Abstract
Accurate predictions of motor performance after stroke are of cardinal importance for the patient, clinician, and health care system. More than ten years ago, the proportional recovery rule was introduced by promising just that: high-fidelity predictions of recovery following stroke based only on the initially lost motor performance, at least for a specific fraction of patients. However, emerging evidence suggests that this recovery rule is subject to various confounds and may apply less universally than assumed by many.
We systematically revisited stroke outcome predictions by casting the data in a less confounded form and employing more integrative and flexible hierarchical Bayesian models. We jointly analyzed n=385 post-stroke trajectories from six separate studies – the currently largest overall dataset of upper limb motor recovery. We addressed confounding ceiling effects by introducing a subset approach and ensured correct model estimation through synthetic data simulations. Finally, we used model comparisons to assess the underlying nature of recovery within our empirical recovery data.
The first model comparison, relying on the conventional fraction of patients called fitters, pointed to a combination of constant and proportional to lost function recovery. Proportional to lost here describes the original notion of proportionality, indicating greater recovery in case of a more pronounced initial deficit. This combination explained only 32% of the variance in recovery, which is in stark contrast to previous reports of >80%. When instead analyzing the complete spectrum of subjects, model comparison selected a composite of constant and proportional to spared function recovery, implying a more significant improvement in case of more preserved function. Explained variance was at 53%.
Therefore, our data suggest that motor recovery post-stroke may exhibit some characteristics of proportionality. However, the levels of explanatory value were substantially reduced compared to what has previously been reported. This finding motivates future research moving beyond solely behavior scores to explain stroke recovery and establish robust single-subject predictions.
Competing Interest Statement
The authors have declared no competing interest.
Funding Statement
AKB’s clinician scientist position is supported by the dean’s office, Faculty of Medicine, University of Cologne. GRF gratefully acknowledges support by the Marga and Walter Boll foundation.
Author Declarations
All relevant ethical guidelines have been followed and any necessary IRB and/or ethics committee approvals have been obtained.
Yes
All necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived.
Yes
Any clinical trials involved have been registered with an ICMJE-approved registry such as ClinicalTrials.gov and the trial ID is included in the manuscript.
Not Applicable
I have followed all appropriate research reporting guidelines and uploaded the relevant Equator, ICMJE or other checklist(s) as supplementary files, if applicable.
Yes
Data Availability
The recovery data as well as jupyter notebooks (python 3.7, primarily software package pymc3) employed in this study are available from the authors on reasonable request.
Abbreviations
- CeC
- Compression enhanced coupling
- FM
- Fugl Meyer
- PPC
- Posterior predictive check
