Quantifying caregiver workload during the COVID-19 pandemic using ambient intelligence

ABSTRACT

Introduction: The COVID-19 pandemic has increased caregiver workload [1]. It is unclear how this workload is distributed across patients with varying presentations. Ambient intelligence (AmI) utilizes neural networks to monitor multiple data points in video feeds, and automatically tracks various aspects of human movement [2]. AmI was used to examine the workload of healthcare staff in relation to temporal and patient characteristics on a COVID ward at a major metropolitan hospital. Method(s) Sensors were deployed in patient rooms on a COVID ward to detect caregiver visits at 5-min intervals. Electronic medical records were used to identify variables hypothesized to contribute to visits. Result(s) 5514 h across 55 patients (mean age 72, range 17-98) were analyzed. The primary reason for admission was medical in 45 cases (81.8%), psychiatric in 8 cases (14.5%) and surgical in 2 cases (3.6%). Medical emergency (MET) calls occurred in 21 (38.2%) cases. As summarized in Fig. 1, visitation was lowest between 0000 and 0400 (27.3 +/- 1.1 min/hour (min/hr)) and highest between 1200 and 1600 (65.5 +/- 1.3 min/hr). The mean +/- SE visitation spent with medical, psychiatric, and surgical patients was 51.7 +/- 0.6, 38.8 +/- 1.3, and 33.7 +/- 3.9 min/hr respectively (p < 0.005). Overall lowest visitation was in surgical patients between 0400 and 0800 (4.5 +/- 4.5 min/hr). Mean +/- SE visitation were 66.7 +/- 4.6 min/hr in the three hours preceding and following MET calls compared to 50.2 +/- 0.5 min/hr in periods without MET calls (p < 0.005). There was no difference in visitation time between patients with respiratory symptoms and those without (50.7 +/- 0.9 vs 48.8 +/- 0.7 min/hr, p = 0.1). Conclusion(s) AmI can help quantify patient workload, potentially improving staff planning. Further studies comparing healthcare attendance between patients on COVID wards and non-COVID wards may provide insight into the impact of unique factors associated with the pandemic.