TILES-2018, a longitudinal physiologic and behavioral data set of hospital workers

Abstract

We present a novel longitudinal multimodal corpus of physiological and behavioral data collected from direct clinical providers in a hospital workplace. We designed the study to investigate the use of off-the-shelf wearable and environmental sensors to understand individual-specific constructs such as job performance, interpersonal interaction, and well-being of hospital workers over time in their natural day-to-day job settings. We collected behavioral and physiological data from n = 212 participants through Internet-of-Things Bluetooth data hubs, wearable sensors (including a wristband, a biometrics-tracking garment, a smartphone, and an audio-feature recorder), together with a battery of surveys to assess personality traits, behavioral states, job performance, and well-being over time. Besides the default use of the data set, we envision several novel research opportunities and potential applications, including multi-modal and multi-task behavioral modeling, authentication through biometrics, and privacy-aware and privacy-preserving machine learning.

Fig. 1

Experimental design. Participants received instructions in a 2-hour on-boarding session, where they completed the first part of the baseline survey and were instructed in the use of sensors and smartphone apps. This session was followed by the second part of the baseline survey and then by 10 weeks of data collection, during which participants wore multiple wearable sensors (wristband, garment and an audio badge) and answered two daily EMAs (Ecological Momentary Assessments) through their personal smartphones. During the off-boarding session, participants handed in their sensors and finished uploading data. After the sensor data collection, they completed a post-study survey.

Fig. 2

Positioning of Owl-in-Ones (Bluetooth hubs) within a nursing unit. Rooms shaded in gray contain an Owl-in-One, while the blue circles show their exact locations. pat denotes a patient room, ns a nursing station, med a medication room, and lounge represents the lounge or break area for the workers.

Fig. 3

Data flow. This diagram shows the data flow from the sensors given to participants, the sensors placed at USC’s Keck Hospital, and smartphones to the research server, where the data is stored for long-term use.

Fig. 4

Daily sensor usage. These histograms show the average number of hours each participant wore the wearable sensors per day throughout the study.

Fig. 5

Owl-in-One network performance. (a) Shows a typical owl setup in a nursing unit. Here, a single packet is transmitted (in this case, by purple square) and received and decoded by several other owls (as blue circles, the numbers correspond to RSSI values of decodings at the receivers). These packets contain the sender and receivers’ directories. Some of these packets were processed and stored by our pipeline with corrupted information due to transmission errors. (b) (top) total number of distinct packets sent daily in the full Owl-in-One network (purple) and total number of decodings by all receiving owls (blue) and (bottom) proportion of packets whose sender directory was corrupt and therefore lost among all packets sent (purple) and receiver directory was corrupt and lost (blue) among all receiver decodings.

Fig. 6

Cronbach’s α for each EMA. The diameter of the circles is linearly related to the number of observations for the calculation of each α value. Each row of scatter plots corresponds to constructs assessed the same number of times during the data collection.

Fig. 7

Number of participants with up to the specified percentage of surveys started for each survey type. PF: Psychological flexibility; PC: Psychological capital.

Fig. 8

Median EMA response times across participants. Each line shows the median of the average response times per question in a given scale for all participants. We include the baseline EMA questions present in the job, health, and personality surveys, which were asked on a daily basis.