Home > Riviste > European Journal of Physical and Rehabilitation Medicine > Fascicoli precedenti > Articles online first > European Journal of Physical and Rehabilitation Medicine 2020 Mar 03

ULTIMO FASCICOLO
 

JOURNAL TOOLS

eTOC
Per abbonarsi PROMO
Sottometti un articolo
Segnala alla tua biblioteca
 

ARTICLE TOOLS

Publication history
Estratti
Permessi
Per citare questo articolo

 

 

European Journal of Physical and Rehabilitation Medicine 2020 Mar 03

DOI: 10.23736/S1973-9087.20.05465-9

Copyright © 2020 EDIZIONI MINERVA MEDICA

lingua: Inglese

Mapping patient reported outcome measures for low back pain to the International Classification of Functioning, Disability and Health using random forests

Kerstin TUECHLER 1 , Elisabeth FEHRMANN 1, 2, Thomas KIENBACHER 1, Patrick MAIR 3, Linda FISCHER-GROTE 1, Gerold EBENBICHLER 4

1 Karl Landsteiner Institute for Outpatient Rehabilitation Research, Vienna, Austria; 2 Department of Psychology, Karl Landsteiner University of Health Sciences, Krems, Austria; 3 Department of Psychology, Harvard University, Cambridge, MA, USA; 4 Department of Physical Medicine and Rehabilitation, Medical University Vienna, Vienna, Austria


PDF


BACKGROUND AND AIM: This proof-of-concept study demonstrates and validates a novel approach that automatically translates information derived from patient reported outcome measures (PROMs) into a standardized functioning report based on the International Classification of Functioning, Disability and Health (ICF).
METHODS: The translation algorithm was developed using data from a total of 244 patients with chronic low back pain (cLBP) who completed the Roland-Morris disability questionnaire as well as the Pain Disability Index, and who were interviewed using the activity and participation component of the ICF brief core set for low back pain. Random forest machine learning models that predict the presence or absence of an impairment at the specific ICF category based on the information extracted from the PROMs as well as the personal factors age and gender were developed and validated. An additional external data set from 19 patients with moderate to severe cLBP was used to avoid overfitting and to prove the generalizability of the fitted prediction models.
RESULTS: Accuracy of the prediction models was found to be acceptable for the most relevant ICF brief core set categories for low back pain if applied at the population level.
DISCUSSION: The findings from this study are of high relevance for the further development of automatic linking programs that would allow the ICF-based classification of functioning properties within the International Classification of Diseases (ICD-11) for any health condition. Future research is needed to identify PROMs that are best suited to the automatic ICF translation process using machine learning algorithms.


KEY WORDS: International Classification of Functioning, Disability and Health; Patient reported outcome measures; Low back pain; Machine learning; Random forests

inizio pagina