ABSTRACT
Patient safety has constituted a huge public health concern for a long period of time. The focus of safety in the healthcare context is around reducing preventable harms, such as medical errors and treatment-related injuries. COVID-19 pandemic, if anything, has act as a wake-up call for health experts to address latent safety problems. Advancements in the field of artificial intelligence have highlighted the use of intelligent systems as a proven means of improving patient safety and enhancing quality of care. This chapter explores trends in quality and safety research, the use of machine learning and natural language processing in the context of improving patient safety and outcomes, the use of patient safety databases as a source of data for machine learning, and the future of artificial intelligence in quality and safety. © Springer Nature Switzerland AG 2022.
ABSTRACT
Parkinson's disease is a movement disorder that increases fall risk. Clinicians administer several validated gait and balance tests for people with Parkinson's disease in person. COVID-19 has reduced healthcare access, and this has disproportionately affected older populations. We tested the reliability of remote gait and balance assessments of people with Parkinson's disease using face-to-face as the comparator. Fifteen people with Parkinson's disease (aged 57-82, 11 males) performed 14 tests of gait and balance twice: (i) face-to-face, and (ii) remotely, via videoconference between 7 and 14 days after. A trained physiotherapist rated participant performance. The tests included items from the Berg Balance Scale, Functional Gait Assessment, and the Timed-Up- And-Go. These assessments have been validated face-to-face for people with Parkinson's disease. The videoconference assessment was recorded. We compared face-to-face and live videoconference performance to obtain assessment reliability. The physiotherapist rated the recording at least two weeks after the live videoconference to obtain intrarater reliability. A second rater assessed the recording, and we compared live and recorded telehealth assessments to obtain inter-rater reliability. Reliabil ity was measured using either intraclass correlation (ICC) two-way mixed with absolute agreement (continuous measures) or Fleiss multi-rater Kappa test (ordinal measures). Most tests showed moderate to very good assessment reliability between face-to-face and live telehealth (ICC=0.5-1), between face-to-face and recorded telehealth (ICC=0.5-1) and good to very good inter-rater reliability between the recorded telehealth assessments (ICC=0.63-1). Reliability appeared to be higher in tests involving quantitative, rather than qualitative, measures of performance. A ceiling effect was noted in some tests where all participants completed tests with maximum scores in both face-to-face and remote assessments. This study supports the feasibility of remote assessment in clinical practice for people with Parkinson's disease. Further research with a larger cohort and adjustment of the assessments to avoid ceiling effects is necessary.
ABSTRACT
The authors apologize for a typing error that occurred in the September 2020 article that changes the meaning of a sentence. Correction: Page 921, right column, 2nd paragraph, line 8, change «match» to «watch» so it reads, «Primary care providers should watch for frailty development due to physical inactivity during the COVID-19 pandemic (47).» In addition, the author listed as “C. Won Won” wishes to be known as “C.W. Won.” © 2020, The Journal of Nutrition, Health & Aging.