ADvAnced PhysioTherapy in MuSculosKeletal Triage: Investigating prognostic factors, healthcare utilisation and clinical outcomes (ADAPT MSK) - a cohort study protocol.

Background: Clinical specialist physiotherapist-led musculoskeletal triage clinics were introduced nationally in Ireland in 2011 to improve patient care and reduce waiting times for secondary care orthopaedics and rheumatology. Evidence has shown them to be effective in reducing waiting lists, however there are currently no data on longitudinal patient outcomes following clinic attendance. The primary aim of this cohort study is to identify predictors of pain and function outcomes up to one year following musculoskeletal triage review. Secondary aims include measuring self-reported use of healthcare resources over the 12-month follow-up period and to explore musculoskeletal phenotypes based on established prognostic factors for musculoskeletal pain. This is a prospective cohort study. Methods: ADvAnced PhysioTherapy in MuSculosKeletal Triage (ADAPT MSK) will recruit a cohort of 252 adults through musculoskeletal triage clinics across five secondary care sites in Ireland. The STrengthening the Reporting of Observational studies in Epidemiology (STROBE) guidelines will be adhered to for future reporting. Adults (≥ 18 years old) attending physiotherapist-led musculoskeletal triage clinics with musculoskeletal pain, who do not require surgical or consultant-led medical care will be considered for participation. Participant demographics, health literacy, healthcare utilisation, and self-report questionnaires on pain, function, musculoskeletal health, musculoskeletal risk stratification, fear of movement, and psychological distress will be obtained at baseline, with follow-ups at three, six, and 12 months. Primary outcomes are pain intensity and function. Secondary outcomes include musculoskeletal risk stratification status, musculoskeletal health, healthcare utilisation, and work-related factors. Descriptive statistics will be used to profile the participants and predictors of outcome will be assessed using multivariable linear regression. Musculoskeletal phenotypes will be explored using latent class analysis. Results: Results will be disseminated via peer-reviewed journal publication and presentation at national and international conferences. Engagement with a public patient involvement (PPI) panel will explore dissemination strategies for public and service user engagement.


Introduction
Musculoskeletal (MSK) pain, which includes conditions such as low back pain, neck pain or osteoarthritis is recognised as one of the leading causes of disability worldwide 1 , resulting in increased healthcare expenditure and longer waiting times for orthopaedic and rheumatology outpatient services 2,3 .Adult orthopaedic services represent the largest waiting list in Ireland (June 2023) with a total of 64,867.Up to 25% of patients are waiting more than 12 months for orthopaedic (22%) and rheumatology (25%) appointments in secondary care 4 .
In 2011, to reduce outpatient Orthopaedic and Rheumatology waiting times in Ireland, the Health Service Executive (HSE) National Clinical Programmes for Trauma and Orthopaedics (NCPTOS), and Rheumatology (NCPR) established the National MSK Triage Initiative, consisting of 24 clinical specialist physiotherapist (CSP) posts in 18 Acute Hospital sites nationwide.In these MSK Triage clinics, CSPs triage patients on outpatient orthopaedic and rheumatology waiting lists, who are unlikely to require consultant care, onto appropriate care pathways.In a national audit, over 80% of patients presenting to MSK-triage clinics in Ireland were managed independently by the CSP, with 71% discharged at their initial appointment 5 and 23% referred to physiotherapy 6 .From 2012 to 2018, 125,852 patients on orthopaedic and rheumatology waiting lists were managed through MSK triage services 7 .Access to primary care physiotherapy also presents a barrier to patients, with 56,200 on primary care waiting lists and 22% (12,502) waiting greater than one year to access primary care physiotherapy services in 2022 8 .Longer waiting times to access physiotherapy can negatively affect patients' quality of life, psychological wellbeing, healthcare utilisation, health outcomes and economics 3,9,10 .
Whilst the National MSK Triage Initiative has been successful in reducing acute hospital outpatient orthopaedic and rheumatology waiting lists, the high discharge rate of 71% at initial appointment 5 warrants further examination to explore the patient journey and potential reasons why patients are not referred to the right service at the right time, in line with the Irish government health reform plan (Sláintecare) 11 .It is possible that suboptimal access to primary care services, may be influencing referrer behaviour and decision making.
Several predictors of pain and functional outcomes in MSK conditions have previously been identified, including baseline function, pain intensity, mental well-being, co-morbidities, age, body mass index (BMI), duration of symptoms, workers' sick leave, education level 12 and altered pain processing 13 ; which can also predict non-response to physiotherapy 14 .Recently, MSK core outcome sets, and prognostic stratification tools (such as the Subgroups for Targeted Treatment Back (STarT Back) and Subgroups for Targeted Treatment MSK (STarT MSK)), have been developed, based on established prognostic factors 12,15 , and validated to identify earlier, those at risk of developing persistent MSK pain 16,17 .Research to date has shown that MSK triage is an effective waiting list initiative with good service user and healthcare professional satisfaction 5,7,[18][19][20][21] .
However, currently, patient outcomes, prognostic stratification, and predictors of outcome up to 1-year later have not been consistently studied in patients attending MSK triage clinics in Ireland or internationally.

Objectives
The primary aim of this prospective, cohort study is to identify predictors of clinical outcome (pain and function) at three-, six-, and 12-months post MSK-triage appointment.
Secondary aims are to: 1. Measure self-reported use of healthcare resources over the 12-month follow-up period post MSK-triage appointment.

Study design
ADAPT MSK is a prospective, observational, cohort study.The STROBE standardised reporting guidelines will be used to guide the reporting of this study 22 .Adults with MSK pain attending CSP-led MSK triage clinics will be recruited from five sites across Ireland.Baseline assessment will consist of baseline demographics, work-related factors, healthcare utilisation and self-report questionnaires..

Setting
This study will be based in MSK Triage clinics across five urban and regional secondary care sites in Ireland.These clinics are run by CSPs with more than five years clinical experience and the majority achieving a postgraduate MSc or PhD degree, in the field of MSK physiotherapy 5 .They provide expert assessment, diagnosis and education to patients and identify the most appropriate management pathway for patients with MSK disorders.Patients on orthopaedic and rheumatology waiting lists, deemed unlikely to require orthopaedic surgeon or rheumatology consultant care are triaged to these MSK triage clinics, which improves service efficiency by reducing secondary care waiting lists and directing patients towards the appropriate care pathway 7 .

Participants
A consecutive sample of patients presenting to orthopaedic and rheumatology MSK-triage clinics with pain will be recruited.Participants will be eligible if they are aged 18 years or over, are triaged for non-consultant care at one of the five participating MSK triage services across Ireland, and have sufficient English language proficiency for the completion of selfreported questionnaires.
Patients will be ineligible to participate if they've been triaged by the CSP for orthopaedic surgical or rheumatologist assessment, are unable to communicate in English (written and spoken word), along with those who present with clinical indicators of suspected 'red flag' pathology (e.g.recent trauma with significant injury; acute, red, hot, or swollen joints; suspected fracture; joint infection; cancer) 24 ; or a diagnosed systemic inflammatory MSK condition (such as rheumatoid arthritis) or a diagnosis of dementia or terminal illness.

Sample size
The estimated sample size is based on our primary aim.Approximately 18 predictor variables will be included in univariate analysis and with 10 events required per predictor variable 25 , a sample of 180 participants is required.An additional 40% has been added to allow for drop-out at the 12-month follow-up, resulting in a final sample size of 252.

Recruitment and data collection
The MSK triage physiotherapist will identify and screen prospective participants for eligibility.If eligible, they will provide a participant information leaflet, briefly explain the objective of the study, and obtain written consent to be contacted by the primary investigator (FC).This allows the primary investigator to contact prospective participants to answer any questions about the study and if interested in participating, obtain informed written or electronic consent.
Once recruited, each participant will undergo a baseline assessment with the primary investigator, capturing participant demographics and healthcare utilisation, via Microsoft Teams or telephone, depending on participant preference.Thereafter, participants will complete a number of self-report questionnaires based on established prognostic factors i.e., baseline function, pain intensity, mental wellbeing, symptom duration, fear avoidance/catastrophising, quality of life/self-efficacy, widespread pain, age, co-morbidities, work absence duration, and education level 15,26,27 .This data will be collected through Research Electronic Data capture (REDCap) software 28,29 , hosted at RCSI, on their personal device, or via posted paper questionnaires.Demographic information will include participant gender, age, level of education, presenting MSK complaint, duration of symptoms, number of MSK pain sites, previous physiotherapy/surgery for presenting complaint, and work-related factors (employment status, work classification and duration of any work absence).Co-morbidities will be identified from a list of 12 comorbid conditions, informed by the National Institute of Clinical Excellence (NICE) indicator for multi-morbidity in primary care 30 .Health literacy will be explored using the singleitem literacy screener 31 .Healthcare utilisation will be recorded using a modified version of the Managing of OSteoArthritis In ConsultationS (MOSAICS) trial questionnaire 32 , which captures advice and information received about their condition, selfmanagement, prescribed medications, aids and appliances, private/public health services (e.g., physiotherapy, GP, nursing, occupational therapy, podiatry), treatments, and investigations.
Self-report questionnaires will include the Musculoskeletal Health Questionnaire (MSK-HQ) 33 , STarT MSK tool 17 , and Patient Specific Functional Scale (PSFS) 34 to assess functional and MSK health status; pain intensity through the Numerical Pain Rating Scale (NPRS) 35,36 ; fear of movement through the 11-item Tampa Scale for Kinesiophobia 37 and psychological distress via the Hospital Anxiety and Depression scale (HADS) 38 .
All participants recruited in two sites (Beaumont Hospital and Tallaght University Hospital) will be invited to participate in a once-off baseline physical examination, consisting of grip strength examination, neurological exam, and quantitative sensory testing (Table 1).
Pain hypersensitivity, measured by quantitative sensory testing, has been shown to be a predictor of worse outcome (pain and disability) at follow-up across multiple MSK conditions (e.g., osteoarthritis, low back pain, whiplash, post-operative pain) and different body sites (e.g., hip, knee, low back, shoulder and neck) 13 .Quantitative sensory testing uses standardised testing protocols of somatosensory nerve function, to investigate potential underlying pain mechanisms 39,40 .The International Association for the Study of Pain (IASP) task force clinical criteria and grading system for nociplastic pain involves a stepwise approach to differentiate between predominant nociceptive, neuropathic or nociplastic pain 41 , which, in conjunction with the NeuPSIG guidelines on neuropathic assessment 42 will be used to categorise participants' dominant pain phenotype (Figure 1).A quantitative sensory testing protocol including pressure pain thresholds (PPT), dynamic mechanical allodynia, pinprick, temporal summation and cold pain thresholds will be used to assess pain sensitivity in accordance with IASP and NeuPSIG grading systems 41,42 .
Grip strength is regarded as a biomarker of current health status and has been adopted as a singular indicator of overall body strength [43][44][45] .Grip strength will be assessed isometrically using a calibrated Jamar Plus Digital dynamometer following a standard protocol 46 .

Follow-up assessment
The primary investigator will contact participants at three, six, and 12 months via Microsoft Teams or telephone to collect healthcare utilisation data and work-related factors (employment status, work classification and duration of any work absence).Self-report questionnaires (MSK-HQ, STarT MSK, Patient Specific Functional scale, and NPRS) will be sent electronically via REDCap software or via post.Any participant withdrawals or loss to follow-up will be recorded.

Outcomes
The primary outcomes of interest are pain intensity (NPRS) and function (PSFS).Secondary outcomes are musculoskeletal risk stratification status (STarT MSK), musculoskeletal health

Statistical analysis
Pseudonymised data will be stored in an encrypted and password protected folder on the study's SharePoint site in RCSI.
Secure and encrypted access to the Microsoft SharePoint folder will be assigned to data controllers only.Descriptive statistics will be used to profile the characteristics of the cohort at baseline, three, six, and 12 months.Changes at three, six, and 12 months will be analysed using repeated measures multivariable regression.All models will be adjusted for potential confounding factors, checking for interactions and collinearity.Multivariable linear regression will be used to identify baseline predictors of pain and function outcomes at three, six, and the primary timepoint of 12 months.Variables included in the multivariable regression model will be selected if deemed clinically significant, or, if they have a univariable p-value of <0.2.Latent Class Analysis will be undertaken to explore underlying pain phenotypes within the cohort at baseline, three, six, and 12 months based on a range of observed categorical variables.Statistical significance will be inferred when the P value is <0.05.STATA 17 statistical software (StataCorp, College Station, Tx, USA) will be used for statistical analyses.

Dissemination
Findings from this study will be disseminated via peerreviewed journal publication and presentation at national and international conferences.Engagement with a public patient involvement (PPI) panel will explore dissemination strategies for public and service user engagement.

Study status
Data collection commenced in December 2022, with study completion anticipated in November 2024.

Discussion
The burden of MSK disorders is increasing exponentially worldwide, resulting in significant pressure on healthcare systems.People with MSK pain who present to their GP in Ireland are faced with difficulties accessing first-line public services, such as primary care physiotherapy and subsequently specialised orthopaedic and rheumatology services.To address secondary care waiting lists and improve service efficiency, the National MSK Triage Initiative, MSK triage clinics, run by CSPs under the clinical governance of Orthopaedic and Rheumatology Consultants commenced in Ireland in 2011, and has demonstrated success as a waiting list initiative.However, high discharge rates and onward referral to primary care physiotherapy following MSK triage suggest that these patients may have been managed more appropriately in primary care if sufficiently resourced.Currently, the patient journey and long-term outcomes following their MSK triage attendance are unknown.This longitudinal cohort study aims to identify predictors of pain and function outcomes up to 1 year following MSK triage attendance; measure individuals' self-reported use of healthcare resources and explore MSK phenotypes based on identified prognostic factors.This research has the potential to inform future needs within primary care for those with MSK conditions, as well as the implementation of pathways from primary to secondary care orthopaedics and rheumatology, ensuring that patients receive the 'right care, at the right place, at the right time' in line with SláinteCare principles 11 .

Introduction
Overall could include more information to expand on the gap this research is addressing/what is known already and why this research is important While the background contains interesting info on the MSK triage initiative it's not clear from the Introduction why this is an important question?Why is knowing predictors of outcome for this particular patient cohort at 1 year is important?What are the implications of better predicting outcomes?Could it be used for early identification of those at risk of persistent pain and subsequent prioritisation by physiotherapy services?How will this information be used to improve or better current MSK services in Ireland?(some mention of this in Discussion) Also, while the authors refer to previous SRs of predictors of pain and function in MSK conditions, it would be useful to know what type of patient population these relate to (primary care or secondary care).It is stated that MSK triage clinics have not been studied, which is true, however this cohort are essentially very similar to cohort attending secondary care elective othopaedic clinics, what are the main findings from these studies?

Methods
Study Design-Follow-up at 3, 6 and 12 months-is there a pre-specified primary end-point?
How many are estimated to undergo a baseline physical exam?Will there be sufficient data from these two sites to carry out the anticipated data analysis e.g.latent class analysis or multi-variable analysis?Or is this info solely to categorise the pain phenotype?Should make reference to this sub-group in the data analysis section.
Pain phenotyping-using IASP criteria and NeuPSIG based on baseline physical exam.Is this different to the Latent Class Analysis to explore underlying pain phenotypes "based on a range of categorical variables"?Clarify this.
How soon will the physical exam be after the patient is recruited?What is the anticipated duration of this examination eg. 1 hour?
QST-Would have expected in a protocol to see more detail here in order that your testing procedures are reproducible.For example how do you intend to choose anatomical testing sites (related to the primary pain complaint?)What is the order of testing?Instruments-Device you are using for pressure pain algometry, what thermal testing devices?.What is pinprick testing?References (41, 42)

Karin Samsson
University of Gothenburg, Gothenburg, Sweden This is an important aspect to address in research, to improve on the knowledge regarding the patients that are referred for MSK triage.Furthermore, to be able to screen patients prior to referral could reduce waiting times as well as improve care on the right level.You state in the introduction that 71% of patients were discharged at their initial appointment and 23% referred to physiotherapy, which makes me wonder.The 71% who were discharged, do you think that they would have needed different care than what they received?And since this large number of patients were discharged, and not referred to physiotherapy, why do you think predictors of clinical outcome and prognostic factors for MSK pain. is important to investigate for this patients group? 1.
The aims are not consistently reported throughout the paper.

2.
You have a great number of predictors and outcomes, but I struggle to make them out clearly since they are inconsistently reported in introduction and various sections in the methods.It would be highly Regarding patients included in the study -who is responsible for deciding which patients are "unlikely to require consultant care" and based on what?

4.
Are the patients to be included in the study after the triage?When is the patient asked to participate?And why don't you include the patients in need of orthopaedic surgeon to be able to investigate predictors of clinical outcome and prognostic factors for MSK pain for them as well?Considering that 71% are discharged, it would probably be quite beneficial to know if there is a difference between these two groups.

5.
There is no formal sample size calculation.
Recruitment and data collection 6.
You state that the MSK triage physiotherapist will identify and screen prospective participants for eligibility -is this before the actual triage?I think this process needs to be 7.
clearly described (as also addressed in previous question).If I understand correctly you are collecting two written consents?What is the purpose of that?

8.
Why did you chose to have a teams or telephone data collection as well as questionnaires?9.
Inconsistent reporting of outcomes collected and through what outcome measures, order and the use of abbreviations or not.Furthermore not clear which ones are collected via self-reported questionnaires vs telephone/teams.Furthermore, not all outcomes are included in the primary or secondary outcomes.

10.
The outcome health literacy is not mentioned or described prior to data collection.11.
The whole section starting with "Pain hypersensitivity, measured by… / (Figure 1) seems more to belong in the introduction section.Outcomes 12.
Outcomes are not consistent with aims.13.

Are the datasets clearly presented in a useable and accessible format? Not applicable
Competing Interests: No competing interests were disclosed.
Reviewer Expertise: Physiotherapist led orthopaedic triage.Have not focused on the statistics part as that is not my area of expertise.
I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
I focused on the statistical analysis component of the manuscript.Here are my specific concerns regarding this manuscript.1) There is no formal sample size calculation.Rule of thumb of 10 events per variable is used.Software including PASS, SAS can actually provide a more formal sample size estimation.
2) The authors propose using repeated measures multivariable regression.I recommend being more specific in terms of whether it is random-effects model, covariance pattern model or marginal GEE model.
3) The authors have adjusted sample size to account for attrition.But missing data is not addressed in the statistical analysis plan.
4) If the authors are going to run separate models for baseline predictors of pain and function outcomes at three, six and the primary timepoint of 12 months, they should adjust the statistical significance level to adjust for inflated type I error rate.5) There is no detail provided regarding the proposed Latent Class Analysis.6) On page 4, column 2 the authors state "..... multiple MSK conditions (e.g., osteoarthritis, low back pain, whiplash, post-operative pain).It is quite possible that a large majority with postoperative pain may be lost to follow-up at 12 months.If the participant's pain is not chronic, they may drop-out of the study.

Are sufficient details of the methods provided to allow replication by others? Partly
Are the datasets clearly presented in a useable and accessible format?

Not applicable
Competing Interests: No competing interests were disclosed.
Reviewer Expertise: Biostatist I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.

Table 1 .
Overview of primary and secondary outcomes, predictor variables, and time of assessment.Total number of MSK pain sites (number/11 on body chart) NICE multi-morbidity index Single item health literacy screener (SILS) dynamometer Physical ✓ (MSK-HQ), healthcare utilisation and work-related factors (employment status, work classification, work absence ± duration).

Figure 1 .
Figure 1.Screening process for pain classification based on IASP criteria for nociplastic pain 41 and NeuPSIG grading system for Neuropathic pain 42 .
beneficial for the reader to have these clearly and consistently presented regarding what outcome you want to measure and what outcome measure you use.Furthermore, to use the same terms i.e. in some sections you use workers' sick leave, sick leave or work-related factors.Participants 3.

Table 1 -
provided are for grading systems and not for a QST protocol.Did the authors mean to reference the Rolke et al, 2006 (40)DFNS protocol here?If DFNS Rolke 2006 protocol is what the researchers are using, are the researchers following a modified version of DFNS Rolke 2006 protocol or reproducing exactly.For example, this protocol uses testing of 1 pinprick versus 10 repeated pinpricks and repeats this procedure 5 times, this (55 pinpricks !) may be poorly tolerated by patients, esp hyperalgesic.Have the researcher's modified this?How?Perhaps Table 1 could include more detail on protocol or how differs to referenced protocol or include as supplementary info different QST info compared to text-pinprick omitted, heat pain threshold here instead of cold pain thresholds Clinical Neurological Exam-indicate in Table1what will this entail?Is the main purpose to help with pain classification?Patients attending Rheumatology MSK triage clinics are typically multi-site MSK pain, eg.Small joints of the hands plus other several painful area.It is not clear how it's decided which is the index site for the primary outcome pain NRS.Similarly for QST indicate how will select test site?

have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
This is an open access peer review report distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.