Evaluation of an accelerometer-based monitor for detecting bed net use and human entry/exit using a machine learning algorithm.

BACKGROUND: Distribution of long-lasting insecticidal bed nets (LLINs) is one of the main control strategies for malaria. Improving malaria prevention programmes requires understanding usage patterns in households receiving LLINs, but there are limits to what standard cross-sectional surveys of self-reported LLIN use can provide. This study was designed to assess the performance of an accelerometer-based approach for measuring a range of LLIN use behaviours as a proof of concept for more granular LLIN-use monitoring over longer time periods. METHODS: This study was carried out under controlled conditions from May to July 2018 in Liverpool, UK. A single accelerometer was affixed to the side panel of an LLIN and participants carried out five LLIN use behaviours: (1) unfurling a net; (2) entering an unfurled net; (3) lying still as if sleeping; (4) exiting from under a net; and, (5) folding up a net. The randomForest package in R, a supervised non-linear classification algorithm, was used to train models on 20-s epochs of tagged accelerometer data. Models were compared in a validation dataset using overall accuracy, sensitivity and specificity, receiver operating curves and the area under the curve (AUC). RESULTS: The five-category model had overall accuracy of 82.9% in the validation dataset, a sensitivity of 0.681 for entering a net, 0.632 for exiting, 0.733 for net down, and 0.800 for net up. A simplified four-category model, combining entering/exiting a net into one category had accuracy of 94.8%, and increased sensitivity for net down (0.756) and net up (0.829). A further simplified three-category model, identifying sleeping, net up, and a combined net down/enter/exit category had accuracy of 96.2% (483/502), with an AUC of 0.997 for net down and 0.987 for net up. Models for detecting entering/exiting by adults were significantly more accurate than for children (87.8% vs 70.0%; p < 0.001) and had a higher AUC (p = 0.03). CONCLUSIONS: Understanding how LLINs are used is crucial for planning malaria prevention programmes. Accelerometer-based systems provide a promising new methodology for studying LLIN use. Further work exploring accelerometer placement, frequency of measurements and other machine learning approaches could make these methods even more accurate in the future.

The impact of pill counting on resource-limited health facilities: a thematic qualitative analysis in eSwatini.

Research on health systems in resource-limited settings has garnered considerable attention, but the dispensing of individual prescriptions has not been thoroughly explored as a specific bottleneck to effective delivery of care. The rise of human immunodeficiency virus/tuberculosis prevalence and non-communicable diseases in the Kingdom of eSwatini has introduced significant pressures on health facilities to meet patient demands for lifelong medications. Because automated pill counting methods are impracticable and expensive, most prescriptions are made by means of manually counting individual prescriptions using a plastic dish and spatula. The aim of this work was to examine the perceptions of health providers of causes for pill counting errors, and pill counting's impact on clinic workflow. Our study took place in 13 randomly selected public health facilities in eSwatini, stratified by three groups based on monthly patient volumes. Thirty-one participants who count pills regularly and 13 clinic supervisors were interviewed with semi-structured materials and were audio-recorded for later transcription. Interviews were thematically analysed with inductive coding and three major themes emerged: workflow, counting error causes and effect on clinic function. Findings demonstrate large variety in how facilities manage pill counting for prescription making. Due to patient demands, most facilities utilize all available personnel, from cleaners to nurses, to partake in prescription making. Major causes for pill counting errors were distractions, exhaustion and being hurried. Participants mentioned that patients said that they had initially received the wrong quantity of pills and this affected medication adherence measurements based off pill counts. Most participants described how efforts put into pill counting detracted from their work performance, wasted valuable time and increased patient wait times. Future research is needed to quantify prescription accuracy, but our data suggest that interventions directly alleviating the burden of pill counting could lead to improved clinic quality and possibly improve patient outcomes.

Comparison of traditional methods versus SAFEcount for filling prescriptions: A pilot study of an innovative pill counting solution in eSwatini.

BACKGROUND: Packaging medications is a crucial component of health system efficiency and quality. In developing countries, medications often arrive in bulk containers that need to be counted by hand. Traditional counting is time-consuming, inaccurate and tedious. SAFEcount is a novel and inexpensive handheld device that may improve the accuracy and speed of pill-counting in resource limited settings. We designed a head-to-head trial to compare traditional and SAFEcount prescription filling in eSwatini. METHODS: We recruited 31 participants from 13 health facilities throughout eSwatini. Speed and accuracy for each prescription was recorded while each participant filled prescriptions of various quantities using both the traditional and SAFEcount methods. RESULTS: Traditional pill counting resulted in an error rate of 12.6% inaccurate prescriptions compared to 4.8% for SAFEcount (p<0.0001). SAFEcount was 42.3% faster than traditional counting (99.9 pills per minute versus 70.2; p<0.0001). Using SAFEcount was preferred over traditional pill counting by 97% (29/30) of participants. CONCLUSIONS: The SAFEcount device is a preferred alternative by counting personnel and is significantly faster and more accurate compared to traditional counting methods. SAFEcount could help improve the efficiency and quality of health care delivery in place of traditional hand counting.