Partridge, Thomas J.
(2021)
Optimising route comfort indices for neonatal transfers by road.
PhD thesis, University of Nottingham.
Abstract
The risk of severe brain injuries in sick premature infants increases when transferred between hospitals. Causality is uncertain, but stress levels are elevated during ambulance journeys; potentially due to excessive levels of noise and vibration. It has been proposed that reducing these levels would reduce the risk, with one prospective method being comfort-optimised navigation.
An Android app was developed that logs noise level, Inertial Measurement Unit (IMU) and location data during journeys, sampling at the fastest rates possible depending on the hardware and firmware. The smartphone used during development was found to sample noise levels accurate to 0.3 dB up to 80 dB(A) and accelerations accurate to 10\% up to 40~Hz, although considerable jitter was present in the IMU sampling. Recorded data were shown to be repeatable for multiple passes over the same stretch of road (acceleration interquartile range (IQR): 0.14ms^{-2}; noise IQR: 2.8 dB). Data were influenced by both supplementary audio and the smartphone model so an initial idea of gathering data through public engagement was determined unsuitable.
Controlled collection of data was planned, utilising the neonatal ambulances operated by CenTre Neonatal Transport (CenTre). A new smartphone model was identified that was capable of sampling accelerations at a sufficient rate to comply with the "Evaluation of human exposure to whole-body vibration" standard, ISO 2631. This model also had greater processing power than the previous model used during initial testing, resulting in reduced jitter, and was found to provide more accurate accelerations (within 5% up to 55 Hz). Logging of periods before and after each journey was added along with meta-data describing each journey.
Journeys performed by CenTre were recorded over the course of 12 months. Recorded variables were supplemented by calculation of ISO-weighted vibration parameters. The final dataset comprises 1,487 journeys over 81,901 km and 1,318 hours. Strong similarities between meta-data and officially reported transport data suggested there was no bias in the journeys that the staff recorded.
Roads driven between Nottingham City Hospital (NCH) and Leicester Royal Infirmary (LRI) were chosen as a case study. Data from 588 journeys contributed towards the analysis.
A range of metrics, derived from previous studies and adult standards, were used to assess the roads of the NCH to LRI network. Both speed and road classification were found to influence vibration and noise level, however the influence could not be separated due to the inherent link between both parameters. All routes involved either use of motorway or a concrete A-road, with the latter producing worse vibration. Although individual road sections varied, differences were reduced between the routes.
Assessments were also performed of the metrics at each of the 42 hospitals (36 departing; 38 arriving) present in the data. Results were similar between hospitals, but differed between loading and unloading phases. High magnitude shocks were more abundant during the loading phases, whereas low impact vibrations were more frequent during unloading. Both phases registered greater shocks than those found during journeys.
In summary, this work provides a low-cost method of obtaining large amounts of data describing the ambulance environment without requiring any technical knowledge to operate. The theory that the physical environment could be altered through routing has also been confirmed. The data collected during this work could be utilised in the future to aid determination of neonatal responses and subsequently establish optimal routes.
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
