CCOMPLCS

Background: The case complexity of patients with chronic musculoskeletal pain (CMP) is often explained with the bio-psycho-social model. Comorbidities like anxiety, anger and depression, maladaptive cognitions such as catastrophizing and poor coping skills, and functional deficits due to poor physical activity with fear of movement play an important role and play an interdependent role in this model. Ten highly important factors were used to develop the Case Complexity Index (CCI) with predominantly psychosocial factors as the most important factors in the assessment of the CCI and represented for patients in rehabilitation 59.1% of the case complexity.
Objective: The aim of this study is to evaluate the construct validity of the CCI with patient-reported outcomes in patients with CMP admitted to a daycare rehabilitation setting.
Study design: Prospective cohort study with care as usual
Methods: 50 patients with CMP referred to daycare rehabilitation will be admitted. At the primary stage unidemsionality will be assessed with factor analysis. For this purpose principal component analysis will be used with varimax rotations. For each factor of the CCI a Cronbach’s alpha, with 95% confidence intervals, will be calculated. The hypothesis testing of the construct validity will be performed with the Acceptance and Action Questionnaire (AAQ-II), Outcome Questionnaire (OQ-45), Pain intensity measured with a VAS scale, Personal Inventory DSM-5 (PID-5), Zelfinventarisatielijst (ZIL), Central Sensitization Inventory (CSI), Pain Catastrophizing Scale (PCS), Pain Disability Index (PDI), scale of kinesiofhobia (TAMPA), care uptake, widespread pain, sick leave and intensity of the advised therapy. Secondly diagnostic odds ratio’s will be provided with the CCI and a classification system with a regression tree will be provided to determine the contribution of each individual factor of the CCI to the level of care needed.

The main objective will be to assess the construct validity of the individual domains of the CCI with validated questionnaires as well as the total score of the CCI. Secondary objectives will be to assess the diagnostic odds ratio of the CCI in relation to the outcome therapy determined by the interdisciplinary team and a classification system with a regression tree to determine the contribution of each individual factor of the CCI to the outcome.

2.6 Hypothesis testing and sample size
In line with the recommendations for qualitative research by Terwee et al. (2007) hypotheses will be provided beforehand and proof of construct validity will be considered if more than 75% of the hypothesis are met for the total CCI. Fifty patients will be included in (sub)groups. A fair relationship (.3 < r < .6) is expected with the OQ-45 social role, VAS scale, PID-5, OQ-45, PDI, CSI, sick leave, care uptake, widespread pain, TAMPA scale, Personal injury proceedings and a negative moderate correlation with the AAQ-II. Moderate to good correlations (.6 < r < .8) are expected with the PCS, SQ-48, ZIL and the intensity of the advised therapy. Good to excellent correlations (r > .8) are not expected.

Statistical analysis

2.7.1 Flow diagram
A flow diagram of enrolment will provide an overview of the exact numbers in every stage of the study, namely: enrolment, completion of the questionnaires, interdisciplinary consult without the patient. Participants lost to follow-up and who discontinued the diagnostic phase will be described with reasons.

2.7.2 Baseline descriptives
A table will be provided with the baseline descriptives. The number of participants mean or median with the percentage or SD will be provided of the cohort where appropriate. Baseline characteristics will include age, duration of pain, pain intensity, gender, marital status, smoking status, alcohol use, insurance company, occupational status, specialization of the referring physician.

2.7.3 Internal consistency
At the primary stage unidimensionality will be assessed with factor analysis. For this purpose principal component analysis will be used with varimax rotations. For each factor of the CCI a Cronbach’s alpha, with 95% confidence intervals, will be calculated. Eigenvalues of more than 1 will be selected and acceptable internal consistency will be assumed with Cronbach’s alpha between 0.7 - 0.95.61

2.7.4 Hypothesis testing
The hypothesis testing of the construct validity will be performed in Rstudio version 3.6.3 with ETA coefficient testing, library “lsr”, version 0.5. Multiple testing will be controlled with a Bonferroni correction. The handling of missing data will be with Multiple imputation if the data is missing at random (MAR) or missing completely at random (MCAR). A result will be considered statistically significant with a p-value of < 0.05 after the Bonferroni correction. Floor and ceiling effects are considered if more than 15% of the patients score the highest or lowest possible score on a questionnaire.

2.7.5 Discriminative properties with Diagnostic odds ratio and regression trees
Secondly diagnostic odds ratio’s will be provided with the CCI and a classification system with a regression tree will be provided to determine the contribution of each individual factor of the CCI to the level of care needed. For the regression tree a training set of 80% and a test set of 20% will be used. Diagnostic odds ratio’s for the diagnosis of a patient in secondary health care will be calculated for inclusion into secondary care with CCI > 30 compared to CCI < 30, CCI > 40 compared to CCI < 40 and CCI > 50 compared to CCI < 50. Differences in diagnostic odds ratio will be described.

A sensitivity analysis will be performed with the total score of the questionnaires reflecting the domains and the total score of the CCI on the outcome therapy of the interdisciplinary consult with a regression tree. A training set of 80% of the data and a test set of 20% will be used. The difference in the diagnostic odds will be described with Formula 1.

Interdisciplinary Rehabilitation assessment