# Multidomain evaluation and data-driven approaches to predict recurrent neck pain (END-RNP): a study protocol for a multicentre prospective longitudinal cohort study

## TL;DR

This study aims to identify factors that predict the frequency and severity of recurrent neck pain episodes using physical, psychological, and social data collected over 12 months.

## Contribution

The study introduces a novel data-driven approach to predict recurrent neck pain using multidomain factors during symptom remission.

## Key findings

- Baseline measurements of physical, psychological, and social factors will be used to predict future neck pain episodes.
- Two multivariable models will be developed to predict the number of days with pain and its severity.
- Internal validation will assess model performance and clinical utility for personalized prevention strategies.

## Abstract

Neck pain (NP) is a leading cause of disability worldwide and affects more than 200 million people. Incidence and associated economic burden are constantly increasing, and little is known about the factors that promote new and more severe episodes in those individuals with recurrent NP. Current evidence supports that changes in physical, psychological and social factors persist between NP episodes, and these changes might contribute to the development of new episodes. The End Recurrent Neck Pain (END-RNP) study aims to use physical, psychological and social factors tested while in symptom remission to predict, within a 12-month period, the frequency and severity of new NP episodes.

The END-RNP study is a multicentre, prospective cohort study conducted from March 2025 to February 2028 at the University of Birmingham and the University of Essex (UK). 300 adults reporting two or more NP episodes in the previous year will be recruited from September 2025 to form the recurrent NP cohort, and 48 adults without a history of NP will provide normative data. Laboratory testing will be conducted for all participants when pain-free by assessing cervical kinematics and proprioception, neck-muscle strength, endurance and activation, pain processing, psychological and social factors. All recurrent NP participants will complete online questionnaires every 2 weeks for 12 months, recording days with NP, pain intensity/interference, healthcare use and other behavioural and environmental factors. Participants in the recurrent NP cohort who experience an acute NP episode during the 12-month follow-up will repeat the laboratory assessment. To develop the prediction models, candidate predictors will be the baseline measurements of any feature that shows either cross-sectional differences between recurrent NP and control groups or within-subject changes between the pain-free baseline and a pain episode. From the identified candidate predictors, two multivariable models will be developed using penalised regression, with (i) number of days with NP (linear regression) and (ii) NP severity (ordinal regression) as their respective dependent variables. Internal validation will use bootstrap resampling to estimate optimism-adjusted performance (R2, C-statistic and calibration slope), prediction instability and uncertainty, and clinical utility. The models from the END-RNP study will provide clinical prediction tools to help identify those at high risk of frequent and severe NP episodes and to inform the personalised prevention of recurrent NP.

The END-RNP study was approved by the Ethics Committee at the University of Birmingham (ERN_4005-Aug2025) and by the University of Essex (ETH2526-0098) on 2 September 2025, prior to the recruitment of the first participant. The findings will be presented at national and international conferences and submitted for publication in peer-reviewed journals.