Objectives: To assess the effectiveness and cost-effectiveness of a single session compared to multiple sessions of education and exercise for older adults with spinal pain treated conservatively in an advanced practice physiotherapy (APP) model of care.

Methods and Analysis: In this pragmatic randomized controlled trial, 152 adults (≥ 18 years old) with neck or back pain initially referred for a consultation in neurosurgery, but treated conservatively, will be recruited through the APP neurosurgery CareAxis program in the Montreal region (Quebec, Canada). In the CareAxis program, older patients with spinal pain are triaged by an advance practice physiotherapist and are offered conservative care and only potential surgical candidates are referred to a neurosurgeon. Participants will be randomized into one of two arms: 1- a single session or 2- multiple sessions (6 sessions over 12 weeks) of education and exercise with the advance practice physiotherapist. The primary outcome measure will be the Brief Pain Inventory (pain severity and interference subscales). Secondary measures will include self-reported disability (the Neck Disability Index or Oswestry Disability Index), The Pain Catastrophizing Scale, satisfaction with care (VSQ-9 and MedRisk questionnaires), and health-related quality of life (EQ-5D-5L). Participants healthcare resources use, and related costs will be measured. Outcomes will be collected at baseline and at 6, 12 and 26 weeks after enrollment. Intention-to-treat analyses will be performed, and repeated mixed-model ANOVA will assess differences between treatment arms. Cost-utility analyses will be conducted from the perspective of the health care system.

Ethics and dissemination: Ethics approval has been obtained from the Comité d'éthique de la recherche du CIUSS de l'Est-de-l'Île-de-Montréal (FWA00001935 and IRB00002087). Results of this study will be presented to different stakeholders, published in peer-reviewed journals and presented at international conferences.

OBJECTIVES

The aim of this pragmatic randomized controlled trial is to evaluate at 6, 12 and 24 weeks, the clinical effectiveness in terms of patient-centered outcomes and the cost-effectiveness of a single session compared to multiple sessions of education and exercise for older adults with spinal pain treated conservatively in an APP neurosurgery model of care.

METHODS

Study Design This study is a pragmatic randomized controlled trial with cost-effectiveness analyses. This trial will use a pragmatic approach within an existing APP model of care in neurosurgery. Therefore, the rehabilitation intervention in both treatment arms will be offered as they are provided in their current clinical reality (see Figure 1 for flowchart).

Setting and Participants This pragmatic randomized controlled trial will be performed within the CareAxis neurosurgery APP model of care. In 2019, CareAxis, a non-profit organization, developed and implemented an APP model of care for older patients with spinal pain referred to neurosurgery in the province of Quebec (Canada). The model was developed in partnership with the Montreal Neurological Institute and Hospital, Montreal, Quebec, Canada. The aim of this model is to reduce wait time to consult a neurosurgeon, more efficiently triage surgical candidates, and improve conservative and rehabilitation care for patients referred to the program. The current model of care involves a group of trained APPTs, who assess and triage potential surgical candidates, recommend medical care (medication or injection), provide education, and prescribe a self-management exercise program. This is systematically offered to all patients in a one-hour session. Only patients deemed surgical candidates are then referred to see the neurosurgeon. The program is funded through health charities and is offered at no costs to older adults with a referral from a family physician.

This study will take place in physiotherapy clinics (n=9) associated with the CareAxis group in Quebec, Canada. Eligible participants will be recruited when referred to the CareAxis model of care, either directly through family physicians or from the Centre de répartition des demandes de service de Montréal - the centralized intake system for specialist referrals (including neurosurgery) from primary care in Montreal, Quebec. The eligibility criteria in the current trial are based on the intended clientele to be cared by the APPT within this model of care, representing our pragmatic approach to the definition of our sample. Eligibility criteria: 1) adults consulting for a neck or back condition; 2) aged 18 years old or older; 3) referred for a consultation in neurosurgery to the CareAxis group, either directly by family physicians or from the Centre de répartition des demandes de service de Montréal and 4) not considered as a potential surgical candidate based on the initial APPT assessment.

For patients who decline participation in the project, demographic data, such as age, gender and duration of symptoms will be collected to calculate participation proportions and establish comparison between participants and non-participants.

Randomization All eligible and consenting patients will be randomly assigned to either the single session APP arm or the multiple sessions APP arm. A randomization list will be generated prior to the initiation of the study using an online random number generator. A blocked randomization will be performed with block sizes of 4, 6 or 8. Randomization will be stratified for the affected body region (neck or back) and with respect to gender. Allocation will be concealed in opaque and sealed envelopes that will be sequentially numbered. The randomization procedure will be performed by an independent research assistant not involved in other aspects of the current trial.

Interventions Participants will initially consult an APPT of the CareAxis group. In a one-hour consultation, the APPT will assess and triage surgical candidates, recommend medical care (medication or injection), as well as provide education and prescribe a self-management exercise program aimed at impairments and functional limitations identified during the initial assessment. Participants randomized to the single session arm will not receive additional care by the APPT while those randomized to the multiple sessions arm will receive five additional consultations with the APPT within a 12-week period. During the follow-up visits, the APPT will reassess the participants, provide further education and recommendations, review and modify the home exercise program and may provide other rehabilitation interventions. Since this is a pragmatic trial, the interventions in both treatment arms will be offered as they are provided in their current clinical reality and will not be standardized. Information regarding care will be systematically recorded. Patients will be allowed to take their usual medication and this data will be systematically collected through a diary. Information on other co-interventions during follow-up will also be collected.

Data Collection Data collection will take place at baseline and at 6, 12 and 26 weeks after the initial evaluation. Prior to being randomized and seen by the APP, eligible participants will answer a questionnaire regarding socio-demographic characteristics, such as age, sex, gender, education level, household income, and living status. Information on clinical variables such as anthropometric data, affected body area (neck or low back), reason for consultation, duration of symptoms and presence of any comorbidities such as arthritis, high blood pressure and diabetes will also be documented. The STart Back Screening Tool and a modified version for neck pain will be used to evaluate prognosis. All data collection will be done through the CareAxis web data collection portal already in use for clinical follow-up of patients and with the Centre de Recherche de l'Hôpital Maisonneuve-Rosemont (CRHMR) REDCap data collection portal. Both web platforms use Transport Layer Security (SLL/TLS) encryption to secure data.

Outcome Measures The primary outcome measure will be the Brief Pain Inventory (BPI), short form, pain severity scale and pain interference scale. The BPI is a self-reported questionnaire. The pain severity scale includes four numerical pain rating questions (0-10) related to worst, least, average and current pain intensity, while the pain interference scale includes seven numerical rating questions (0-10) related to the impact of pain on various functional activities. The BPI is valid, reliable and responsive to change in MSKDs populations, including spinal pain patients.

The secondary outcome measures will include other validated, reliable and responsive to change self-reported questionnaires. Depending on the affected body region, participants will complete the Neck Disability Index (NDI) for neck-related disorder or the Oswestry Disability Index (ODI), for back-related disorder. Pain catastrophizing will be assessed through the Pain Catastrophizing Scale. To assess satisfaction with received care, patients will be asked to complete a modified validated version of the 9-item visit-specific satisfaction questionnaire (VSQ-9) after the initial assessment. To assess satisfaction following the interventions and follow-up, participants will be asked to complete the MedRisk which was developed to measure patient satisfaction with rehabilitation care. EQ-5D-5L will be used to measured health-related quality of life. The investigators will also collect health care resource use, including medical consultations, diagnostic tests, medication, injections, rehabilitation and surgery using a self-reported questionnaire. Occurrence of any adverse events will be questioned at each time point.

Based on their initial assessment, the APPTs will complete a standardized form, indicating their diagnosis, suggestion for additional medical imaging or laboratory tests (if relevant), treatment plan (e.g., conservative treatment options, medication, injections, or physiotherapy care) and referral to other specialists, if relevant. For the multiple session arm, APPTs will document patients' progress, change in the exercise plan and all treatments provided.

Statistical Analyses Descriptive statistics will be used to present the participants' characteristics. Participants withdrawing from the study and reasons for withdrawal will be analyzed. Characteristics of participants and non-participants will be compared. Baseline demographic data will be compared across groups to establish the comparability across intervention arms (unpaired t-tests and Chi-square tests). If differences are observed, statistical models will be adjusted. Descriptive statistics will be computed for all outcome measures at the different measurement times. Intention-to-treat analyses will be performed. As secondary analyses, per-protocol analyses will also be performed. Missing data will be handled with multiple data imputation. A repeated mixed-model ANOVA will be used to determine whether outcomes differ between treatment arms across time points. Separate analyses will be conducted on each of the primary and secondary outcomes. In particular, if a difference (interaction) between groups is detected (p< 0.05), individual effects will be examined. Sphericity will be tested with Mauchly's test. If sphericity is rejected (variances of the difference are not equal), correction will be used to determine if repeated measures ANOVA test is statistically significant. If Epsilon > 0.75, the Huynh-Feldt correction will be used and if Epsilon ≤ 0.75, the Greenhouse-Geisser correction will be used. Differences in proportions of adverse events, co-interventions (medication, injections or any other treatment), satisfaction scores (VSQ-9 at initial assessment and MedRisk at 12 weeks) between treatment arms will be calculated using Chi-Square tests or Student's t-tests. Alpha level will be set at 0.05. All analyses will be carried out using the SPSS software (IBM Corp. Released 2019. IBM SPSS Statistics for Windows, Version 26.0) and the Excel software (Microsoft Corporation. 2018. Microsoft Excel for Windows).

Economic Evaluation A cost-utility analysis comparing both arms will be conducted. The analysis will be from the perspective of the publicly funded health care system. Secondary analyses including publicly and privately health care system costs will also be performed. Costs and outcomes will be assessed within the follow up period of the trial. Data on services used and the efficacy of each approach will be obtained from the concurrent trial. Health utility value will be derived from the EQ-5D-5L using the Canadian validated algorithm. Quality-adjusted life years (QALYs) will be estimated using the total area under the curve method. Unit costs for each health care resource will be obtained from Canadian sources. Costs will be expressed in 2021 Canadian Dollars.

Results of the cost-utility analysis will be presented as incremental cost per QALY gained. The statistical analysis will be conducted in accordance with current guidelines for a cost-effectiveness analysis alongside randomized control trials. The incremental cost and incremental outcome will be estimated using generalized estimating equations (GEEs) with appropriate links and distributions to account for repeated nature of the trial data. The incremental cost-effectiveness ratio will be obtained through the difference in the mean costs of the two approaches divided by the difference in the mean value of QALYs for the two arms as denoted by the coefficient of the intervention indicator variables. Uncertainty in the analysis will be addressed by estimating 95% confidence intervals using a non-parametric bootstrapping method. For this study, 10,000 estimates of costs and outcomes will be obtained for both arms. Results from the bootstrapping exercise will also be used to show cost-effectiveness acceptability curves, which represent the probability of each APP approach being cost-effective over a range of willingness to pay values that the health system may be willing to pay for an additional unit of QALY. A series of sensitivity analyses will be undertaken to examine the robustness of the trial findings.

Sample Size Calculation The sample size required is based on the primary outcome measure, the BPI short form. The BPI interference scale and the severity scale have estimated minimum clinically important difference of 1 point.The estimated standard deviation is 2.03 for spinal pain. The considered parameters are 0.05 for a type I error (α) with a power of 0.80 (1-β). For an analysis of variance (ANOVA), the sample size required is 76 participants per group. This sample also accounts for a potential 15% loss to follow up at 6 months. The total sample will be 152 participants and will provide adequate power.

Inclusion Criteria:

1. adults consulting for a neck or back condition;
2. aged 65 years old or older;
3. referred for a consultation in neurosurgery to the CareAxis group, either directly by family physicians or from the Centre de répartition des demandes de service de Montréal and
4. not considered as a potential surgical candidate based on the initial APPT assessment.

Exclusion Criteria:

1. Considered as a potential surgical candidate based on the initial advanced practice physiotherapist assessment.

Individual participant data that underlie the results reported in the primary study manuscript after deidentification.