RESUMO
BACKGROUND: In many low and middle-income countries (LMICs), difficulties in patient identification are a major obstacle to the delivery of longitudinal care. In absence of unique identifiers, biometrics have emerged as an attractive solution to the identification problem. We developed an mHealth App for subject identification using pattern recognition around ear morphology (Project SEARCH (Scanning EARS for Child Health). Early field work with the SEARCH App revealed that image stabilization would be required for optimum performance. METHODS: To improve image capture, we designed and tested a device (the 'Donut'), which standardizes distance, angle, rotation and lighting. We then ran an experimental trial with 194 participants to measure the impact of the Donut on identification rates. Images of the participant's left ear were taken both with and without use of the Donut, then processed by the SEARCH algorithm, measuring the top one and top ten most likely matches. RESULTS: With the Donut, the top one identification rate and top ten identification rates were 99.5 and 99.5%, respectively, vs. 38.4 and 24.1%, respectively, without the Donut (P < 0.0001 for each comparison). In sensitivity analyses, crop technique during pre-processing of images had a powerful impact on identification rates, but this too was facilitated through the Donut. CONCLUSIONS: By standardizing lighting, angle and spatial location of the ear, the Donut achieved near perfect identification rates on a cohort of 194 participants, proving the feasibility and effectiveness of using the ear as a biometric identifier. TRIAL REGISTRATION: This study did not include a medical intervention or assess a medical outcome, and therefore did not meet the definition of a human subjects research study as defined by FDAAA. We did not register our study under clinicaltrials.gov .
Assuntos
Identificação Biométrica , Orelha , Saúde Global , Sistemas de Identificação de Pacientes , Adulto , Algoritmos , Feminino , Humanos , Masculino , TelemedicinaRESUMO
Pneumonia is the leading cause of death among children around the world. According to WHO, a total of 740,180 lives under the age of five were lost due to pneumonia in 2019. Lung ultrasound (LUS) has been shown to be particularly useful for supporting the diagnosis of pneumonia in children and reducing mortality in resource-limited settings. The wide application of point-of-care ultrasound at the bedside is limited mainly due to a lack of training for data acquisition and interpretation. Artificial Intelligence can serve as a potential tool to automate and improve the LUS data interpretation process, which mainly involves analysis of hyper-echoic horizontal and vertical artifacts, and hypo-echoic small to large consolidations. This paper presents, Fused Lung Ultrasound Encoding-based Transformer (FLUEnT), a novel pediatric LUS video scoring framework for detecting lung consolidations using fused LUS encodings. Frame-level embeddings from a variational autoencoder, features from a spatially attentive ResNet-18, and encoded patient information as metadata combiningly form the fused encodings. These encodings are then passed on to the transformer for binary classification of the presence or absence of consolidations in the video. The video-level analysis using fused encodings resulted in a mean balanced accuracy of 89.3 %, giving an average improvement of 4.7 % points in comparison to when using these encodings individually. In conclusion, outperforming the state-of-the-art models by an average margin of 8 % points, our proposed FLUEnT framework serves as a benchmark for detecting lung consolidations in LUS videos from pediatric pneumonia patients.
Assuntos
Pulmão , Pneumonia , Ultrassonografia , Humanos , Ultrassonografia/métodos , Pulmão/diagnóstico por imagem , Pneumonia/diagnóstico por imagem , Criança , Pré-Escolar , Interpretação de Imagem Assistida por Computador/métodos , Masculino , Feminino , Lactente , Gravação em VídeoRESUMO
OBJECTIVES: To determine the prevalence of COVID-19 postmortem setting in Lusaka, Zambia. DESIGN: A systematic, postmortem prevalence study. SETTING: A busy, inner-city morgue in Lusaka. PARTICIPANTS: We sampled a random subset of all decedents who transited the University Teaching Hospital morgue. We sampled the posterior nasopharynx of decedents using quantitative PCR. Prevalence was weighted to account for age-specific enrolment strategies. INTERVENTIONS: Not applicable-this was an observational study. PRIMARY OUTCOMES: Prevalence of COVID-19 detections by PCR. Results were stratified by setting (facility vs community deaths), age, demographics and geography and time. SECONDARY OUTCOMES: Shifts in viral variants; causal inferences based on cycle threshold values and other features; antemortem testing rates. RESULTS: From 1118 decedents enrolled between January and June 2021, COVID-19 was detected among 32.0% (358/1116). Roughly four COVID-19+ community deaths occurred for every facility death. Antemortem testing occurred for 52.6% (302/574) of facility deaths but only 1.8% (10/544) of community deaths and overall, only ~10% of COVID-19+ deaths were identified in life. During peak transmission periods, COVID-19 was detected in ~90% of all deaths. We observed three waves of transmission that peaked in July 2020, January 2021 and ~June 2021: the AE.1 lineage and the Beta and Delta variants, respectively. PCR signals were strongest among those whose deaths were deemed 'probably due to COVID-19', and weakest among children, with an age-dependent increase in PCR signal intensity. CONCLUSIONS: COVID-19 was common among deceased individuals in Lusaka. Antemortem testing was rarely done, and almost never for community deaths. Suspicion that COVID-19 was the cause of deaths was highest for those with a respiratory syndrome and lowest for individuals <19 years.
Assuntos
COVID-19 , Criança , Humanos , COVID-19/diagnóstico , COVID-19/epidemiologia , Zâmbia/epidemiologia , Prevalência , SARS-CoV-2 , Reação em Cadeia da Polimerase , Teste para COVID-19RESUMO
OBJECTIVE: To directly measure the fatal impact of coronavirus disease 2019 (covid-19) in an urban African population. DESIGN: Prospective systematic postmortem surveillance study. SETTING: Zambia's largest tertiary care referral hospital. PARTICIPANTS: Deceased people of all ages at the University Teaching Hospital morgue in Lusaka, Zambia, enrolled within 48 hours of death. MAIN OUTCOME MEASURE: Postmortem nasopharyngeal swabs were tested via reverse transcriptase quantitative polymerase chain reaction (PCR) against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Deaths were stratified by covis-19 status, location, age, sex, and underlying risk factors. RESULTS: 372 participants were enrolled between June and September 2020; PCR results were available for 364 (97.8%). SARS-CoV-2 was detected in 58/364 (15.9%) according to the recommended cycle threshold value of <40 and in 70/364 (19.2%) when expanded to any level of PCR detection. The median age at death among people with a positive test for SARS-CoV-2 was 48 (interquartile range 36-72) years, and 69% (n=48) were male. Most deaths in people with covid-19 (51/70; 73%) occurred in the community; none had been tested for SARS-CoV-2 before death. Among the 19/70 people who died in hospital, six were tested before death. Among the 52/70 people with data on symptoms, 44/52 had typical symptoms of covid-19 (cough, fever, shortness of breath), of whom only five were tested before death. Covid-19 was identified in seven children, only one of whom had been tested before death. The proportion of deaths with covid-19 increased with age, but 76% (n=53) of people who died were aged under 60 years. The five most common comorbidities among people who died with covid-19 were tuberculosis (22; 31%), hypertension (19; 27%), HIV/AIDS (16; 23%), alcohol misuse (12; 17%), and diabetes (9; 13%). CONCLUSIONS: Contrary to expectations, deaths with covid-19 were common in Lusaka. Most occurred in the community, where testing capacity is lacking. However, few people who died at facilities were tested, despite presenting with typical symptoms of covid-19. Therefore, cases of covid-19 were under-reported because testing was rarely done not because covid-19 was rare. If these data are generalizable, the impact of covid-19 in Africa has been vastly underestimated.
Assuntos
COVID-19/mortalidade , SARS-CoV-2/isolamento & purificação , Adulto , Distribuição por Idade , Fatores Etários , Idoso , Autopsia , COVID-19/diagnóstico , COVID-19/virologia , Teste de Ácido Nucleico para COVID-19/estatística & dados numéricos , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Nasofaringe/virologia , Estudos Prospectivos , Fatores de Risco , SARS-CoV-2/genética , Fatores Sexuais , População Urbana/estatística & dados numéricos , Zâmbia/epidemiologiaRESUMO
Patient identification in low- to middle-income countries is one of the most pressing public health challenges of our day. Given the ubiquity of mobile phones, their use for health-care coupled with a biometric identification method, present a unique opportunity to address this challenge. Our research proposes an Android-based solution of an ear biometric tool for reliable identification. Unlike many popular biometric approaches (e.g., fingerprints, irises, facial recognition), ears are noninvasive and easily accessible on individuals across a lifespan. Our ear biometric tool uses a combination of hardware and software to identify a person using an image of their ear. The hardware supports an image capturing process that reduces undesired variability. The software uses a pattern recognition algorithm to transform an image of the ear into a unique identifier. We created three cross-sectional datasets of ear images, each increasing in complexity, with the final dataset representing our target use-case population of Zambian infants (N=224, aged 6days-6months). Using these datasets, we conducted a series of validation experiments, which informed iterative improvements to the system. Results of the improved system, which yielded high recognition rates across the three datasets, demonstrate the feasibility of an Android ear biometric tool as a solution to the persisting patient identification challenge.