MAISON-LLF: Multimodal Sensor Dataset for Monitoring Older Adults Post Lower-Limb Fractures in Community Settings

Dataset Description The MAISON-LLF dataset comprises longitudinal multimodal data collected from 18 community-dwelling older adults living alone following lower limb fractures. Each participant was monitored for eight weeks starting from the first week after hospital discharge. In total, the dataset contains 1008 days of continuous sensor-based measurements, complemented by biweekly clinical questionnaire assessments. Version 4 of the MAISON-LLF dataset is organized using the following directory structure: ├── maison-llf/│ ├── maison-llf-demographics.csv│ ├── maison-llf-features.csv│ ├── maison-llf-feature-descriptions.xlsx Dataset Contents The file maison-llf-demographics.csv contains participant-level demographic information for participants 1 through 18. Variables include sex, age, fracture type, relationship status, education, employment status, and ethnicity. The file maison-llf-features.csv aggregates all daily sensor-derived features and clinical assessment data into a single CSV file. The data represent post-processed features following missing value imputation. Missing values were imputed at the sensor data level prior to feature extraction. For days with missing sensor recordings for a participant, the corresponding sensor values were imputed using the mean of that participant’s other days within the same week and biweekly clinical assessment period. Biweekly clinical questionnaire variables include individual item level scores for the Social Isolation Scale items sis-01 to sis-06, the Oxford Hip Score items ohs-01 to ohs-12, and the Oxford Knee Score items oks-01 to oks-12, along with their corresponding total scores sis, ohs, and oks. The file additionally includes results from the Timed Up and Go test and the 30-second chair stand test. Each participant contributed four sets of clinical assessment data. Within each set, all biweekly clinical variables share the same timestamp, corresponding to the calendar date in yyyy-MM-dd format on which the assessment was conducted. The file maison-llf-feature-descriptions.xlsx provides detailed definitions and descriptions of all daily features included in maison-llf-features.csv. This file represents an updated version of Table 3 from the paper. Citation If you use this dataset, please cite the following paper: Abedi, A., Chu, C.H. & Khan, S.S. Multimodal sensor dataset for monitoring older adults post lower limb fractures in community settings. Sci Data 12, 733 (2025). https://doi.org/10.1038/s41597-025-05069-7