis by time period was completed for 1990–1999, 2000–2009, 2010–2015 and 2016-Present. The period from 2010-Present was divided into 2 categories to represent post checklist publication periods and allow for diffusion of use (and subsequent assessment of its effect on reporting). Across the 4 time points, reporting assessed via TIDieR was variable, however overall completeness of reporting assessed via CERT got worse (Table 5).
Table 4Completeness of reporting pre/post checklist
We found the overall reporting quality of exercise interventions was typically poor as assessed by the TIDieR and the CERT checklist. Underlying theories and parameters (incl. type, dose, duration) of the interventions were generally well-reported, but reporting on specific elements such as materials, infrastructure, training of therapists and adaptations was poor. Reporting quality has not improved over time nor since the introduction of reporting checklists, as would be expected.
Previous research into the reporting quality of interventions for musculoskeletal conditions has found variable results [
Interventions to increase adherence to therapeutic exercise in older adults with low back pain and/or hip/knee osteoarthritis: a systematic review and meta-analysis.
The Consensus on Exercise Reporting Template (CERT) applied to exercise interventions in musculoskeletal trials demonstrated good rater agreement and incomplete reporting.
. Nascimento, et al. (18) reported “moderately-well” described interventions in 18 studies of non-specific LBP in older adults. These interventions included complimentary health approaches, percutaneous electrical nerve stimulation, exercise and pharmacological agents. Other studies have investigated reporting quality of Tai Chi for chronic musculoskeletal pain [
Interventions to increase adherence to therapeutic exercise in older adults with low back pain and/or hip/knee osteoarthritis: a systematic review and meta-analysis.
Interventions to increase adherence to therapeutic exercise in older adults with low back pain and/or hip/knee osteoarthritis: a systematic review and meta-analysis.
] examined psychological therapy trials using the TIDieR tool. They found relative high reporting of item 2 (Rationale), similar to our study, however variable rates across the other TIDieR items. One other review assessed 20 trials of exercise interventions for a range of musculoskeletal conditions using CERT. Similar to the findings from our review, the authors found poor reporting, with 60% of papers missing information for more than half of the CERT items [
The Consensus on Exercise Reporting Template (CERT) applied to exercise interventions in musculoskeletal trials demonstrated good rater agreement and incomplete reporting.
Our review is the most comprehensive assessment of reporting quality of LBP exercise interventions; we assessed 180 exercise interventions from 100 trials. We used two assessment checklists to enable robust assessment of reporting quality. A limitation is that we could not feasibly assess all trials eligible for inclusion. However, we are confident our sample is representative as we randomly selected trials to assess and confirmed the distribution of exercise types with the full set of trials. Another potential limitation is our assessment of change in reporting quality. The checklists are not designed to assess change in reporting quality over time and may not provide sensitive scores to measure this. As we assessed change in reporting quality in a number of ways and the findings were consistent - that reporting quality has not improved - we have confidence in the findings. We did not assess risk of bias of the included studies, as our outcome of interest was reporting quality and not the estimation of treatment effects, we did not feel this was needed.
5.3 Implications for practice and policy
Without appropriate reporting, clinicians may not be able to translate research about exercise interventions into practice [
. While reporting of factors like the underlying theory behind the exercise, adherence rates, and motivation strategies are useful to consider the validity of a trial, reporting of exercise parameters is needed to replicate the intervention clinically. Unintended deviations from exercise programs tested in clinical trials, due to poor reporting, may lead to suboptimal benefits or potential harm to patients. Although researchers have been encouraged to improve their reporting and numerous checklists have been published to support this, reporting quality has not improved.
Research waste from poor reporting of clinical trials is an ongoing issue [
] completed in 2005 identified 61 RCTs investigating exercise interventions for LBP. We found 582 trials of exercise interventions for LBP, meaning that in the last 15 years a significant number of new studies have been published. Our review shows that most of the included trials may not be capable of informing clinical practice as critical details regarding the exercise interventions are missing from the reports. Given that LBP is the leading cause of disability worldwide, more accountability is required to improve reporting quality to enable better use of evidence-based interventions and reduce research waste [
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.
Currently, little is known about the best way to influence authors to use, and journals to enforce the use of, reporting checklists. At present only a small number of journals require authors to complete the TIDieR checklist with submission of their paper, prior to acceptance for publication. However, as it was not designed specifically for exercise interventions, more detailed information is usually required for clinical use. Up to 40% of trials in our review did not report adequate details about dose or tailoring of interventions, information that would be required to replicate that intervention in clinical practice. Further research about what improves reporting quality is needed to support the translation of evidence-based treatments into practice. This should aim to understand targets to change author behavior, for example, investigating why authors omit certain details of interventions or understanding barriers and enablers to using checklists. These questions should be supplemented with better understanding about what information clinicians need to implement trial interventions.
6. Conclusion
Despite ongoing attention and checklists targeted at the issue, poor reporting of clinical trials continues. This review examined published descriptions of 180 exercise interventions for studies using exercise for LBP and found poor reporting using the TIDieR and the CERT reporting checklists. Further work is required to improve reporting by researchers and increase uptake of checklist use in journals, as a requirement for publication.
Declarations
Ethical approval
Not applicable.
Author contributions
SD, CW and SK developed and designed the project. RH assisted with data management and analysis. ZY, ER, CG, AW, PVdS assisted with screening and data extraction. SD drafted this manuscript and all authors revised it critically for important intellectual content and approved the final version. SD attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted. SD: Conceptualization, Methodology, Investigation, Formal Analysis, Data Curation, Writing- Original Draft, Project administration. SK: Conceptualization, Methodology, Writing- Review and Editing, Supervision. RH: Formal Analysis, Writing- Review and Editing, Supervision. ER: Investigation, Data Curation, Writing- Review and Editing. CG: Investigation, Data Curation, Writing- Review and Editing. PVdS: Investigation, Data Curation, Writing- Review and Editing. AW: Investigation, Data Curation, Writing- Review and Editing. ZY: Investigation, Data Curation, Writing- Review and Editing. CW: Conceptualization, Methodology, Writing- Review and Editing, Supervision.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. CW and SK received salary funding from the National Health and Medical Research Council of Australia.
Availability of data and materials
Data used for analysis will be made available on reasonable request. Proposals for data use may be submitted to the principal/corresponding author.
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.
Clinical practice guidelines for the noninvasive management of low back pain: A systematic review by the Ontario Protocol for Traffic Injury Management (OPTIMa) Collaboration.
Interventions to increase adherence to therapeutic exercise in older adults with low back pain and/or hip/knee osteoarthritis: a systematic review and meta-analysis.
The Consensus on Exercise Reporting Template (CERT) applied to exercise interventions in musculoskeletal trials demonstrated good rater agreement and incomplete reporting.
00172-4&id=gr1.jpg){kind=link}