ABSTRACT
Symptom-based machine learning models for disease detection are a way to reduce the workload of doctors when they have too many patients. Currently, there are many research studies on machine learning or deep learning for disease detection or clinical departments classification, using text of patient's symptoms and vital signs. In this study, we used the Long Short-term Memory (LSTM) with a fully connected neural network model for classification, where the LSTM model was used to receive the patient's symptoms text as input data. The fully connected neural network was used to receive other input data from the patients, including body temperature, age, gender, and the month the patients received care in. In this research, a data preprocessing algorithm was improved by using keyword selection to reduce the complexity of input data for overfitting problem prevention. The results showed that the LSTM with fully connected neural network model performed better than the LSTM model. The keyword selection method also increases model performance.
ABSTRACT
This report aims to provide practical advice about the implementation of a public health monitoring system using both geographic information system technology and mobile health, a term used for healthcare delivery via mobile devices. application amongst household residents and community stakeholders in the limited resource community. A public health monitoring system was implemented in a semi-rural district in Thailand. The challenges encountered during implementation were documented qualitatively in a series of monthly focus group discussions, several community hearings, and many targeted interviews. In addition, lessons learned from the expansion of the program to 75 other districts throughout Thailand were also considered. All challenges proved solvable yielding several key pieces of advice for future project implementation teams. Specifically, communication between team members, anticipating technological challenges, and involvement of community members are critical. The problems encountered in our project were mainly related to the capabilities of the data collectors and technical issues of mobile devices, internet coverage, and the GIS application itself. During the implementation phase, progressive changes needed to be made to the system promptly, in parallel with community team building in order to get the highest public health impact.
Subject(s)
Mobile Applications , Telemedicine , Delivery of Health Care , Humans , Technology , ThailandABSTRACT
BACKGROUND: Thoracolumbar junction pure bilateral facet joint dislocation without facet fracture is an extremely rare injury. A current review of thoracolumbar junction pure bilateral facet joint dislocation reported less than 15 cases in which the surgeon had a difficulty with the dissection of surgical planning using only 2D radiographic film and the axial or coronal view in computerized tomography. Bilateral pure facet joint dislocation in the thoracolumbar junction without facet fracture is difficult to understand the morphology of bone injuries. CASE PRESENTATION: A 25-year-old Thai gentleman presented with paraplegia and loss of sensation in the lower extremity (ASIA A) following a fall from a high lorry. Radiographic film and computed tomography scan revealed pure facet dislocation T11-T12 without facet fracture. The patient's thoracolumbar junction of the spine is presented to describe the three-dimensional (3D) printing technique for surgical preoperative planning. After the patient underwent open reduction, decompression and instrumentation with posterolateral fusion, the patient's thoracolumbar junction was described in the three-dimensional (3D) printing again for follow-up and in order to help the surgeon understand about the morphology and alignment after surgery. CONCLUSION: Pure facet dislocation is rarely seen at the thoracolumbar junction; it is a very unstable injury. In this case, we performed an early investigation using a 3D digital printing model in order to help with orthopedic surgical planning, emergency early open reduction and instrumentation with fusion. Neurological status was recovered. The 3D digital printing model should be a standard investigation in rare cases of orthopedic surgical planning.
ABSTRACT
Early warning systems for outbreak detection is a challenge topic for researchers in the epidemiology and biomedical informatics fields. We are proposing a new method for detecting disease epidemics using a symptom-based approach. The data was collected from developed mobile applications which include users' demographic information and a list of chief complaint symptoms. Deliberated outbreaks are differentiated from seasonal outbreak by specific symptoms that represent a sign of infection. These symptoms were grouped, classified, and then converted to a time-series digital signal using the consensus scoring approach. Through the syndromic grouping method, the system digitized each data package into a single independent variable that is ready for further one-dimensional signal processing to predict disease outbreaks in the future.
