Relationship Between Motor Level and Wheelchair Transfer Ability in Spina Bifida: A Study From the National Spina Bifida Patient Registry

Abstract

Objective: To identify the specific features that contribute to the variability in baseline wheelchair transfer and the changes in transfer ability (gain or loss) over time for a large cohort of patients with spina bifida (SB) in the National Spina Bifida Patient Registry.

Design: Longitudinal cohort study.

Setting: A total of 35 United States outpatient SB clinic sites.

Participants: Individuals (N=1687) with SB ages 5-73 (median, 13.33) years who were therapeutic ambulators or nonambulators.

Intervention: Not applicable.

Main outcome measure: Ability to transfer from a wheelchair to another level surface.

Results: Bayesian Network Analysis was used to reduce the initial variable set to the following predictors: SB subphenotype, motor level, age, insurance, sex, race, ethnicity, surgical procedures, and number of visits. We used a multinomial logistic model with Wald Chi-square analysis of effects to examine the relationships between transfer ability and predictors. A total of 295 of 1687 eligible patients (17.56%) with myelomeningocele (MMC) and 6 of 58 eligible patients (10.32%) with non-MMC experienced changes in transfer ability during the period of the study. For those with MMC and non-MMC, the highest number of individuals exhibiting changes in motor level had changes from thoracic to high-lumbar, high-lumbar to thoracic, high-lumbar to midlumbar, and midlumbar to high-lumbar lesion levels. Results of the Bayesian Network Analysis revealed that motor level was the predominant factor associated with baseline transfer ability followed by age. The combination of SB sub phenotype, motor level, age, insurance status, number and type of surgical procedures, and time point accurately classified the loss, gain, or no change in transfer ability 82.7% of the time.

Conclusions: Motor level was the predominant factor associated with baseline transfer ability, and the change in transfer ability was directly related to a corresponding change in motor level that might be explained by changes in muscle strength of the iliopsoas and quadriceps.

Keywords: Machine learning; Myelomeningocele; Registries; Rehabilitation; Spinal dysraphism; Walking; Wheelchairs.

Fig 1

Participant selection CONSORT diagram. Abbreviation: CONSORT, Consolidated Standards of Reporting Trials.

Fig 2

Baseline motor level.

Fig 3

Change in motor level (increase in function).

Fig 4

Change in motor level (decrease in function).

Fig 5

Prediction of transfer ability by subphenotype, baseline motor level, age, surgical procedures, and number of visits.

Fig 6

Relationship between age and probability of transfer.