Development of a Smart Hospital Assistant: Integrating Artificial Intelligence and a Voice-User Interface for Improved Surgical Outcomes.

Patient safety and efficiency are top priorities in any surgical procedure. One effective way to achieve these objectives is to automate the logistical and routine tasks that occur in the operating suite. Inspired by smart assistant technology already widely used in the consumer sector, we engineered the Smart Hospital Assistant (SHA), a smart, voice-controlled virtual assistant that handles natural speech recognition while executing non-surgical functions to aid any surgery. In simulated procedures, the SHA reduced operating time, optimized surgical staff resources, and reduced the number of major touch-points that can lead to surgical site infections. The SHA holds promise not only for use in the operating theater, but also in understaffed healthcare environments where automation can improve healthcare delivery.

Evaluating the Effects of Cerebrospinal Fluid Protein Content on the Performance of Differential Pressure Valves and Antisiphon Devices Using a Novel Benchtop Shunting Model.

BACKGROUND: Hydrocephalus is managed by surgically implanting flow-diversion technologies such as differential pressure valves and antisiphoning devices; however, such hardware is prone to failure. Extensive research has tested them in flow-controlled settings using saline or de-aerated water, yet little has been done to validate their performance in a setting recreating physiologically relevant parameters, including intracranial pressures, cerebrospinal fluid (CSF) protein content, and body position. OBJECTIVE: To more accurately chart the episodic drainage characteristics of flow-diversion technology. A gravity-driven benchtop model of flow was designed and tested continuously during weeks-long trials. METHODS: Using a hydrostatic pressure gradient as the sole driving force, interval flow rates of 6 valves were examined in parallel with various fluids. Daily trials in the upright and supine positions were run with fluid output collected from distal catheters placed at alternating heights for extended intervals. RESULTS: Significant variability in flow rates was observed, both within specific individual valves across different trials and among multiple valves of the same type. These intervalve and intravalve variabilities were greatest during supine trials and with increased protein. None of the valves showed evidence of overt obstruction during 30 d of exposure to CSF containing 5 g/L protein. CONCLUSION: Day-to-day variability of ball-in-cone differential pressure shunt valves may increase overdrainage risk. Narrow-lumen high-resistance flow control devices as tested here under similar conditions appear to achieve more consistent flow rates, suggesting their use may be advantageous, and did not demonstrate any blockage or trend of decreasing flow over the 3 wk of chronic use.

DEVELOPMENT OF VOICE-CONTROLLED SMART SURGICAL BED.

Wasted time in the operating room results in higher operating costs and greater post-operative complications. One effective way to reduce operation time is automating basic processes that occur during surgery. Given the rise of smart-home devices, implementation of virtual assistants became a feasible solution in many medical settings. With a consumer smart-home device and off-the-shelf components, we engineered a voice-controlled smart surgical bed that adjusts the bed configuration via a voice input. The resulting device is expected to optimize human resources and reduce surgical site infection by eliminating the need of a traditional touch control mechanism. Future work is needed to develop its proprietary hardware and software, and continuous collaboration with medical personnel to bring this device into market.