Trial design and registration
Two-parallel-group (allocation ratio 1:1), single-centre, sham-controlled, randomised blinding feasibility study [18]. This trial protocol is reported in accordance with the SPIRIT 2013 guidance for protocols of clinical trials (Additional file 1) [19]. Our trial was preregistered on ClinicalTrials.gov: NCT05778396. Figure 1 presents the trial design.
Trial setting
Data will be collected, and interventions delivered at a community-based chiropractic research clinic in Zurich, Switzerland—CHIROMED Praxis im Seefeld (Dufourstrasse 101, 8008 Zurich). This research clinic actively collaborates and is affiliated with the EBPI-UWZH Musculoskeletal Epidemiology Research Group, within the Epidemiology, Biostatistics and Prevention Institute of the University of Zurich and the University Spine Centre Zurich of Balgrist University Hospital.
Participants
Eligibility criteria
We will include participants aged 18 years or older with or without experience of SMT or current low back pain, to maximise feasibility of this study. Candidate participants—“candidates” hereafter—will be excluded if they have serious spinal pathology (e.g., spinal fracture, cancer or infection), a history of lumbar spine surgery, are currently under care or in consultation with a specialist, chiropractor, physiotherapist, or osteopath for current low back pain, are a manual medicine health care provider (i.e., chiropractor, physiotherapist, osteopath, massage therapist, manual medicine trained physician), have a serious comorbidity preventing them from attending the research clinic and receiving the interventions, are pregnant or breastfeeding, are involved in pending litigation related to back pain, or are already participating in another research study related to back pain.
Recruitment, screening, and informed consent
Candidates will be recruited by convenience and purposive sampling via word-of-mouth, snowballing, and mass email advertisements through the University of Zurich and Balgrist University Hospital mailing lists. Candidates will be approached for interest directly by study team members in and around the Zurich area. Candidates that express preliminary interest will receive an invitation email with study details. Those expressing willingness to consider participation will be able to pre-book their study visit appointments and will receive a REDCap study eligibility survey [20]. Eligible candidates will be directed to review the study participant information form and complete an electronic signature in REDCap to express interest and willingness to participate. The study information and consent form masks the blinding feasibility study objectives and states that the aim of the study is to find out whether two SMT interventions are “practical and acceptable” for a future study in patients with back pain. Active SMT is described in the study intervention form as an “active or real treatment, with an unknown exact benefit.” Sham-control SMT is described as “other manual therapy procedure that is a comparison or control treatment that is not known to have a benefit.”
At the first clinic appointment, the trial coordinator reviews the participant information form with the candidate, answers any questions, confirms candidate understanding and agreement to participate, and requests the candidate to sign the official study informed consent form. Candidates completing these steps are deemed enrolled study participants.
Interventions
SMT is conceptualized as a pragmatic hands-on treatment directed towards the spine that includes manipulation, mobilization, or both [16]. Spinal manipulation is manual therapy applied to the spine that involves a high velocity, low amplitude impulse or thrust applied at or near the end of the passive range of motion of the spinal joints [21]. Spinal mobilization is manual therapy involving passive movement applied to a spinal region or segment that incorporates movements within spinal passive range of motion with the aim of achieving a therapeutic effect [22]. Mobilization includes spinal traction, which is application of an intermittent or continuous distractive force within the passive range of motion.
Up to 10 licensed chiropractic treating clinicians (“treating clinicians” hereafter) with at least three years of clinical experience will be trained in the delivery of active and sham-control SMT interventions according to prespecified intervention descriptions and standard operating procedures (see Additional File 2A) for consistent study interactions with participants and high-fidelity intervention delivery. Treating clinicians are distinguished from outcome assessors—the latter being clinicians or clinicians-in-training strictly in charge of outcome assessment and brief, standardised interactions with participants within the RCT setting (see Additional File 2A for details).
Participants will receive two sessions of the assigned intervention, one to two weeks apart, to maximise retention and minimise the risk for information loss (missing data). Participants who are lost to follow-up after study visit one will be asked for reasons for discontinuation. Due to our blinding feasibility primary objective, concomitant care is not restricted outside of the trial.
Active SMT protocol
Active SMT includes (1) side-lying lumbar spine manipulation, (2) prone lumbar spine mobilisation, and (3) prone thoracic spine manipulation—all delivered with therapeutic intent. These components are chosen based on their acceptability in clinical practice and their validated sham-control counterparts [23]. The treating clinician will deliver (1) a high-velocity, low-amplitude (HVLA) thrust to the L4-L5 or L5-S1 spinal motion segments. Side-lying lumbar spine manipulation is performed bilaterally, with the treating clinician choosing any suitable technique and with or without occurrence of the characteristic audible joint cavitation associated with spine manipulation. The treating clinician then delivers (2) prone lumbar mobilisation by placing the contact hand and applying downward pressure over L4-L5 or L5-S1 with the other hand guiding a manual flexion-distraction piece through 3 sets of 5 full range oscillations. The treating clinician then delivers (3) prone thoracic manipulation by applying a HVLA thrust with therapeutic intent in a posterior-to-anterior direction at the T5-T6 or T6-T7 spinal motion segments. Details on the active SMT protocol—consistent with the 12-item TIDieR checklist [24]—are found in Additional file 2B.
Sham-control SMT protocol
Sham-control SMT involves: (1) control side-lying lumbar spine manipulation, (2) control prone lumbar spine mobilisation, and (3) control prone thoracic spine manipulation—all performed without therapeutic intent. The three included sham-control procedures are informed by previous sham-control validation work in RCT settings [23]. (1) Sham-control side-lying lumbar spine manipulation is operationalized as the application of a low-velocity broad push manoeuvre to the gluteal region with an inferior-to-superior non-therapeutic line of drive [23]. The treating clinician then delivers (2) sham-control prone lumbar mobilisation, consisting of a manual manoeuvre involving 3 sets of 10 minimal oscillations (0 to ± 2°) of the flexion-distraction piece with light touch over the lumbar spine region. The treating clinician then delivers (3) sham-control prone thoracic mobilisation consisting of two-handed bilateral superior-to-inferior scapulae push manoeuvres [23]. Additional file 2C details the sham-control SMT protocol [25].
Outcomes
Primary outcome
Our primary outcome is participant blinding immediately after each intervention session, as measured by the Bang blinding index (BI) [14, 26]. The blinding assessment question elaborates “As mentioned at the beginning of the study, one of the two study interventions had unknown benefits (genuine treatment) and the other intervention was not known to be beneficial (control treatment)” and then asks “Which treatment do you believe you received?”. There are five response options: “Strongly believe I received the genuine treatment”, “Somewhat believe I received the genuine treatment”, “Somewhat believe I received the control treatment”, “Strongly believe I received the control treatment”, and “I do not know which treatment I received”. Bang BI estimates (between − 1 to + 1, with 0 suggesting satisfactory blinding—random guess) can be interpreted as the proportion of correct guesses beyond chance within an intervention arm. From arm-specific estimates, a sum Bang BI can be calculated measuring the difference in proportions in the same guess (0 suggesting an equal proportion of participants in both arms believed they received active treatment). We deem an absolute arm-specific value score of ≤ 0.3 (i.e., -0.3 to 0.3) as suggestive of satisfactory blinding. Since arm-specific value scores within this range are desirable, we discuss arm blinding scenario interpretations taking into account arm-specific and sum Bang BIs [14].
Secondary outcomes
Our secondary outcomes are a second BI metric—the James BI [27]—which provides a measure of study-level blinding. The James BI is a modification of the kappa statistic that measures disagreement beyond chance and returns a value between 0 and 1. A value of 1 is returned when all responses are “don’t know” (complete blinding), 0 is returned when all responses are correct (complete unblinding), and 0.5 when responses are randomly distributed (50% correct, 50% incorrect). Secondary outcomes measured after each intervention session include: outcome assessor blinding (Bang and James BIs), and factors influencing perceived intervention assignment among participants and outcome assessors (thematic analysis [28]; answers to the open-ended questions “Why do you believe you received this treatment?”, and “Why do you believe this participant received this treatment?”, respectively). Other outcomes—included to mask the study objective from participants—were lumbar spine range of motion (ROM, maximum active total flexion and extension), self-rated general health, satisfaction with care, self-rated back flexibility, pain intensity and function (NRS, 0–10). In addition, participant-reported credibility of the interventions and expectancies are measured after the second intervention session with a modified and adapted version of the Credibility/Expectancy Questionnaire (CEQ) [29]. Lastly, to complement the main study objectives and provide exploratory information for the future trial, treating clinician outcomes are also collected after each session and include clinician-reported intervention component delivery fidelity and self-rated quality of delivery relative to the protocol. See Table 1 for the schedule of study procedures related to screening, enrolment, interventions, and assessment.
Table 1. Schedule of screening, enrolment, interventions, and assessments
Trial event
|
Screening
|
Enrolment
|
Study visit 1
|
Study visit 2
|
Study visit period
|
|
|
Pre
|
Post
|
Pre
|
Post
|
Time point
|
-1
|
0
|
0
|
0
|
+2 to 14d
|
+2 to 14d
|
Enrolment
|
|
|
|
|
|
|
Inclusion and exclusion criteria (eligibility)
|
X
|
|
|
|
|
|
Patient information and informed expression of interest
|
X
|
|
|
|
|
|
Demographics and study entry form
|
X
|
|
|
|
|
|
Informed consent in-person signature at clinic
|
|
X
|
|
|
|
|
Randomisation
|
|
X
|
|
|
|
|
Intervention assessments
|
|
|
|
|
|
|
Interventions delivered
|
|
|
X
|
X
|
Intervention fidelity (treating clinicians)
|
|
|
|
X
|
|
X
|
Participant tolerability (treating clinicians)
|
|
|
|
X
|
|
X
|
Quality of intervention delivery (treating clinicians)
|
|
|
|
X
|
|
X
|
Assessments
|
|
|
|
|
|
|
Range of motion (outcome assessor)
|
|
|
X
|
X
|
X
|
X
|
Self-rated health
|
|
|
X
|
X
|
X
|
X
|
Pain intensity
|
|
|
X
|
X
|
X
|
X
|
Back function
|
|
|
X
|
X
|
X
|
X
|
Satisfaction with care
|
|
|
|
X
|
|
X
|
Blinding assessment – participants
|
|
|
|
X
|
|
X
|
Blinding assessment – outcome assessors
|
|
|
|
X
|
|
X
|
Factors influencing beliefs about intervention assigned
|
|
|
|
X
|
|
X
|
Intervention component fidelity (treating clinicians)
|
|
|
|
X
|
|
X
|
Quality of intervention delivery (treating clinicians)
|
|
|
|
X
|
|
X
|
Credibility/expectancy of interventions
|
|
|
|
|
|
X
|
Abbreviations: d = days; wk = weeks
Sample size consideration
A precision-based approach has been used to decide sample size using the primary outcome measure of the Bang BI estimate (i.e., width of the 95% CI) rather than statistical hypothesis testing/power given the purpose of this blinding feasibility randomised trial [30]. We estimate the Bang BI within each intervention arm and present a 95% CI for each mean BI point estimate. For a sample size of 26 participants per arm, the 95% CI is the observed BI estimate ± 0.225 points (0.45 points width of the 95% CI) for the primary outcome measure of the intervention arm-specific Bang BI, according to Thompson’s method (Eq 1), as described by Landsman and colleagues [31]. Conservatively allowing for attrition of participants of up to 15%, we aim to recruit at least 30 participants per arm (at least 60 participants in total) for this blinding feasibility RCT.
If the minimum recruitment targets for the desired precision are met, and trial participation interest is sufficient to aim for a larger study size, we may enrol and conduct the trial in up to a maximum of 50 participants per arm (100 participants in total). This additional enrolment will yield more precise and informative estimates of the arm-specific Bang BI primary outcome.
Randomisation
Allocation sequence generation
A computer-generated randomisation sequence will be generated by a trial statistician at the EBPI and used to randomly allocate participants to active SMT or sham-control SMT (1:1 ratio). Randomisation will be stratified by SMT experience as a patient or recipient in the past three months, and low back pain in the past four weeks (3 or greater on a 0 out of 10, NRS), and blocked with randomly varying block sizes, which will not be disclosed to ensure concealment of allocation of participants, outcome assessors, data analysts, and investigators.
Concealment mechanism and implementation
Participants will be allocated to intervention arms through a concealed, central, web-based randomisation system via REDCap, once eligibility is determined, informed consent will be obtained and baseline data will be entered in an electronic case report form. Treating clinicians will assign participants to one of the two SMT interventions—coded “A” and “B”—at study visit one, immediately before delivering the intervention. They will perform randomisation in a private study treatment room with an electronic device and shield the screen from the participants’ view.
Blinding
Per protocol, participants, outcome assessors, data analysts, and investigators will be blinded to intervention assignment after randomisation. This will be achieved through appropriate trial processes keeping treating clinicians and outcome assessors in separate spaces in the study clinic and emphasising the importance of no discussion of study aspects between clinicians, restricting REDCap user right privileges, and by not breaking the code of the intervention variable until blinded data analyses are completed. Due to the physical nature of both the active and sham-control SMT interventions, treating clinicians will not be able to be blinded to the interventions they are delivering. Yet, they will be trained and asked not to disclose the assigned intervention, nature of the interventions, nor trial objectives to trial participants or other members of the trial team.
Data collection and monitoring
The inclusion and exclusion criteria (eligibility), patient information and informed expression of interest forms, will be completed online, prior to the first study visit. The informed consent form, study entry survey, and pre- and post- intervention assessments will be completed face to face. Trial coordinators will provide participants with a tablet or laptop with the REDCap survey [20] in the study clinic waiting area. Trial coordinators will have completed Good Clinical Practice (GCP) training prior to the start of the trial. The study entry survey includes questions on height, weight, gender, experience with SMT (in the past three months and lifetime), low back pain in the past four weeks and average pain in the past four weeks. The following outcomes with their corresponding instruments will be captured:
- Range of motion of the lumbar spine: measured by placing a measuring mobile phone at T12 with the participant standing, using the iOS application Measure® (iOS version 16.0.2, iPhone® model X, Apple Inc., California)—a suggested valid and reliable method [32–34]
- Self-rated health: measured with an item from the RAND 36 Item Short Form Survey Instrument [35], with acceptable psychometric properties
- Pain intensity: measured with the NRS [36]—a widely used method for this purpose
- Back function: measured with selected items from the International Fitness Scale [37] and the Cornell Musculoskeletal Discomfort Questionnaire [38]
- Satisfaction with care: measured by degree of agreement with the statement “I am satisfied with the treatment I received today”, with five Likert-type answers—“Strongly agree”, “Agree”, “Uncertain”, “Disagree”, and “Strongly disagree”
- Beliefs about intervention assignment: measured with proposed blinding questionnaires [14]—a standardized format for blinding assessment
- Factors influencing beliefs about intervention assignment: assessed with the open-ended questions “Why do you believe you received this treatment?” (participants), and “Why do you believe this participant received this treatment?” (outcome assessors)
- Credibility/expectancy of active and control interventions: measured with the CEQ [29]
- Intervention fidelity: assessed by asking treating clinicians to record in case report forms information pertaining to participant tolerance, potential procedure alteration, and occurrence of cavitation sounds immediately after intervention delivery
- Self-rated quality of intervention delivery: measured with the question “What number do you feel best describes the quality of the intervention delivery for this participant relative to the intervention protocol you were trained to deliver?” with 11 possible answers on a 0 to 10 NRS
Adverse events
All adverse events are appropriately documented and reported following the Clinical Trials Ordinance applicable to the Federal Act on Research involving Human Beings.
Statistical methods
Blinding analyses will be conducted using an intention-to-treat approach with the R package BI, version 1.1.0 [39,40]. The current implementation of this package supports both Bang and James BI approaches, the presence of two intervention arms, and three intervention assignment beliefs by participants (i.e., “Genuine”, “Control”, and “Do not know”). Information from the 5-level blinding assessment items will be descriptively analysed and tabulated with numbers and percentages. To evaluate the impact of SMT experience on blinding feasibility of participants and outcome assessors, the responses from the blinding assessment will be described as counts and percentages, by levels of SMT experience. Within- and between-group changes in pain intensity and lumbar ROM will be described, although no formal statistical testing is performed given that these outcomes were collected to mask the study objective from participants and are not relevant to the blinding feasibility assessment. Factors contributing to perceptions about intervention arm assignment will be interpreted using qualitative thematic analysis [28], with three investigators (blinded to assigned interventions) independently assessing responses and grouping them thematically by consensus.
To assess the impact of possible deviations from allocated intervention (e.g., potential lack of adherence or human errors by treating clinicians), an as-treated analysis will be conducted, if applicable and appropriate, to assess blinding outcomes.