BACKGROUND: Functional limitations are associated with poor clinical outcomes, higher mortality, and disability rates, especially in older adults. Continuous assessment of patients' functionality is important for clinical practice; however, traditional questionnaire-based assessment methods are very time-consuming and infrequently used. Mobile sensing offers a great range of sources that can assess function and disability daily. OBJECTIVE: This work aims to prove the feasibility of an interpretable machine learning pipeline for predicting function and disability based on the World Health Organization Disability Assessment Schedule (WHODAS) 2.0 outcomes of clinical outpatients, using passively collected digital biomarkers. METHODS: One-month-long behavioral time-series data consisting of physical and digital activity descriptor variables were summarized using statistical measures (minimum, maximum, mean, median, SD, and IQR), creating 64 features that were used for prediction. We then applied a sequential feature selection to each WHODAS 2.0 domain (cognition, mobility, self-care, getting along, life activities, and participation) in order to find the most descriptive features for each domain. Finally, we predicted the WHODAS 2.0 functional domain scores using linear regression using the best feature subsets. We reported the mean absolute errors and the mean absolute percentage errors over 4 folds as goodness-of-fit statistics to evaluate the model and allow for between-domain performance comparison. RESULTS: Our machine learning-based models for predicting patients' WHODAS functionality scores per domain achieved an average (across the 6 domains) mean absolute percentage error of 19.5%, varying between 14.86% (self-care domain) and 27.21% (life activities domain). We found that 5-19 features were sufficient for each domain, and the most relevant being the distance traveled, time spent at home, time spent walking, exercise time, and vehicle time. CONCLUSIONS: Our findings show the feasibility of using machine learning-based methods to assess functional health solely from passively sensed mobile data. The feature selection step provides a set of interpretable features for each domain, ensuring better explainability to the models' decisions-an important aspect in clinical practice.

Centro de Investigacion en Salud Mental Carlos 3 Institute of Health Madrid Spain

Department of Psychiatry Centre Hospitalier Universitaire Nîmes France

Department of Psychiatry General Hospital of Villalba Madrid Spain

Department of Psychiatry Madrid Autonomous University Madrid Spain

Department of Psychiatry Universidad Catolica del Maule Madrid Spain

Department of Psychiatry University Hospital Infanta Elena Madrid Spain

Department of Psychiatry University Hospital Jimenez Diaz Foundation Madrid Spain

Department of Psychiatry University Hospital Rey Juan Carlos Móstoles Spain

Department of Signal Theory and Communications Universidad Carlos 3 de Madrid Leganés Spain

Evidence Based Behavior S L Leganés Spain

Faculty of Information Technology Brno University of Technology Brno Czech Republic

Grupo de Tratamiento de Señal Gregorio Marañón Health Research Institute Madrid Spain

Kempelen Institute of Intelligent Technologies Bratislava Slovakia

See more in PubMed

Chamberlain AM, Rutten LJF, Jacobson DJ, Fan C, Wilson PM, Rocca WA, Roger VL, St Sauver JL. Multimorbidity, functional limitations, and outcomes: interactions in a population-based cohort of older adults. J Comorb. 2019;9:2235042X19873486. doi: 10.1177/2235042X19873486. 10.1177_2235042X19873486 PubMed DOI PMC

Miller ME, Rejeski WJ, Reboussin BA, Ten Have TR, Ettinger WH. Physical activity, functional limitations, and disability in older adults. J Am Geriatr Soc. 2000;48(10):1264–1272. doi: 10.1111/j.1532-5415.2000.tb02600.x. PubMed DOI

Torres-Castro R, Solis-Navarro L, Sitjà-Rabert M, Vilaró J. Functional limitations post-COVID-19: a comprehensive assessment strategy. Arch Bronconeumol. 2021;57:7–8. doi: 10.1016/j.arbres.2020.07.025. S0300-2896(20)30260-X PubMed DOI PMC

Roberts P, Wertheimer J, Park E, Nuño M, Riggs R. Identification of functional limitations and discharge destination in patients with COVID-19. Arch Phys Med Rehabil. 2021;102(3):351–358. doi: 10.1016/j.apmr.2020.11.005. S0003-9993(20)31265-X PubMed DOI PMC

Fernández-de-Las-Peñas C, Martín-Guerrero JD, Navarro-Pardo E, Rodríguez-Jiménez J, Pellicer-Valero OJ. Post-COVID functional limitations on daily living activities are associated with symptoms experienced at the acute phase of SARS-CoV-2 infection and internal care unit admission: a multicenter study. J Infect. 2022;84(2):248–288. doi: 10.1016/j.jinf.2021.08.009. S0163-4453(21)00391-1 PubMed DOI PMC

Goates S, Du K, Arensberg MB, Gaillard T, Guralnik J, Pereira SL. Economic impact of hospitalizations in US adults with sarcopenia. J Frailty Aging. 2019;8(2):93–99. doi: 10.14283/jfa.2019.10. PubMed DOI

Frontera WR, DeLisa JA, Gans BM, Robinson LR. DeLisa's Physical Medicine and Rehabilitation: Principles and Practice. Philadelphia: Lippincott Williams & Wilkins Health; 2019.

Patterson TL, Goldman S, McKibbin CL, Hughs T, Jeste DV. UCSD performance-based skills assessment: development of a new measure of everyday functioning for severely mentally ill adults. Schizophr Bull. 2001;27(2):235–245. doi: 10.1093/oxfordjournals.schbul.a006870. PubMed DOI

McKibbin C, Patterson TL, Jeste DV. Assessing disability in older patients with schizophrenia: results from the WHODAS-II. J Nerv Ment Dis. 2004;192(6):405–413. doi: 10.1097/01.nmd.0000130133.32276.83.00005053-200406000-00002 PubMed DOI

Miguelez-Fernandez C, de Leon SJ, Baltasar-Tello I, Peñuelas-Calvo I, Barrigon ML, Capdevila AS, Delgado-Gómez D, Baca-García E, Carballo JJ. Evaluating attention-deficit/hyperactivity disorder using ecological momentary assessment: a systematic review. Atten Defic Hyperact Disord. 2018;10(4):247–265. doi: 10.1007/s12402-018-0261-1.10.1007/s12402-018-0261-1 PubMed DOI

Porras-Segovia A, Díaz-Oliván I, Barrigón ML, Moreno M, Artés-Rodríguez A, Pérez-Rodríguez MM, Baca-García E. Real-world feasibility and acceptability of real-time suicide risk monitoring via smartphones: a 6-month follow-up cohort. J Psychiatr Res. 2022;149:145–154. doi: 10.1016/j.jpsychires.2022.02.026.S0022-3956(22)00107-8 PubMed DOI

Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA. Cochrane Handbook for Systematic Reviews of Interventions. Hoboken, NJ: Wiley Blackwell; 2019.

Baumhauer JF, Bozic KJ. Value-based healthcare: patient-reported outcomes in clinical decision making. Clin Orthop Relat Res. 2016;474(6):1375–1378. doi: 10.1007/s11999-016-4813-4. 10.1007/s11999-016-4813-4 PubMed DOI PMC

Üstün TB, Chatterji JR, Rehm J, editors. Measuring Health and Disability: Manual for WHO Disability Assessment Schedule WHODAS 2.0. Geneva, Switzerland: World Health Organization; 2010.

ICHOM. 2018. [2021-12-07]. https://www.ichom.org/

Paton M, Lane R. Clinimetrics: World Health Organization disability assessment schedule 2.0. J Physiother. 2020;66(3):199. doi: 10.1016/j.jphys.2020.03.002. S1836-9553(20)30019-9 PubMed DOI

Majumder S, Deen MJ. Smartphone sensors for health monitoring and diagnosis. Sensors (Basel) 2019;19(9):2164. doi: 10.3390/s19092164. s19092164 PubMed DOI PMC

Hornstein S, Forman-Hoffman V, Nazander A, Ranta K, Hilbert K. Predicting therapy outcome in a digital mental health intervention for depression and anxiety: a machine learning approach. Digit Health. 2021;7:20552076211060659. doi: 10.1177/20552076211060659. 10.1177_20552076211060659 PubMed DOI PMC

Nahum-Shani I, Smith SN, Spring BJ, Collins LM, Witkiewitz K, Tewari A, Murphy SA. Just-in-time adaptive interventions (JITAIs) in mobile health: key components and design principles for ongoing health behavior support. Ann Behav Med. 2018;52(6):446–462. doi: 10.1007/s12160-016-9830-8. 10.1007/s12160-016-9830-8 PubMed DOI PMC

Berrouiguet S, Ramírez D, Barrigón ML, Moreno-Muñoz P, Carmona Camacho R, Baca-García E, Artés-Rodríguez A. Combining continuous smartphone native sensors data capture and unsupervised data mining techniques for behavioral changes detection: a case series of the Evidence-Based Behavior (eB2) study. JMIR Mhealth Uhealth. 2018;6(12):e197. doi: 10.2196/mhealth.9472. v6i12e197 PubMed DOI PMC

Evidence-Based Behavior (eB2) [2023-10-05]. https://eb2.tech/?lang=en .

Barrigón ML, Berrouiguet S, Carballo JJ, Bonal-Giménez C, Fernández-Navarro P, Pfang B, Delgado-Gómez D, Courtet P, Aroca F, Lopez-Castroman J, Artés-Rodríguez A, Baca-García E, MEmind study group User profiles of an electronic mental health tool for ecological momentary assessment: MEmind. Int J Methods Psychiatr Res. 2017;26(1):e1554. doi: 10.1002/mpr.1554. PubMed DOI PMC

Ferri FJ, Pudil P, Hatef M, Kittler J. Comparative study of techniques for large-scale feature selection. Mach Intell Pattern Recognit. 1994;16:403–413. doi: 10.1016/b978-0-444-81892-8.50040-7. DOI

Hastie T, Tibshirani R, Tibshirani R. Best subset, forward stepwise or lasso? Analysis and recommendations based on extensive comparisons. Statist Sci. 2020;35(4):579–592. doi: 10.1214/19-sts733. DOI

Orrù G, Monaro M, Conversano C, Gemignani A, Sartori G. Machine learning in psychometrics and psychological research. Front Psychol. 2019;10:2970. doi: 10.3389/fpsyg.2019.02970. PubMed DOI PMC

Li P, Liang X, Song H. A survey on implicit bias of gradient descent. 14th International Conference on Computer Research and Development (ICCRD); January 7-9, 2022; Shenzhen, China. IEEE; 2022.

Hawkins DM. The problem of overfitting. ChemInform Wiley. 2004;44(1):1–12. doi: 10.1002/chin.200419274. PubMed DOI

Chai T, Draxler RR. Root mean square error (RMSE) or mean absolute error (MAE)?—arguments against avoiding RMSE in the literature. Geosci Model Dev. 2014;7(3):1247–1250. doi: 10.5194/gmd-7-1247-2014. DOI

Brassington G. Mean absolute error and root mean square error: which is the better metric for assessing model performance?. EGU General Assembly; Proceedings From the Conference 19th EGU General Assembly, EGU2017; April 23-28, 2017; Vienna, Austria. 2017.