141 lines
4.1 KiB
Markdown
141 lines
4.1 KiB
Markdown
# Smart healthcare
|
|
|
|
<!--toc:start-->
|
|
|
|
- [Smart healthcare](#smart-healthcare)
|
|
- [Intro](#intro)
|
|
- [Reasons](#reasons)
|
|
- [Implementation](#implementation)
|
|
- [Research: data driven health accessment](#research-data-driven-health-accessment)
|
|
- [Background](#background)
|
|
- [Data](#data)
|
|
- [Wearables](#wearables)
|
|
- [Patient Health Records](#patient-health-records)
|
|
- [Structured vs. unstructured](#structured-vs-unstructured)
|
|
- [Smart digital health platform](#smart-digital-health-platform)
|
|
- [Diagram](#diagram)
|
|
- [Use Cases](#use-cases)
|
|
- [Conclusion](#conclusion)
|
|
|
|
<!--toc:end-->
|
|
|
|
## Intro
|
|
|
|
- History: 
|
|
- Patient centric
|
|
- Precision medicine
|
|
- Atrial digital twins
|
|
|
|
## Reasons
|
|
|
|
- Current problems:
|
|
- Global Health Disparities
|
|
- Unequal shortage of healthcare workers
|
|
- Increasing rate of burnout in healthcare
|
|
- Mental health
|
|
- Sustainable Development Goal 3
|
|
- What smart healthcare can help
|
|
- Remote doctor consult
|
|
- Reduce un-necessary hospital visits
|
|
- Democratize healthcare access
|
|
- Remote emergency
|
|
- 5G powered emergency channel give doctor real time view
|
|
- Real time access of vital signs like ECG, ultrasound image, and blood
|
|
pressure
|
|
- Access to medical history
|
|
- Doctors guide paramedics in ambulance
|
|
- Drug discovery
|
|
- Cost saving and agility methods
|
|
- All fancy words including (ML, DL, ANN)
|
|
- Offset a shortage of specialists
|
|
- AI can help in aiding diagnosis
|
|
- Before smart hospitals
|
|
- Data presented is overwhelming
|
|
- hindering the ability of using it effectively
|
|
- People trip and die in hospitals
|
|
- Too much administration work
|
|
- Shortage of qualified staff
|
|
- Spending wasted in managing assets
|
|
- Nurses waste time in searching for medical equipment
|
|
- Benefits of smart hospitals
|
|
- Save money on asset searching
|
|
- Reminder
|
|
- Home-based health monitoring reduces re-admission
|
|
- Eliminate emergency
|
|
- Warning of deterioration
|
|
|
|
## Implementation
|
|
|
|
- EMR: Electronic medical records
|
|
- Diagnosis, Image Data, Unstructured text data, Medications, Procedures,
|
|
Lab tests
|
|
- Data is challenging to use: high dimensionality, sparse, noisy, irregular,
|
|
biased
|
|
- IoMT: lots of data, lots of money going in
|
|
- Healthcare technology: Better quality of life, with less money required
|
|
- 
|
|
|
|
## Research: data driven health accessment
|
|
|
|
### Background
|
|
|
|
- Role of AI
|
|
- Detection of aliment
|
|
- Improve decision making
|
|
- Help in treatment
|
|
- Superior experience
|
|
- Tech revolution
|
|
- Automated steps in diagnostic process
|
|
- Can acquire different types of data, representing the physiology of humans
|
|
- Images: MRI, CT
|
|
- Wearable devices
|
|
- Electronic health records
|
|
- Robotics
|
|
- ML learn from data, so that it can make decisions like human brain
|
|
|
|
### Data
|
|
|
|
#### Wearables
|
|
|
|
- Versatile monitoring
|
|
|
|
#### Patient Health Records
|
|
|
|
- Personal demographics: age, gender, ethnicity, weight, height, BMI
|
|
- Medical parameters: blood, pressure, fat, blood sugar, plasma
|
|
- Non-medical parameters: physical activity, diet plan
|
|
- Questionnaires filled
|
|
- Text mining can be performed
|
|
|
|
#### Structured vs. unstructured
|
|
|
|
- Structured: Time, date and code in diagnosis logs, can be stored in RDBMS
|
|
- Semi-structured: medication taken, procedures, allergic reactions
|
|
- Unstructured: Clinical notes, need NLP tools
|
|
|
|
### Smart digital health platform
|
|
|
|
#### Diagram
|
|
|
|
|
|
|
|
#### Use Cases
|
|
|
|
- Heart disease detection
|
|
- Computer aided retinal disease diagnosis
|
|
|
|
### Conclusion
|
|
|
|
- Future of healthcare technology involves:
|
|
- Self monitoring
|
|
- AI for health management
|
|
- Research is being developed to integrate real-time data with technology
|
|
- Require future clinical trails
|
|
- Explainable artificial intelligence:
|
|
- Need model transparency for clinicians and stakeholders
|
|
- Need collaboration in all fields:
|
|
- Academics
|
|
- Clinicians
|
|
- Industries
|
|
- Stakeholders
|