From acute to chronic: the PREVENT protocol for understanding pain progression after emergency department visits

Conceptual framework & analytic rationale

Building upon the framework described above, the Cognitive Risk Model for CMP conceptualizes pain beliefs and attitudes as the primary independent variable, with the presence of CMP as the primary outcome, and functional status and pain intensity as secondary outcomes. By applying this theoretical lens, our study aims to identify how pain beliefs and behavioural responses constrain meaningful action and contribute to pain persistence. This model informed the selection of variables, timing of follow-up assessments, and the structure of our analytic approach. Specifically, it supports the inclusion of psychosocial and behavioural mediators (e.g., catastrophizing, fear-avoidance, physical activity, pain self-efficacy, recovery expectations), and contextual modifiers (e.g., substance use, social determinants of health, mental health conditions), and pathway modelling to explore both direct and indirect effects.

The study design was further refined and strengthened through consultation with ED clinicians, pain researchers, and a patient partner with lived experience of chronic pain, ensuring alignment with clinical priorities and patient-centred outcomes.

Recruitment will occur through direct, in-person approaches in the ED based on screening via the electronic medical record (EMR) tracking system (Figure 2).

"V体育平台登录" Inclusion criteria

Adult ED patients ≥18 years of age will be eligible for participation if they meet specific criteria: (1) have acute neck and/or low back pain (onset ≤4 weeks); (2) have incurred a recent (≤4 weeks) trauma (e.g., motor vehicle crash, falls, bicycle accidents); (3) have no definitive cause of pain related to unstable fracture, malignancy, active infection, or inflammatory conditions (e.g., ankylosing spondylitis); (4) have no co-occurring progressive neurological symptoms (e.g., spinal cord injury or cauda equina syndrome); (5) are not currently pregnant or planning pregnancy within the next 6 months; (6) have no impairments precluding informed consent (e.g., cognitive dysfunction); (7) are English-speaking; and (8) have a willingness to participate in follow-up assessments.

The ≤4-week trauma window ensures that participants are enrolled during the acute phase of pain and that pain onset can be clearly linked to a specific inciting event. This design feature enhances the precision of prospective tracking of pain progression and aligns with established chronic pain definitions, which require symptoms to persist for ≥3 months on most days or every day (Dahlhamer et al., 2018; Nahin et al., 2023; Ray et al., 2024). This approach also mirrors methodologies used in post-surgical pain studies, where researchers follow patients from a well-defined index event to assess long-term pain outcomes (Katz & Seltzer, 2009). By anchoring enrolment to a recent traumatic event, we minimize misclassification of subacute or chronic pain and increase the validity of modelling. This approach is consistent with prior ED-based longitudinal research (Beaudoin et al., 2020) and supports reliable identification of acute-to-chronic pain transitions.

Recruitment

After EMR screening, study staff will coordinate with the patient’s primary ED provider to confirm eligibility and ensure that approaching the patient will not interfere with clinical care. To maintain confidentiality and minimize disruption, recruitment discussions will take place in private or semi-private areas (e.g., examination rooms). Study staff will introduce the study, explaining its purpose, procedures, and voluntary nature, while providing the opportunity for patients to ask questions. If the patient expresses interest, staff will proceed with the informed consent process. Consent will only be obtained once the patient is clinically stable, ensuring that they are in a condition where they can fully comprehend the study details. Participants will be given ample time to review the consent document and ask any questions before signing. Research staff will be available to provide explanations and answer any questions the participant or their family may have. Participants will also have the option to discuss the study with their healthcare provider or family members before deciding. To ensure participants fully understand the research, the research staff will ask comprehension questions after the participant has read the consent form. This step will help assess whether the participant has understood the study, its risk, and benefits. If a participant is unable to demonstrate an understanding of the study, they will not be enrolled. If the treating clinical team or research staff suspects that a patient may have diminished capacity to consent, the patient will not be approached for participation. This ensures that all participants can make an informed decision. Participants will receive a PDF copy of the signed consent form immediately after providing consent via the REDCap (Research Electronic Data Capture) e-consent system. This ensures they have a record of their participation and the study details. All signed consent documents will be securely stored in REDCap. REDCap provides secure electronic storage, ensuring the confidentiality and integrity of participant information in compliance with Health Insurance Portability and Accountability Act (HIPAA) regulations. Since this study is under HIPAA guidelines, all consent documents and related research records will be maintained for a minimum of 6 years following study closure. During this time, the documents will remain securely archived in REDCap, which offers audit trails and data encryption to safeguard participant privacy.

Since January 2025, screening efforts at our primary ED site have identified 217 patients with musculoskeletal pain, of whom 28 met all inclusion criteria, including the ≤4-week trauma requirement. Among those eligible, 25 participants consented and were enrolled, while 3 declined participations, yielding an acceptance rate of approximately 89% and an overall enrolment rate of 11.5% based on screened patients. These preliminary data support the feasibility of enrolling the target sample of 246 participants within the planned 24 – 36-month recruitment period.

Financial reimbursement

Patient compensation is allocated to incentivize participation, enhance retention, and reduce attrition during the study’s follow-up assessments. Given the longitudinal nature of the study, financial incentives will support continued engagement at key time points, ensuring high-quality data collection and completeness of outcome measures. The distribution of compensation includes:

• Baseline: completion of Neck-Pain & Attitude Questionnaire (Neck-PAQ) validation study: $10

• 1-month follow-up: $50 raffle (four winners; ∼1 in 62 chance)

• 3-month follow-up: $100 raffle (four winners; ∼1 in 62 chance)

• 6-month follow-up: $200 raffle (one winner; ∼1 in 246 chance)

Compensation will be provided through prepaid debit cards or electronic payment methods, ensuring timely and convenient distribution to participants. The incentive structure aligns with best practices for participant retention in longitudinal studies, while remaining compliant with NIH and institutional research (Teague et al., V体育安卓版 - 2018).

Quantitative outcomes & assessment tools

To ensure standardization and data harmonization with other studies, this study aligns with the HEAL Common Data Elements (CDE) repository (Wandner et al., 2021). Key measures, including pain intensity (VAS), pain frequency (NHIS Pain Chronicity), psychological assessments (PCS, TSK, PHQ-2, GAD-2), functional status (ODI, NDI), and substance use screening (ASSIST), were selected from HEAL’s CDE repository whenever possible (Table 1). Participant follow-up assessments will be conducted at 1-, 3-, and 6-months post-ED visit to track pain progression and related outcomes.

The primary outcome is the development of CMP in the neck and/or low back, defined as pain in the same anatomical region as the acute presenting complaint, present on most days or every day for at least three months. Secondary outcomes will include changes in pain beliefs (assessed through B-PAQ/N-PAQ scores), pain catastrophizing (PCS scores), and fear-avoidance behaviours (TSK scores), and healthcare utilization, including recurrent ED visits and outpatient care. Additional secondary measures will evaluate changes in physical activity levels (IPAQ-SF scores), pain self-efficacy (PSEQ-2 scores), and pain-related disability, measured through changes in ODI and NDI scores over time. While this study applies the Cognitive Risk Model for CMP specifically to post-traumatic neck and/or low back pain, the model is designed to be adaptable to other anatomical regions and chronic pain types – including both chronic primary and secondary musculoskeletal pain – in future research (Treede et al., 2015).

In addition to baseline enrolment forms, we will perform a chart audit post-ED discharge to assess ED management, including ED imaging (x-ray, CT, MRI), medication management (including opioids, NSAIDs, acetaminophen, and muscle relaxants), ED disposition (admission vs discharge), diagnostic codes (ICD-10 for musculoskeletal and neurological conditions), and referrals provided (e.g., primary care, pain specialist, physical therapy). After 1- and 6-month follow-ups, additional audits will be completed utilizing an automated process to query the EMR using the Financial Identification Number (FIN) and Medical Record Number (MRN) from the index ED visit. Using these patient-specific identifiers collected during enrolment, a report can be generated for pre-specified data fields in the EMR which exactly matches the data dictionary in the case report form (CRF) built in REDCap. Data can be imported into REDCap and will be checked for fidelity and missingness. An advantage of this process is reduction in bias inherent in manual chart review (Gilbert et al., 1996). As a further check, we will query the regional health information exchange for relevant healthcare utilization at regional outside hospitals.

Sample size & quantitative data analysis

Our primary outcome – pain frequency – is assessed as a categorical variable based on participants’ reported number of days per week experiencing neck or low back pain (0–1 days, 2–3 days, 4–5 days, or 6–7 days). Pain frequency will be assessed at 1, 3, and 6 months. For analysis, responses will be dichotomized to classify CMP status: participants reporting pain on 4–5 or 6–7 days per week will be categorized as having CMP (coded as 1), while those reporting 0–1 or 2–3 days per week will be categorized as not having CMP (coded as 0). Based on prior studies reporting a 12%–27% transition rate from acute to chronic pain (Beaudoin et al., 2020; Office of the Assistant Secretary for Planning and Evaluation, 2021; Bourassa et al., 2020; Barnett et al., 2017), we conducted a power analysis to estimate the incidence of CMP with acceptable precision. Using a small effect size (Cohen’s d 0.2), with α = 0.05 and power = 0.80, a minimum sample size of 197 participants is required. To account for an estimated 25% attrition rate, we will recruit 246 participants to ensure adequate study power. This sample size will also provide sufficient power to explore the association between baseline pain beliefs and subsequent development of CMP, as well as to conduct preliminary analyses for establishing potential clinical cut-offs on pain relief measures. Higher scores may indicate individuals at elevated risk for developing CMP and who may benefit from targeted educational interventions. However, these exploratory analyses will require validation in independent samples.

Quantitative statistical analysis will involve predictive model development using multivariable logistic regression to identify significant predictors of CMP as a binary outcome. Pain beliefs will be modelled as the primary independent variable, with adjustment for potential confounding variables identified in the conceptual framework (e.g., demographics, anxiety, depression, substance use, and social determinants of health). Mediation analyses will explore indirect pathways through cognitive-affective constructs (e.g., pain catastrophizing, fear-avoidance, recovery expectations, and pain self-efficacy) and behavioural responses (e.g., physical activity), using structural equation modelling or mediation analysis to estimate direct and indirect effects. Interaction terms between pain beliefs and demographic variables will be tested to evaluate potential effect modification. Model validation will incorporate cross-validation techniques, such as k-fold cross-validation, to assess robustness. Receiver operating characteristic (ROC) curve analysis will determine the model’s discriminative ability using area under the curve (AUC) estimates, and calibration analysis (e.g., Hosmer–Lemeshow test) will assess model fit. Preliminary clinical cut-offs for pain beliefs scores will also be explored by identifying thresholds that maximize sensitivity and specificity for predicting CMP developed based on ROC analysis. Sensitivity analyses will evaluate the stability of findings under alternative variable definitions and measurement criteria.

Validation of the neck-pain and attitudes questionnaire (Neck-PAQ)

During patient screening in the ED, individuals presenting with acute neck pain will be identified as potential participants for the Neck-PAQ validation study. Eligible and consenting participants will be invited to complete the Neck-PAQ and participate in a Think Aloud interview to evaluate the tool’s clarity, interpretability, and response process validity (Scott et al., 2021). The Think Aloud methodology involves asking participants to verbalize their thoughts in real-time as they complete the questionnaire (Padilla & Leighton, 2017). This process enables the research team to assess how respondents interpret and process each item, and whether the intended constructs are being accessed during response formulation. Participants will be reminded that the goal is to capture their natural reasoning and will complete a short practice session to increase comfort with the process (Padilla & Leighton, 2017). All Think Aloud interviews will be audio-recorded, securely stored in REDCap and assigned an identification number. Audio recordings will be transcribed using Whisper (OpenAI, 2023), an automated speech recognition system that operates locally without requiring internet access, thereby minimizing risks to protected health information (PHI). Manual verification will be conducted to ensure transcription accuracy. Anonymized transcripts will then be uploaded to NVivo (QSR International Pty Ltd, 2020) for qualitative analysis. The original audio files will be securely stored on password-protected drives accessible only to study personnel. Upon study completion, all audio recordings will be permanently deleted. Audio will be used solely for data analysis and will not be included in any presentation, publication, or non-research setting. Only transcribed, de-identified excerpts will be used for dissemination purposes.

Qualitative sample size & data analysis

The target sample size is guided by the concept of information power, which considers the study aim, sample specificity, theoretical grounding, quality of dialogue, and chosen analytical strategy (Malterud et al., 2016). As the Neck-PAQ is a newly developed instrument and the aim is narrow and well-defined, a sample of approximately 20 participants is expected to yield sufficient information power to evaluate the appropriateness of the tool for capturing maladaptive beliefs about the neck, pain, management, and recovery. Verbal data from the Think Aloud interviews will be analysed using deductive content analysis guided by a predefined cognitive model of response processes (Figure 3). This model includes five domains: Comprehension, Knowledge Recall, Evaluation of Truthfulness, Decision-Making and Response Selection, and Reflection and Adjustment (see appendix Neck-PAQ Coding Manual). Each distinct thought unit will be assigned to one or more of these categories. Coding will be conducted independently by two trained raters. Inter-rater reliability will be calculated using Cohen’s Kappa, with a threshold of ≥0.60 required to indicate acceptable agreement (McHugh, 2012).

Discrepancies will be discussed and resolved collaboratively, with updates to rater training or the coding manual as necessary. Although this is a deductive process, qualitative insights may emerge during analysis. These will be descriptively reported, but the core analytic strategy focuses on the structured categorization of verbalized response processes to validate whether items in the Neck-PAQ are functioning as intended.