Low reproducibility of randomized clinical trials methodology related to sampling: a systematic methodological review

ABSTRACT

Objectives: The reporting of sampling methods in Randomized Clinical Trials (RCTs) allows for research quality assessment, determination of sampling bias, and assures the presence of details necessary for reproducibility in future trials. The purpose of this study was to: (1) determine if sampling methodology was reproducible in RCTs related to musculoskeletal physical therapy (MSKPT) interventions to treat non-specific low back pain (NSLBP) and (2) establish if there was a relationship between sample reproducibility and established measures of research quality.

Methods: Data were collected through a systematic review by a professional librarian. The identified RCTs were assessed for methodological quality by two blinded individual reviewers. Data analysis was performed by a third, blinded researcher; additional comparisons were made based on Journal Impact Factor and PEDro score.

Results: Ninety-nine published peer-reviewed RCTs were identified that met inclusion criteria. Only 29% of the articles were judged to be reproducible based on the reported sampling methodology. There were meaningful correlations between two out of ten of the sampling reporting criteria and the judgement made if the sample was reported in significant detail to allow for replication. There was no relationship between sampling reporting criteria, Journal Impact Factors (JIFs), and Physiotherapy Evidence Database (PEDro) scores.

Discussion: The reporting of sampling methodology needs to be considered to ensure reproducibility and avoid sampling bias. Despite the proliferation of measures of research quality, the overall reporting quality of RCTs continues to be inadequate to allow widespread reproducibility of trials.

Level of Evidence: 1a

Introduction

The accuracy and completeness of reporting in randomized clinical trials (RCTs) provides the foundation that allows for a determination of the generalizability of a study’s findings, and also for verification of a study’s findings through replication [1]. The CONSORT statement was created in order to improve accuracy of reporting in RCTs [2–4] and allow for critical appraisal and interpretation of findings [5]. The most recent CONSORT statement was published in 2010 [6] and endorsement of its use by journals has increased over time. While the CONSORT checklist explicitly details the need to report sample size and subsequent randomization details, there is no requirement to report the methods used to recruit the sample. In addition to the CONSORT, Journal Impact Factors (JIFs) are used to determine the influence of the journal that a RCT is published in; previous research has suggested that JIF may be used as an indicator of overall journal quality in the medical literature [7]. The Physiotherapy Evidence Database (PEDro) scale is utilized to assess the quality of RCTs; previous studies have shown the PEDro scale to be reliable [8] and valid as a measure of RCT quality [9]. However, neither of these tools assures that the sampling process is adequately described.

Selection bias has been defined by Portney and Watkins [10] as, ‘A threat to internal validity in which bias is introduced by initial differences between groups, when these differences are not random.’ Selection bias may occur during the sampling process of a RCT, and concerns of selection bias in RCTs have been discussed in the medical literature [11,12]. A non-random relationship between a specific condition and the sample included in the study may be created if participants are pre-screened prior to enrollment in the study using the studies inclusion/exclusion criteria [13]. The creation of this relationship between the condition being studied and the sample may result in a study sample that is no longer an accurate representation of the general population. If this occurs, it threatens the internal validity and subsequently the external validity of the study [13,14], as the generalizability of the findings to the general population may no longer be appropriate [15,16]. This suggests that it is important that sampling methodology is reported in enough detail to ensure that selection bias did not occur and to allow for replication of the sampling methods in future trials.

The objectives of this systematic review were to: (1) determine if sampling methodology was reproducible in RCTs related to musculoskeletal physical therapy (MSKPT) interventions to treat non-specific low back pain (NSLBP) and (2) establish if there was a relationship between sample reproducibility and established measures of research quality.

Methods

Protocol and registration

The purpose of performing the systematic review for this study was to create an unbiased sample of RCTs that would represent a commonly addressed musculoskeletal issue in physical therapy (NSLBP). We chose NSLBP in an attempt to attain a broad range of journals that are commonly used by physical therapists that see this as a topic of interest. Since this is a systematic methodology review without an outcome of direct patient or clinical relevance, it was not possible to register the protocol in PROSPERO. This methodology has been previously published [17,18].

Eligibility criteria

For practical purposes searches were limited to RCTs published in the English language related to MSKPT interventions to treat NSLBP from January 2012 to December 2016. This date range was selected in order to obtain studies that utilized the most recent CONSORT statement and to capture a time frame during which journals endorsed its use [19]. The date of last search was 15 December 2016 and no search updates were conducted later in the process.

Information sources

The methods outlined by Furlan [20] and Lefevre [21] were used to guide the development of the search strategies. Controlled vocabulary and free text terms related to MSKPT interventions to treat NSLBP were used to search the following databases and registries for RCTs: PubMed, which includes MEDLINE (1946–15 December 2016), Embase (embase.com 1974 to 15 December 2016), and Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library via EBSCOHost to January 2012 Issue 1 to December 2016 Issue 12).

Search

The search strategies are presented in Appendix 1. These database searches were supplemented by citation searching using bibliographies of reviews and research articles that met the inclusion and inclusion criteria.

The PubMed search strategy was developed by a professional librarian experienced in systematic literature reviews, and peer reviewed by another librarian using the PRESS standard [22]. The PubMed search was then adapted for the other databases. For identification of RCTs in PubMed, the Cochrane Highly Sensitive Search Strategy for identifying randomized trials in MEDLINE: sensitivity-maximizing version (2008 revision); PubMed format [21] was used. For identification of RCTs in Embase, the RCT filter based on the Scottish Intercollegiate Guidelines Network (SIGN) strategy and amended to embase.com format was used, as provided by the Cochrane Work Review Group [23]. For identification of RCTs in CENTRAL, a therapy filter was adapted that was developed for PsycInfo by the Health Information Research Unit of McMaster University [24].

Study selection

Article titles were screened using EndNote™ software (Thomson Reuters Endnote, Philadelphia, PA, Version X7.7.1 Bld 10036) and were exported to Covidence™ (Covidence systematic review software, Veritas Health Innovation, Melbourne, Australia. Available at www.covidence.org). Covidence was used by two independent reviewers to assess abstracts and full texts.

Data collection process

The articles were evaluated for sampling reporting quality. Journal Impact factors (JIF) for the year of publication were attained through the 2015 Journal Citation Reports, Clarivate Analytics, 2017 at https://jcr.incites.thomsonreuters.com/. The PEDro scores for all studies were attained through the PEDro Physiotherapy Evidence Database website at https://www.pedro.org.au/.

Data items

The identified RCTs were assessed for 10 sampling reporting criteria by two individual reviewers without knowledge of each other’s work. This work was then sent to a third reviewer to assess agreement. In instances where there was a disagreement, the third reviewer identified these instances and sent them back to the initial reviewers to determine if a consensus could be reached. If there were any instances where consensus could not be attained the third reviewer made the final decision.

Assessment of reporting of sampling methodology

To assess the overall quality of reporting of sampling methodology in RCTs, the following criteria were considered: 1. Prospective power analysis described and performed (Internal validity); 2. Target population described (External validity); 3. Inclusion and exclusion criteria listed; 4. Subject recruitment process described; 5. Type of sampling methodology (probability or non-probability) described; 6. Specific type of probability or non-probability sampling method described; 7. Sampling location described; 8. Sample size reported; 9. CONSORT Diagram provided; and 10. Time frame of data collection described. For a study to be deemed reproducible, it was determined that the study must at a minimum report: (1) the sample size; (2) inclusion/exclusion criteria; (3) have an identifiable sampling methodology; and (4) the target population. Additionally, the presence of basic sample demographic data, including age, gender, symptom duration, and whether the proportion of individuals who were enrolled compared to those screened could be determined were considered. A criterion explanation and examples are provided for each sampling scoring item in Table 1. The primary and secondary authors created this list of items prospectively based on the description of what a sample is, how its size is determined, how it is created and evolves from recruitment to data analysis, and the methodology involved in probability and non-probability sampling as described in Portney and Watkins [10].

Table 1.

Criteria used and examples of the ten items used to determine the reproducibility of sampling methodology.

Criterion	Definition	Example
1. Prospective power analysis described and performed	Prospective sample size calculations described and reported in the text.	“A power analysis was conducted a priori, in which it was determined that a sample size of 70 participants would provide 80% power for detecting a 7% difference between groups for TrA muscle thickness assessed with ultrasonography.” [26]
2. Target population described	The population of interest is clearly identified in the text.	“Consecutive presentations of people with a referral for LBP treatment to the Physiotherapy Department were invited to participate in the trial.” [26]
3. Inclusion and exclusion criteria listed	Defined in a comprehensive, specific fashion without generalizations.	“Inclusion criteria were a history of LBP of 3 months or greater and a directional preference observed with a mechanical assessment based on the McKenzie method. The area of pain could be localized between the 12th rib and the buttock crease. Patients reporting referred pain into one or both legs, with or without sensory and/or motor changes, were also included.” [26] “Participants were excluded if they were classified as not having a directional preference, were under 18 or over 70 years of age, or lacked adequate comprehension of English to comply with treatment or read the study materials.” [26]
4. Subject recruitment process described	Clear statement of the recruiting methodology (i.e. radio/tv, newspaper, flyers, word of mouth)	“Recruitment was done through internal hospital announcements, via an institutional web site and via notice boards.” [27]
5. Type of sampling described (probability or non-probability) described	Described in enough details to determine if the sampling methodology involves probability or nonprobability sampling	“Consecutive presentations of people with a referral for LBP treatment to the Physiotherapy Department were invited to participate in the trial.” [26] “Sample was selected from a list of patients being seen at the Department of Orthopedics of University Hospital, University of Sao Paulo. We included patients with LBP for more than 3 months (pain felt between T12 and the gluteal fold) who agreed to participate in the study.” [28]
6. Specific type of probability or non-probability sampling method described	If it is probability sampling is it simple random sampling, systematic sampling, stratified sampling, disproportional sampling, or cluster sampling? If nonprobability sampling is it a sample of convenience with consecutive sampling, sample of convenience with self-selection, quota sampling, purposive sampling, or snowball sampling?	“Consecutive presentations of people with a referral for LBP treatment to the Physiotherapy Department were invited to participate in the trial.” [26]
7. Sampling location described	Study location(s) clearly defined to include geographical location (i.e. country, region, or state) and setting (i.e. hospital vs out-patient).	“The study was conducted in the Physiotherapy Department of Concord Repatriation General Hospital, which is a major metropolitan blicpu hospital in Sydney, Australia.” [26]
8. Sample size reported	Final sample size attained reported in the results section.	“A total of 133 people referred to physical therapy for treatment of LBP underwent an assessment for inclusion. Seventy participants met all the inclusion criteria and progressed to data collection and randomization. Thirty-five participants were allocated to each treatment group.” [26]
9. CONSORT Diagram provided	CONSORT diagram is present to include the number of patients assessed, randomized, allocated, present at follow-up, and available for final analysis.	See CONSORT flow Diagram (http://www.equator-network.org/reporting-guidelines/consort/)
10. Time frame of data collection described	Dates of when the study started and stopped are identifiable in the text.	“Recruitment commenced in April 2011 and was completed by March 2013.” [26]

Summary measures

The criteria outlined above were used to make the determination if the sampling methodology was reported in sufficient detail to reproduce the sample in future trials. If the RCT did not meet these criteria, the quality of the sampling methodology could not be assessed to determine the presence of any sampling bias.

Synthesis of results

Data were collated in a Microsoft Excel spreadsheet. Descriptive statistics and inferential analyzes were performed with MedCalc Statistical Software version 17.9.7 (MedCalc Software bvba, Ostend, Belgium; http://www.medcalc.org) and GraphPad InStat version 3.06 for Windows (GraphPad Software, San Diego California USA, www.graphpad.com). Almost all correlations passed the Runs test for a linear relationship. However, all correlations had non-normal distribution of residuals, and the non-parametric Spearman rho correlation analysis was employed. Outliers in data sets were identified using the Generalized ESD test with MedCalc.

Additional analyses

Associations were determined between PEDro scores, JIF, the Number of 10 sampling criteria items present, and the determination of Study Judged to be Reproducible.

Results

Study selection

In total, 4336 records were identified through the database search with 2914 records retained for screening after duplicate records were removed. After screening by title, a total of 210 abstracts records were identified and screened. We retained 176 records for full text review, with 99 records that met the inclusion criteria included in the final analysis. The full text articles which were excluded, are detailed in Figure 1 with the reason for exclusion.

Figure 1.

PRISMA diagram.

Study characteristics

The 99 peer-reviewed research articles selected in this study were published between January, 2012 and December, 2016. Sixteen of the research articles were published in journals with no reported JIF scores through the year 2016. Of the 83 research reports with JIF scores, the mean and standard deviation was 2.83 ± 4.08. There was one article published in JAMA, which was a prominent outlier for JIF score (37.68). With this outlier removed, mean and standard deviations were 2.41 ± 1.29 and 2.74 ± 1.37, respectively. PEDro scores were identified and analyzed for 90 of the 99 articles through the PEDro website, with a mean and standard deviation of 6.74 ± 1.38.

Quality of reporting of sampling methodology

As summarized in Table 2, the reporting of the10 individual criteria for the 99 research articles ranged from 44% to 100%. The two evaluators made a judgement as to whether sufficient information was provided in each research article to allow independent investigators to repeat the research study (Study Judged to be Reproducible) based on the reporting of the sample size, inclusion/exclusion criteria, have an identifiable sampling methodology, and target population. Only 29% of the articles were judged to be reproducible from this perspective.

Table 2.

Correlations between the ‘Judged to be Reproducible’ variable vs. (i) manuscript evaluation criteria, (ii) JIF scores and (iii) PEDro scores, organized as three families of analyzes.

Manuscript Criterion/Variable	% Criteria Reported in Articles	N	Spearman rho
1 Prospective power analysis described and performed	80.8	99	0.257
2 Target population described	100.0	99	undefined
3 Inclusion and exclusion criteria listed	98.0	99	0.092
4 Subject recruitment process described	66.7	99	0.455
5 Type of sampling described (probability or non-probability) described*	51.5	99	0.624
6 Type of non-probability sampling method described*	44.4	99	0.675
7 Sampling location described	89.9	99	0.216
8 Sample size reported	99.0	99	0.065
9 CONSORT Diagram provided	92.9	99	0.178
10 Time frame of data collection described	58.6	99	0.226
JIF Year of Publication	N/A	83	0.212
PEDro Score	N/A	90	0.066

‘*’ = Variables with meaningful correlations are listed in bolded text;

Synthesis of results

The findings of the correlation analysis are summarized in Table 2. The associations between PEDro scores and JIF did not attain statistical significance. There were two meaningful correlations between studies Judged to be Reproducible and two of the sampling reporting criteria (5. Type of sampling described (probability or non-probability) described) (Spearman rho = 0.624) and (6. Type of non-probability sampling method described) (Spearman rho = 0.675).

Discussion

Of the 10 sampling reporting criteria assessed, two had a meaningful, statistically significant relationships with the judgement that the sample was reproducible. There were several aspects of reporting that were consistently included (1,2,3,7,8,9 all >80%), which may be areas positively indluenced by current reporting guidelines. However, it appears that important aspects of sampling are not being reported adequately. While not reaching statistical significance, we feel that criteria #4, ‘description of subject recruitment process’ is a key element needed to understand potential selection bias, which was only reported in 67% of manuscripts assessed. In combination with the significant variables of Type of sampling described (probability or non-probability) described (51%) and Type of non-probability sampling method described (44%), the overall reporting of important aspects of reproducibility was limited.

There were no meaningful relationships between the 10 sampling reporting criteria, JIF, and PEDro scores. This suggests that current measures of journal and research quality do not adequately consider the importance of complete reporting of sampling methodology. Until journals, and the peer-reviewers who assess RCTs, recognize the importance of the reporting of sampling methodology as a threat to the internal and external validity of RCTs, this will continue to be a glaring obstacle in the path towards appropriate evidence informed practice based on reproducible RCTs.

Tools such as the CONSORT suggest a minimal set of data to be reported for a trial; this limited data set may not be adequate to recreate a given study, thereby limiting its external validity. The PEDro scale also assesses the overall quality of trial administration. Similar to the CONSORT checklist, there is no specific category to assess the process utilized to generate the study sample. Despite achieving PEDro scores demonstrating moderate to high trial quality, overall reporting of sampling methodology was poor. This suggests that a set of guidelines similar to the Template for Intervention Description and Replication (TiDieR) [25], which has recently been adopted to help make study interventions more reproducible, may be needed to address the reporting of sampling methodology. Of the 10 sampling reporting criteria that were identified, items 5 ‘Type of sampling described (probability or non-probability)’ and 6 ‘Type of non-probability sampling method described’ shared a meaningful relationship with the final judgement if the study would be reproducible or not based on the reporting of sampling methodology. These two items were also the two items that were reported least frequently, with values of 51.5% and 44.4, respectively. We suggest that standardized reporting of sampling methodology should be considered to improve both reproducibility of research and appropriate interpretation of findings.

Future studies are warranted to investigate if the reporting of sampling methodology is an important component in assessing the quality and reproducibility of RCTs. Additionally, tools such as the CONSORT checklist and PEDro scale should be evaluated further for their sensitivity to capture complete reporting of sampling methodology, risk of sampling bias, the relationship between these variables, and if they are a valid measure of the quality of RCTs. Finally, more research is needed to appraise if the quality of RCTs, and the quality of their reporting, is improving.

Summary of evidence

The reporting of sampling methodology appears to be largely ignored in reporting guidelines. There may be no relationship between the accuracy of reporting of sampling methodology and tools used to assess the quality and potential sources of bias in RCTs.

Limitations

Generalizability of our findings is a limitation of this study. To attain a reasonable number of research articles to allow for a practical literature review and statistical analysis we limited our literature search to MSKPT interventions to treat NSLBP in the English language by publication date. Our findings therefore cannot be generalized to RCTs outside our search criteria.

Conclusions

The findings of this study suggest that, based on the current reporting of sampling methodology, fewer than one third of published trials of MSKPT interventions for NSLBP are reproducible. Due to limited reporting of methodology, a determination of sampling bias cannot be assessed for the majority of trials. In addition, there appears to be no meaningful relationship between established measures of journal quality, such as such as JIF and the PEDro scale, and the risk of sampling bias. Despite the proliferation of measures of research quality, the overall reporting quality of RCTs continues to be inadequate to allow widespread reproducibility and straightforward interpretation of results.

Biographies

Sean P. Riley, PT, DPT, ScD is an Assistant Professor in the Physical Therapy program at Sacred Heart University in Fairfield, CT, USAand faculty in the Sacred Heart University Orthopedic Physical Therapy Residency program. He graduated from the UConn Physical Therapy Program in 2004 and completed his Doctorate in Physical Therapy at Simmons College in 2006. He completed his Doctor of Science from Texas Tech University Health Science Center in 2013. Sean became board certified in Orthopedics through the American Board of Physical Therapy Specialties in 2007 and completed his sports certification in 2009. He completed his Fellowship in Manual Therapy from Regis University in 2018.

Brian T. Swanson, PT, DSc is an Assistant Professor of Physical Therapy in the Department of Rehabilitation Sciences at the University of Hartford in West Hartford, CT, USA and faculty in the Select Medical Orthopedic Physical Therapy Residency program. Dr. Swanson graduated from Quinnipiac University with a Bachelor of Science degree in Physical Therapy in 1996 and earned a Doctorate of Science in Physical Therapy from Andrews University in 2012. Brian became board certified in Orthopedics through the American Board of Physical Therapy Specialties in 2005. He completed two-year residency and Fellowship in the Kaltenborn-Evjenth Norwegian Approach to Orthopaedic Manual Physical Therapy at the Institute of Orthopaedic Manual Therapy, Woburn, MA in 2006.

Jean-Michel Brismée, PT, ScD is professor in the Doctor of Science Program in Physical Therapy at Texas Tech University Health Sciences Center in Lubbock, Texas. Dr. Brismée graduated from the Catholic University of Louvain-la-Neuve in Belgium with Bachelor of Science degrees in Physical Education and Physical Therapy in 1985. He graduated from Texas Tech University in 1996 with a Master of Sciences in Sports Health and earned a Doctor of Science degree in Physical Therapy in 2003. He is the Chair of the Research Committee and Director of the Fellowship program in Orthopaedic Manual Therapy of the IAOM-US, Editor-in-Chief of the Journal of Manual and Manipulative Therapy, and maintains clinical practical at University Medical Center in Lubbock, Texas.

Steven F. Sawyer, PT, PhD, is professor and chair of the Department of Rehabilitation Sciences, Associate Dean for Faculty Development, and director of the Clinical Postural Control Research Laboratory in the Center for Rehabilitation Research, all within the School of Health Professions at Texas Tech University Health Sciences Center. His PhD in Neuroscience was earned from the University of California at San Diego in 1989. He graduated from PT school at the Texas Tech University Health Sciences Center in 1997, and has been on the faculty since then. He received certification in Vestibular Rehabilitation from Emory University in 2001.

Elizabeth J. Dyer is Associate Dean of Library Services and a Research and Teaching Librarian at the University of New England in Portland, Maine where she liaisons with the Doctor of Physical Therapy program. She has a BA from Cornell University, an MLIS from University of South Carolina, and is a Senior Member of the Medical Library Association's Academy of Health Information Professionals (AHIP). She attended the University of Pittsburgh’s Systematic Review Workshop in 2016 and was a 2015 fellow at the National Library of Medicine’s Biomedical Informatics Course. She is Chair-Elect of the North Atlantic Health Sciences Libraries chapter of the MLA.

Appendix 1: Electronic search strategies

PubMed

((((((((((((((((((((((low back pain[mh:noexp]) OR low back pain/rehabilitation[mh]) OR low back pain/therapy[mh]) OR (low back pain/prevention and control[mh])) OR back muscles[mh:noexp]) OR back muscles/injuries[mh]) OR back muscles/therapy[mh]) OR back injuries[mh:noexp]) OR (back injuries/prevention and control[mh])) OR back injuries/rehabilitation[mh]) OR back injuries/therapy[mh]) OR spinal injuries[mh:noexp]) OR (spinal injuries/prevention and control[mh])) OR spinal injuries/rehabilitation[mh]) OR spinal injuries/therapy[mh]) OR lumbosacral region[mh:noexp]) OR lumbosacral region/injuries[mh]) OR lumbosacral region/therapy[mh])) OR ((((((((((((low back pain[tiab]) OR lower back pain[tiab]) OR LBP[tiab]) OR backache[tiab]) OR dorsalgia[tiab]) OR facet joints[tiab]) OR lumbar pain[tiab]) OR lumbago[tiab]) OR oswestry[tiab]) OR roland-morris[tiab]) OR low back pain[ot]) OR lower back pain[ot]))) AND ((((((((((((((((((((((((((physical therapy modalities[mh:noexp]) OR electric stimulation therapy[mh:noexp]) OR pulsed radiofrequency treatment[mh]) OR transcutaneous electric nerve stimulation[mh]) OR exercise movement techniques[mh]) OR exercise therapy[mh:noexp]) OR muscle stretching exercises[mh]) OR plyometric exercise[mh]) OR resistance training[mh]) OR musculoskeletal manipulations[mh:noexp]) OR manipulation, orthopedic[mh]) OR manipulation, osteopathic[mh]) OR manipulation, spinal[mh]) OR therapy, soft tissue[mh:noexp]) OR exercise[mh:noexp]) OR circuit-based exercise[mh]) OR physical conditioning, human[mh]) OR running[mh]) OR jogging[mh]) OR swimming[mh]) OR walking[mh]) OR warm-up exercise[mh]) OR cryotherapy[mh] OR traction[mh]))) OR (((((((((((((((((((((((((((((physical therap*[tw]) OR physiotherap*[tw])) OR exercise*[tiab]) OR motor control[tiab]) OR mobilization[tiab]) OR manipulation[tiab]) OR manual therapy[tiab]) OR stabili*[tiab]) OR training[tiab]) OR running[tiab]) OR jogging[tiab]) OR swimming[tiab]) OR walking[tiab]) OR electrical stimulation[tiab]) OR pulsed radiofrequency[tiab]) OR transcutaneous electric nerve stimulation[tiab]) OR TENS[tiab]) OR dance[tiab]) OR tai ji[tiab]) OR tai chi[tiab] OR yoga[tiab]) OR breathing exercise*[tiab]) OR qigong[tiab]) OR muscle stretching[tiab]) OR physical conditioning[tiab]) OR soft tissue therapy[tiab]) OR cryotherapy[tiab] OR traction[tiab]))))) AND ((((((((((randomized controlled trial[pt]) OR controlled clinical trial[pt]) OR randomized[tiab]) OR placebo[tiab]) OR drug therapy[sh]) OR randomly[tiab]) OR trial[tiab]) OR groups[tiab])) NOT (animals[mh] NOT humans[mh])) AND (‘2012/01/01’[PDat]: ‘2017/12/31’[PDat]) AND English[lang]

Embase

((‘low back pain’:de OR ‘back muscle’:de OR backache:de OR ‘low* back pain’:ab,ti OR ‘back muscle’:ab,ti OR backache:ab,ti) AND (‘physiotherapy’:de OR ‘home physiotherapy’:de OR ‘joint mobilization’:de OR ‘kinesiotherapy’:de OR ‘exercise’:de OR ‘aerobic exercise’:de OR ‘anaerobic exercise’:de OR ‘aquatic exercise’:de OR ‘arm exercise’:de OR ‘athletic performance’:de OR ‘breathing exercise’:de OR ‘circuit training’:de OR ‘closed kinetic chain exercise’:de OR ‘dynamic exercise’:de OR ‘endurance training’:de OR ‘exercise intensity’:de OR ‘isokinetic exercise’:de OR ‘muscle exercise’:de OR ‘isometric exercise’:de OR ‘isotonic exercise’:de OR ‘stretching exercise’:de OR ‘plyometrics’:de OR ‘resistance training’:de OR ‘static exercise’:de OR ‘warm up’:de OR ‘manipulative medicine’:de OR ‘craniosacral therapy’:de OR ‘massage’:de OR ‘orthopedic manipulation’:de OR ‘soft tissue therapy’:de OR ‘physical activity’:de OR ‘jogging’:de OR ‘jumping’:de OR ‘lifting effort’:de OR ‘running’:de OR ‘stretching’:de OR ‘swimming’:de OR ‘walking’:de OR ‘weight lifting’:de OR ‘nerve stimulation’:de OR ‘electroacupuncture’:de OR ‘functional electrical stimulation’:de OR ‘magnetic stimulation’:de OR ‘neuromuscular electrical stimulation’:de OR ‘electrostimulation therapy’:de OR ‘pulsed radiofrequency treatment’:de OR ‘cryotherapy’:de OR ‘traction therapy’:de OR physiotherap*:ab,ti OR ‘physical therap*’:ab,ti OR ‘kinesiotherap*’:ab,ti OR ‘exercis*’:ab,ti OR ‘manipulat*’:ab,ti OR ‘mobilization*’:ab,ti OR ‘manual therap*’:ab,ti OR ‘motor train*’:ab,ti OR ‘stabilization*’:ab,ti OR ‘training’:ab,ti OR ‘running’:ab,ti OR ‘jogging’:ab,ti OR ‘swimming’:ab,ti OR ‘walking’:ab,ti OR ‘electric* stimulat*’:ab,ti OR ‘pulsed radiofrequency’:ab,ti OR ‘electroacupuncture’:ab,ti OR ‘magnetic stimulat*’:ab,ti OR ‘craniosacral therap*’:ab,ti OR ‘massage’:ab,ti OR ‘danc*’:ab,ti OR ‘tai chi’:ab,ti OR ‘tai ji’:ab,ti OR ‘yoga’:ab,ti OR ‘qigong’:ab,ti OR ‘stretching’:ab,ti OR ‘physical conditioning’:ab,ti OR ‘soft tissue therap*’:ab,ti OR ‘cryotherap*’:ab,ti OR ‘traction’:ab,ti OR ‘resistance train*’:ab,ti OR ‘muscle energy technique’:ab,ti OR ‘endurance train*’:ab,ti OR ‘physical activit*’:ab,ti OR ‘physical fitness’:ab,ti OR ‘strengthen*’:ab,ti) AND (‘clinical trial’/exp OR ‘clinical trial’/de OR ‘randomized controlled trial’/exp OR ‘randomized controlled trial’/de OR ‘randomization’/exp OR ‘randomization’/de OR ‘single blind procedure’/exp OR ‘single blind procedure’/de OR ‘double blind procedure’/exp OR ‘double blind procedure’/de OR ‘crossover procedure’/exp OR ‘crossover procedure’/de OR ‘placebo’/exp OR ‘placebo’/de OR ‘prospective study’/exp OR ‘prospective study’/de OR ‘randomi?ed controlled’ NEXT/1 trial* OR rct OR ‘randomly allocated’ OR ‘allocated randomly’ OR ‘random allocation’/exp OR ‘random allocation’ OR allocated NEAR/2 random OR single NEXT/1 blind* OR double NEXT/1 blind* OR (treble OR triple) NEAR/1 blind* OR placebo*) AND [english]/lim) AND (2012:py OR 2013:py OR 2014:py OR 2015:py OR 2016:py OR 2017:py)

CENTRAL

((SU ‘low back pain’ OR TI ‘low* back pain’ OR AB ‘low* back pain’) OR (SU ‘back muscle*’ OR TI ‘back muscle*’ OR AB ‘back muscle*’) OR (SU ‘back injur*’ OR TI ‘back injur*’ OR AB ‘back injur*’) OR (SU ‘spinal injur*’ OR TI ‘spinal injur*’ OR AB ‘spinal injur*’) OR (SU ‘lumbosacral region’ OR TI ‘lumbosacral region’ OR AB ‘lumbosacral region’) OR (SU backache OR TI backache OR AB backache) OR (TI dorsalgia OR AB dorsalgia) OR (TI ‘facet joints’ OR AB ‘facet joints’) OR (TI ‘lumbar pain’ OR AB ‘lumbar pain’) OR (TI lumbago OR AB lumbago) OR (TI oswestry OR AB oswestry) OR (TI ‘roland morris’ OR AB ‘roland morris’)) AND ((SU ‘physical therap*’ OR TI ‘physical therap*’ OR AB ‘physical therap*’) OR (SU ‘physiotherap*’ OR TI ‘physiotherap*’ OR AB ‘physiotherap*’) OR (SU ‘kinesiotherap*’ OR TI ‘kinesiotherap*’ OR AB ‘kinesiotherap*’) OR (SU ‘joint mobilization*’ OR TI ‘joint mobilization’ OR AB ‘joint mobilization’) OR (SU ‘electric stimulation therapy’ OR TI ‘electric* stimulat*’ OR AB ‘electric* stimulat*’) OR (SU ‘pulsed radiofrequency’ OR TI ‘pulsed radiofrequency’ OR AB ‘pulsed radiofrequency’) OR (SU ‘transcutaneous electric nerve stimulation’ OR TI ‘transcutaneous electric nerve stimulation’ OR AB ‘transcutaneous electric nerve stimulation’) OR (SU exercis* OR TI exercis* OR AB exercis*) OR (SU ‘resistance train*’ OR TI ‘resistance train*’ OR AB ‘resistance train*’) OR (TI training OR AB training) OR (SU manipulat* OR TI manipulat* OR AB manipulat*) OR (SU ‘soft tissue therapy’ OR TI ‘soft tissue therap*’ OR AB ‘soft tissue therap*’) OR (SU ‘physical conditioning’ OR TI ‘physical conditioning’ OR AB ‘physical conditioning’) OR (SU running OR TI running OR AB running) OR (SU jogging OR TI jogging OR AB jogging) OR (SU swimming OR TI swimming OR AB swimming) OR (SU swimming OR TI swimming OR AB swimming) OR (SU cryotherapy OR TI cryotherap* OR AB cryotherap*) OR (SU traction OR TI traction OR AB traction) OR (SU ‘craniosacral therapy’ OR TI ‘craniosacral therap*’ OR AB ‘craniosacral therap*’) OR (SU massage OR TI massage OR AB massage) OR (SU ‘physical activit*’ OR TI ‘physical activit*’ OR AB ‘physical activit*’) OR (SU ‘nerve stimulation’ OR TI ‘nerve stimulation’ OR AB ‘nerve stimulation’) OR (SU electroacupuncture OR TI electroacupuncture OR AB electroacupuncture) OR (TI stimulat* OR AB stimulat*) OR (TI mobilization OR AB mobilization) OR (TI ‘manual therap*’ OR AB ‘manual therap*’) OR (TI ‘motor train*’ OR AB ‘motor train*’) OR (TI stabilization OR AB stabilization) OR (TI danc* OR AB danc*) OR (TI (‘tai chi’ OR ‘tai ji’) OR AB (‘tai chi’ OR ‘tai ji’)) OR (TI yoga OR AB yoga) OR (TI stretching OR AB stretching) OR (TI qigong OR AB qigong) OR (TI ‘endurance train*’ OR AB ‘endurance train*’) OR (TI ‘muscle energy technique’ OR AB ‘muscle energy technique’) OR (TI ‘physical fitness’ OR AB ‘physical fitness’) OR (TI strengthen* OR AB strengthen*)) AND (TX ‘randomized controlled trial*’ OR TX ‘double blind’ OR TX random*)

Limiters - Published Date: 20120101-20161231; Language: English

Funding Statement

No funding was provided for the development of this manuscript. I certify that neither I nor the co-investigators on this research project have any conflicts of interest related to this work.

Disclosure statement

No potential conflict of interest was reported by the authors.