Dosage prediction in Pediatrics

Project Title

Dosage prediction in Pediatrics

Project Area of Specialization Artificial IntelligenceProject Summary

Irrational use of medicines is a major problem worldwide, World Health Organization (WHO) estimates more than 50% of all medicines are prescribed, dispensed or sold inappropriately. The overuse, underuse or misuse of medicines not only waste medical resources but also lead to significant patient harm in terms of medication error and adverse drug effect. As the pharmacokinetics and pharmacodynamics of the pediatric group is highly dynamic, it is great challenge to determine the rational dosage for pediatrics patients. Traditional clinical decision support systems for dosage guidance largely rely on manual collection of medication information, which usually suffers from incomplete and missing evidences for the pediatric group. In this project we propose a data-driven approach to accurately predict/prescribe medication dosages. More specifically we first identify two relevant factors of pediatric medication doages i,e the physiology factors including patients body weight and age group, and the indication factors that effect clinical dosage patterns. Children are at higher risk of irrational use of medicines that adults. Due to lack of dedicated medication instructions and standards for children, such pediatric medication information is usually incomplete. The rules employed to guide medication dosages for adults are usually very simple e.g. 1 capsule 3 times a daily, which do not take into consideration for dosage adjustment based on children's age, body weight and indication information. 

Project Objectives

 We propose a pediatric dosage prediction framework, at the learning stage, by conducting emperical studies on historical we first identify two relevant factors characterized by age group and body weight, and the indication factor noted in diagnosis. Propose a hybrid learning model to adaptively integrate two heterogeneous features. Finally, we exploit the rational dosages from historical prescription data to train a model for each medicine. At the prediction/prescription stage we employ the trained hybrid learning model to predict the rational dosages for each medicine listed in the prescription and provide dosage suggestions to physicians and pharmacists. We elaborate the details of the key components in the following sections. 

Dosage relevant factor identification: Pediatric medication dosage calculation is more complicated than that of adults. In clinical practice, most pediatric medicines are dosed according to patient's body weight (mg/kg) or body surface are (mg/m2). Moreover, dosages also vary by the patient’s symptoms and indication, therefore diagnostic information is helpful when calculating dosages. Based upon this prior knowledge, we first identify two relevant factors of pediatric medication dosages, and then extract the corresponding features from prescription data via correlation analysis. 

To this end we conduct correlation analysis of the rational dosage against the age group and body weight. These high correlation analysis indices indicate that rational pediatric dosages are highly correlated to patients age group and body weight. 

One of the intuitive methods is to concatenate the physiology and indication features into a vector and build a regression or classification model to predict the rational dosage. 

Project Implementation Method

Dosage Prediction in Peds can be implemented in many ways like Mobile Application, Desktop application and in hospitals to support traditional clinical decision system. But our aim is to develop a Mobile application, which is easy to use, easily reachable. 

Mobile Application Will get information from user. User can be patient or doctor. 

User will fill required fields of mobile application. User will enter physical features i.e. age, weight and indication features. 

Separate fields are created for getting input of age, body weight and for indication features.

Benefits of the Project

In this project, we investigate one of the key problem pediatric medication i.e. rational dosage prediction. We proposed data driven approach to accurately predict pediatric dosages. More specifically, we first identify two relevant factors of pediatric medication dosages, and then extract two corresponding features i.e. the physiology features and the indication features. We then proposed a hybrid learning based model for accurate dosage prediction. We evaluate our method on real world prescription datasets. 

As estimated by World Health Organization 50% of all medicines are misused, overused and underdosed, so with the accurate dosage Prediction can save huge amount of medical resources can be saved.

Medication errors (ME) can be reduced.

Irrational use of medicines also causes the adverse drug effect, drug effects are overcome. 

We exploit the traditional clinical pediatric medication dosing experiences of physicians and pharmacists from historical prescriptions to train an adaptive model for rational dosage prediction. 

Medicines play an important role in healthcare delivery, and when used properly, can help cure diseases, relieve symptoms, and alleviate patient suffering. 

Almost three quarters of CDS alerts were overridden in clinical practice, and 40% of the overrides were not appropriate, to address these issues, in this project, we propose to facilitate dosing recommendation directly from historical medication data instead of using medication information and rules. 

Technical Details of Final Deliverable

As mentioned, this project is artificial intellegence machine learning based. Hybrid Learning based model is trained via various algorithms so it requires high performance microprocessor. 

We are using microsoft azure cloud for the training and further processing of Datasets, and storing data of patients and their old prescriptions, for doctors. 

Compatible scanner is used for the scanning of old prescriptions of patients. 

Webcam is used for the clicking pictures, storing picutes as a patients record. 

Stationary includes printed pages, pens and notebooks for taking reviews about final product. 

Poster Printing which consists of Flow chart and basic explanation of project. 

Awareness session for "Irrational Use of Medicines", which help in understand and avoiding irrational use of medicines being misused, overused and underused, and how we can contribute towards sustainable development goals achievement by promoting good health and peoples well being. 

Final Deliverable of the Project Software SystemCore Industry MedicalOther Industries IT Core Technology Artificial Intelligence(AI)Other Technologies OthersSustainable Development Goals Good Health and Well-Being for PeopleRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	60800
Canon Lide 300 Scanner	Equipment	1	9800	9800
AI Processor	Equipment	1	20000	20000
Azure Datasets	Equipment	1	18000	18000
Poster Printing	Miscellaneous	2	1000	2000
Brostures	Miscellaneous	100	10	1000
Stationary	Miscellaneous	5	300	1500
Session about Irrational use of medicines	Miscellaneous	1	5000	5000
Logitech C270 HD Webcam,720p	Equipment	1	3500	3500