We perform research that improves health outcomes and strengthens the sustainability of our health care system. Our progress is founded on artificial intelligence (AI), machine learning and data mining methodologies.
Manual data collection, and prediction for both patient management and research from length patient records is a tedious job for nurses, paramedical staff and doctors. We can develop and evaluate NLP systems to summarize unstructured patient notes in a structured form to help clinical decision making and reducing time of experts in diagnosis and prognosis of the patients. We can also build an automated disease and prognosis based model. In short, a natural language processing tool can be modified to align it for use in the clinical domain and can evaluate the applicability and feasibility in chronic disease management process.
In our approach, firstly the terms that are potential clues for a risk factor are discovered, then we extract additional information about each term (like the subject of a term) and we finally apply our rule module in order to determine if a patient is diagnosed for a specific risk factor. This tool can be user specific and we can develop this for different departments and clinical domains based on user needs. We have already developed and evaluated the tool in chronic renal patients and in dialysis dataset.
1. We are looking to develop, evaluate and determine the effectiveness and measure adverse outcome of adjuvant physiotherapy along with sham or usual care vs sham or usual care alone in chronic non-specific low back adult patients. Specifically, we will: i) Identify the dosage, intensity, frequency and duration of laser therapy and ii) Determine the effectiveness of adjuvant laser therapy in these patients.
Chronic non-specific low back pain is one of the leading causes of disability in the world. Several studies and country and regional specific guidelines have recently shown conflicting evidence and lack of research, highlighting the need for further research on available treatment options for chronic low back pain. To address these gaps in the literature we proposed adjuvant Photobiomodulation therapy (laser therapy) along with sham or usual care vs sham or usual care alone in chronic non-specific low back adult patients. First step will be to develop dosage intensity and frequency along with treatment period and duration of laser therapy. Second step will be to evaluate the effectiveness of the adjuvant laser therapy along with any adverse outcomes.
We explore acceptable dosage, intensity, frequency and duration of laser therapies for different conditions, diagnosis and treatments from the available data. For each use, a range will be develop and side effects will be noted based on retrospective data. Few most related conditions and use of laser therapy will be selected and dosage, intensity, frequency and duration will be noted. Both the identification of retrospective data and exploration of data will be systematic and regression and meta-analysis will be used along with adapting measures of heterogeneity and funnel plots to reduce and minimize error and bias. This will include determining effectiveness and safety of the treatment. A clinical trial will be conducted for this purpose.
2. We are looking to develop, evaluate and determine the effectiveness of antioxidant multivitamins for the treatment of multiple sclerosis. Specifically, we will: i) Identify the dosage, intensity, frequency and duration of antioxidant multivitamins and ii) Determine the effectiveness of antioxidant multivitamins in these patients.
Multiple sclerosis (MS) is a highly unpredictable disease of the central nervous system that commonly causes relapses of neurological symptoms and worsening of symptoms over time. Antioxidant multivitamins are a group of organic compounds that promote cell function, growth and development. Considering the role of antioxidant multivitamins, we will explore use of antioxidant multivitamins for the treatment of multiple sclerosis. First step will be to develop dosage, frequency and duration of treatment along with safety, followed by second step of determining the effectiveness and safety of the different combinations and dosages, frequency and duration of treatment for multiple sclerosis.
We explore acceptable dosage, frequency and duration of antioxidant multivitamins for different conditions, diagnosis and treatments from the available data. For each antioxidant multivitamin their use will be develop and side effects will be noted, based on retrospective data. Few most related conditions and use of antioxidant multivitamins will be selected and dosage, frequency and duration will be noted. Both the identification of retrospective data and exploration of data will be systematic and regression and meta-analysis will be used along with adapting measures of heterogeneity and funnel plots to reduce and minimize error and bias. This will include determining effectiveness and safety of the treatment. A clinical trial will be conducted for this purpose.
3. We are looking to develop a platform where we can have a range of individual variables (demographic, clinical, family etc.) for each patient and this information that can be used for predictive modelling for different diseases and their progression within individual patient.
Predictive analysis is an important component of clinical research. It involves both disease prediction and severity or prognosis. In summary, we will be working on developing a chronic disease management platform that will have data description, analysis and visualization for each patient as well as a group or subgroup of patients.
Several supervised machine learning algorithms such as Import Dataset into Jupyter Notebook Environment, Dataset Preprocessing, Classifier Selection, Parameter Selection and Optimizations, Testing and Performance Metrics and Visualize Results will be used. The best algorithm will be used for predictive analysis.
4. We understand that manual data entry of patient records for reporting to different organizations is a tedious task and often leads to quality issues. Therefore in order to find a solution, we will use a machine learning approach for anomaly detection using similarity measures in complex chronic disease patient data.
We make recommendations for possible values based on predictive models to the clinicians who can either accept or decline the recommendations. In summary, retrospective data of chronic disease patients will be used for anomaly detection problems using supervised, unsupervised and semi-supervised approaches. The historical data will be used to train the supervised machine learning model and will serve as the gold standard.
The common machine learning algorithms for anomaly detection may include Support Vector Machines (SVM), Multi-Layer Perceptron (MLP) and k-Nearest Neighbour (kNN) with local outlier factor. Their performance will be evaluated using a predefined testing framework to identify the best-performing methods. The best method will be used as an effective decision support tool for the respective chronic condition.