Dataset Information

Using Speech Data From Interactions With a Voice Assistant to Predict the Risk of Future Accidents for Older Drivers: Prospective Cohort Study.

ABSTRACT:

Background

With the rapid growth of the older adult population worldwide, car accidents involving this population group have become an increasingly serious problem. Cognitive impairment, which is assessed using neuropsychological tests, has been reported as a risk factor for being involved in car accidents; however, it remains unclear whether this risk can be predicted using daily behavior data.

Objective

The objective of this study was to investigate whether speech data that can be collected in everyday life can be used to predict the risk of an older driver being involved in a car accident.

Methods

At baseline, we collected (1) speech data during interactions with a voice assistant and (2) cognitive assessment data-neuropsychological tests (Mini-Mental State Examination, revised Wechsler immediate and delayed logical memory, Frontal Assessment Battery, trail making test-parts A and B, and Clock Drawing Test), Geriatric Depression Scale, magnetic resonance imaging, and demographics (age, sex, education)-from older adults. Approximately one-and-a-half years later, we followed up to collect information about their driving experiences (with respect to car accidents) using a questionnaire. We investigated the association between speech data and future accident risk using statistical analysis and machine learning models.

Results

We found that older drivers (n=60) with accident or near-accident experiences had statistically discernible differences in speech features that suggest cognitive impairment such as reduced speech rate (P=.048) and increased response time (P=.040). Moreover, the model that used speech features could predict future accident or near-accident experiences with 81.7% accuracy, which was 6.7% higher than that using cognitive assessment data, and could achieve up to 88.3% accuracy when the model used both types of data.

Conclusions

Our study provides the first empirical results that suggest analysis of speech data recorded during interactions with voice assistants could help predict future accident risk for older drivers by capturing subtle impairments in cognitive function.

SUBMITTER: Yamada Y

PROVIDER: S-EPMC8063093 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundOur aim was to evaluate Spanish family doctors' knowledge about medications that increase the risk of traffic accidents involving older drivers, and to obtain data about the involvement of family doctors in accident prevention activities and the associations between these factors and their demographic and workplace characteristics.MethodsA cross-sectional study of 1888 family doctors throughout Spain was carried out from 2016 to 2018. Participants completed a previously validated self-administered questionnaire that explored whether family doctors distinguished between medications associated with a high or low risk of involvement in a traffic accident, investigated the appropriateness of advice given to older patients, and physicians' involvement in preventive activities. Multiple regression models were used to estimate the adjusted association of these variables with each other and with characteristics of family doctors in the sample.ResultsOn a scale of 1 (never or hardly ever) to 4 (always), the indexes constructed to evaluate how often family doctors believed they should oversee the use of high-risk and low-risk medications yielded values of 3.38 for the former and 2.61 for the latter (p < 0.001). Only 24% responded correctly to all three items that inquired about the appropriateness of the advice they gave to older patients. On a scale of 1 to 4, the frequency at which family doctors gave older patients advice about preventive measures was 2.85, and only 43% reported allocating time during appointments to provide this advice. These latter two variables were directly associated with appropriate values for the index used to evaluate physicians' oversight of medications associated with a high risk. The perception of risk associated with medications and involvement in preventive activities were both greater among female participants.ConclusionsFamily doctors correctly identified medications according to their risk of playing a role in traffic accidents, although the recommendations they gave to their patients were not always appropriate. These findings, along with physicians' infrequent involvement in preventive activities, suggest a need to improve family doctors' competencies and increase the resources available to them so that they can provide their older patients with advice on ways to prevent involvement in traffic accidents.

Project description:BackgroundPain is difficult to manage in older adults. It has been recommended that pain management in older adults should include both nonpharmacologic and pharmacologic strategies. Unfortunately, nonadherence to pain medication is more prevalent than nonadherence to any other chronic disease treatment. Technology-based reminders have some benefit for medication adherence, but adherence behavior outcomes have mostly been verified by self-reports.ObjectiveWe aimed to describe objective medication adherence and the latency of medication use after a voice assistant reminder prompted participants to take pain medications for chronic pain.MethodsA total of 15 older adults created a voice assistant reminder for taking scheduled pain medications. A subsample of 5 participants were randomly selected to participate in a feasibility study, in which a medication event monitoring system for pain medications was used to validate medication adherence as a health outcome. Data on the subsample's self-assessed pain intensity, pain interference, concerns and necessity beliefs about pain medications, self-confidence in managing pain, and medication implementation adherence were analyzed.ResultsIn the 5 participants who used the medication event monitoring system, the overall latency between voice assistant reminder deployment and the medication event (ie, medication bottle cap opening) was 55 minutes. The absolute latency (before or after the reminder) varied among the participants. The shortest average time taken to open the cap after the reminder was 17 minutes, and the longest was 4.5 hours. Of the 168 voice assistant reminders for scheduled pain medications, 25 (14.6%) resulted in the opening of MEMS caps within 5 minutes of the reminder, and 107 (63.7%) resulted in the opening of MEMS caps within 30 minutes of the reminder.ConclusionsVoice assistant reminders may help cue patients to take scheduled medications, but the timing of medication use may vary. The timing of medication use may influence treatment effectiveness. Tracking the absolute latency time of medication use may be a helpful method for assessing medication adherence. Medication event monitoring may provide additional insight into medication implementation adherence during the implementation of mobile health interventions.

Project description:ObjectiveDigital voice assistant technology provides unique opportunities to enhance clinical practice. We aimed to understand factors influencing pediatric providers' current and potential use of this technology in clinical practice.MethodsWe surveyed pediatric providers regarding current use and interest in voice technology in the workplace. Regression analyses evaluated provider characteristics associated with voice technology use. Among respondents not interested in voice technology, we elicited individual concerns.ResultsAmong 114 respondents, 19 (16.7%) indicated current use of voice technology in clinical practice, and 51 (44.7%) indicated use of voice technology for nonclinical purposes. Fifty-four (47.4%) reported willingness to try digital voice assistant technology in the clinical setting. Providers who had longer clinic visits (odds ratio [OR], 3.11, 95% confidence interval [CI], 1.04, 9.33, p?=?0.04), fewer patient encounters per day (p?=?0.02), and worked in hospital-based practices (OR, 2.95, 95% CI, 1.08, 8.07, p?=?0.03) were more likely to currently use voice technology in the office. Younger providers (p?=?0.02) and those confident in the accuracy of voice technology (OR, 3.05, 95% CI, 1.38, 6.74, p?=?0.005) were more willing to trial digital voice assistants in the clinical setting. Among respondents unwilling or unsure about trying voice assistant technology, the most common reasons elicited were concerns related to its accuracy (35%), efficiency (33%), and privacy (28%).ConclusionThis national survey evaluating use and attitudes toward digital voice assistant technology by pediatric providers found that while only one-eighth of pediatric providers currently use digital voice assistant technology in the clinical setting, almost half are interested in trying it in the future. Younger provider age and confidence in the accuracy of voice technology are associated with provider interest in using voice technology in the clinical setting. Future development of voice technology for clinical use will need to consider accuracy of information, efficiency of use, and patient privacy for successful integration into the workplace.