Individuals (n=180) with subacute (<1 month, >3 months) WAD grade I and II with medium to high-risk symptoms of working age will be randomized into three groups with block randomization in a prospective, assessor-blinded randomized controlled trial. Two primary intervention groups (A and B) will receive manual therapy (MT) in the same out-patient clinic. In addition, group A will receive a remote, novel, computer-based cervical kinaesthetic exercise program starting at visit two, whereby quality and quantity of exercise performance, as well as compliance (frequency and duration), will be registered into the physical therapy clinic's system for evaluation. Group B will receive neck exercises (not computer-based) provided by the corresponding physical therapist. Group A will continue remote exercise therapy until 6 months post baseline measurements regardless of whether they are still being treated in-clinic or discharged.

Hypothesis 1.a.: Internet-based neck-specific CKE combined with in-clinic manual therapy and exercises will be superior to manual therapy and exercises alone at a physiotherapy clinic (i.e., treatment A is superior to treatment B) for self-reported and measured (movement performance) outcome measures.

Hypothesis 1.b.: Improvements in self-reported outcomes will positively correlate with outcomes of movement performance testing, as will the pre- to post-intervention changes from baseline to follow-up assessments.

The groups will be compared to a "treatment as usual" group (C). Objective measures include measurements for motor control, proprioception, and cervical range of motion. Neck disability and pain intensity, general health, self-perceived handicap, and physical, emotional and functional difficulties due to dizziness will be measured using questionnaires. Short-term effects will be measured at 10-12 weeks and long-term effects at 6- and 12-months post baseline measurements.

Hypothesis 2.a.: Participants of groups A and B will improve significantly more than those in group C for subjective and objective outcome measures.

QUALITY ASSURANCE PLAN:

PhD student will monitor the data and guarantee high data quality. Data will be protected according to the Act on Data Protection and the Processing of Personal data (Icelandic law) and stored at the University Hospital or at a secure NeckCare database. Data will be registered into Microsoft Excel and shared through a secure network with other researchers to promote transparency.

DATA CHECKS:

Data will be checked for outliers.

SOURCE DATA VERIFICATION:

N/A

DATA DICTIONARY:

WAD: whiplash-associated disorders; MVC: motor vehicle collision; VR: virtual reality; NDI: neck disability index; ROM: range of motion; PTs: physiotherapists; MT: manual therapy; CKE: cervical kinaesthetic exercise; IT: information technology; IoT: internet of things; IMU: inertial monitoring unit; VAS: visual analog scale; SF-36: short form 36 item health survey; DHI: dizziness handicap inventory; CSI: central sensitisation inventory; CS: central sensitization; ToT: time on target; CI: confidence interval; MCIC: minimal clinically important changes.

STANDARD OPERATION PROCEDURES:

A total of 180 subjects (60 from each of the three groups) will be recruited through the databases of The Emergency Department of Landspitali University Hospital, Reykjavik, Iceland via phone call.

Subjective and objective outcome measures will be collected by a PhD student. In addition to questionnaire and clinical outcome measures, the pharmaceutical database of the Directorate of Health in Iceland will be accessed for data on subjects' drug use.

Data will be analysed by the PhD student using Microsoft Excel and R, a free software environment for statistical computing and graphics.

The PhD student will be responsible for reporting adverse events and change management.

SAMPLE SIZE ASSESSMENT:

Sample size and power regarding group differences will be calculated based on the primary outcome NDI. To detect a clinically relevant improvement of 8% in the NDI, 40 participants are needed in each group for 80% power. For non-inferiority tests, the significance level will be set to 5% (p > 0.05) which corresponds to the one-sided confidence interval (95% CI). To ensure that enough people are in each group after dropouts, for prediction analyses and the opportunity for subgroup analyses, 60 participants will be included in each group (n=180).

PLAN FOR MISSING DATA:

If data will be found to be missing by any reason, each incident will be assessed by the PhD student and her advisors. Imputation methods may be used, the subject involved may be asked to come back in for a measurement or answer questionnaires again (depending on the time that has passed from the measurement/questionnaires until data is discovered missing) or the subject involved will be removed from the study.

STATISTICAL ANALYSIS PLAN:

Data analyses and statistics Database monitoring will be performed by the primary researcher and statisticians involved, independent of sponsors and competing interests. Background data will be assessed after all subjects have completed their baseline measurements.

Background data and association of self-reported and clinical measures Background data will be evaluated with descriptive statistics and differences in baseline data determined using t-tests and (M)ANOVA (mean and standard deviation) or non-parametric tests where appropriate. Association of self-reported measurements and clinical measures will be assessed for baseline, 10-12-week, 6-month and 12-month measurements using Pearson or Spearman correlation depending on the distribution of the data. Subgroup analyses of CPS may possibly be performed.

Effectiveness of therapy Analyses will be performed primarily on an intention-to-treat basis (as individuals being randomized into the groups) and secondarily on a per protocol basis (individuals who fulfilled the programme for at least 50%). Imputation methods may be used when deemed to have additional value. Linear mixed models with time as a repeated factor will be used to assess the effect of treatment for each outcome. Correlation over time, within the PT and prognostic factors will be accounted for. Estimates of the effect of the intervention will be obtained by constructing linear contrasts to compare the mean change in outcome for each measurement (subjective and objective) to each time point between treatment groups and between each treatment group compared to control group. The main dependent variable will be score on NDI with independent fixed factors time (baseline, 10-12-week, 6-month, and 12-month) and group (A, B and C). If the non-inferiority of C to A or of C to B is concluded based on the 95% CI, a test of the superiority of C to A or of C to B will be performed as suggested by Lesaffre (see reference list). The variation in response to intervention (heterogeneity of treatment effect) will be evaluated using regression analysis.

Inclusion Criteria:

• subacute (>1 month, <3 months) neck problems corresponding to WAD grades I-II verified by WhipPredict
• medium to high risk symptoms (NDI score >32% OR subject is >36 years old) OR VAS score >4/10
• within daily reach of a computer/tablet/smartphone and Internet
• showed neck symptoms within the first week following the car collision (i.e. neck pain, neck stiffness)

Exclusion Criteria:

• WAD grades III-IV
• considerable degree of known or suspected physical pathology (Myelopathy, Spinal tumours, Spinal infection, Ongoing malignancy, Cervical spine surgery, Severe neck problems within their medical history which resulted in sick leave for more than a month in the year before the current whiplash injury, Other illness/injury that may prevent full participation from being feasible, Lack of ability to either understand or write Icelandic, Severe obesity (body mass index; BMI > 35), Pregnancy)
• unable to complete the assessment