ByDarko Savic
AI-powered "what you should know" distractions
3 years ago
6
Create newCreate new
Idea
Contrived _voice Apr 23, 2022
Using an AI to assign qualitative and quantitative values to a patient's health and offer assistance to the most critical first.
Image credit: Photo by LinkedIn Sales Navigator: https://www.pexels.com/photo/two-men-standing-inside-a-room-2182976/
Please leave the feedback on this idea
Originality
Is it original or innovative?
Feasibility
Is it feasible?
Necessity
Is it targeting an unsolved problem?
Conciseness
Is it concisely described?
Bounty for the best solution
Provide a bounty for the best solution
Bounties attract serious brainpower to the challenge.
Create a national database of patients connected to an AI. The AI then uses data from yearly checkups, Family history, in-patient data and personal medical history to extrapolate on when the next medical visit should be. The AI also acts as a personal health consultant when you feel under the weather. It would also void the need for an in-person consultation and consequently make for shorter waiting lines at hospitals. It would schedule appointments based on the level of need a patient has, helping save lives and relieve doctors of decision fatigue.
Why?
- Doctors are only human, Misdiagnosis happens all the time. The US alone has an estimated 12 million misdiagnoses per year . Anything that can be done to lower that number should be done.
- This system allows both private and public facilities to pool their data for the well being of their patients.
- There are too many online services that offer medical consultations. I hypothesise that they would have a higher probability of misdiagnosis due to a lack of information on what is going on around individual patients. Hospitals can work around this because most patients come from the same area and by talking to alot of them they can recreate the full picture relatively quicker.
- With the emergence of global pandemics, big data might offer us a shot at fighting back and protect the ones the most at risk. Being able to schedule and reschedule health drives to combat the virality of disease might reduce the total loss of life.
- Could help identify genetic based ailments and direct younger members to make life choices that reduce the likelyhood of the condition manifesting.
- The global reach means nations can't underplay diseases and everyone is well informed on how at risk they are of being affected.
- Could allow labs to run as hospitals to offset thier expenses without having to need as much invest in relevant facilities by only dealing with cases they are equiped to handle.
- If multiple nations would be willing to place all their databases and practices available for annonymous cross-reference by an AI with the explicit intention of making healthcare available, It could make getting specialized procedures abroad so much easier and so much cheaper.
- Could allow clinics and small practices to relieve strain from major practices if they have the relevant facilities to do so.
How it works
The government creates a national database and all hospitals have to upload all relevant data they have on their patients. The data is private, only accesible to you and your doctor. The patient feeds their present ailments to the database and the AI annonymously cross-refrences them against known family members, coworkers, friends and neighbors to find probable causes of the ailment. Once that is done, It cross-references the results against a database of doctors and hospitals with a history of success in dealing with your present condition. The AI then shows you their fees and the soonest you can schedule an appointment.
In cases where there isn't enough information,the AI can help schedule appointments for bloodwork, X-rays or MRI at the nearest and fastest place and ussing the results direct you to a doctor or hospitals to get treatment
Statistically, at any given moment there should be an emergency case requiring attention. Major hospitals have an emergency ward for this exact purpose. I propose a similar arrangement here where a few well equiped practices refuse to take appointments and instead recieve immediate patients as they are directed. coordination is key here and by keeping the flow of patients distributed among as many hospitals as necessary it could be possible to significantly reduce the strain on existing facilities.
The national database could also link to relevant insuarance providers or double as a medical insuarance provider. This way by principle of pooling and diversifying it could make it possible to reduce the cost of healthcare for common citizens.
In the event of a pandemic or local outbreak, the database keeps track of the spread, calculates your individual probability of infection and gives you a percentage value as well as preventive measures you can follow to protect yourself.
Working case
artificial intelligence (AI)-assisted module XIAO YI
Using a retrospective cohort study from August 1, 2019 to January 31, 2020, they compared the waiting times of patients of who used the AI first before an in person consultation and those who went to the hospital first. The results speak for themselves.
The median waiting time was 0.38 (interquartile range: 0.20, 1.33) hours for the AI-assisted group compared with 1.97 (0.76, 3.48) hours for the conventional group
You would think that it would increase costs but the data showed the total cost for the AI assisted group to be 335.97 CNY compared with the 364.58 of the control group. These numbers aresignifant since the test population was only 12,342 . Imagine the results when implimented at larger scales.
Note
Privacy should be a key principle when making this system. It becomes impossible to trust such a system when it starts being used as an ad platform to sell drugs or special services. The end goal shouldn't be advancement of capitalistic goals but the use of existing technologies to improve human life and eliminate strife and suffering.
[1]https://www.docpanel.com/blog/post/how-common-misdiagnosis-infographic
[2]https://bmchealthservres.biomedcentral.com/articles/10.1186/s12913-021-06248-z
[3]https://bmchealthservres.biomedcentral.com/articles/10.1186/s12913-021-06248-z/tables/1
2
Creative contributions
Cold storage of private data and cloud-based computation
Software that analyses uploaded medical history data of an individual, correlates and analyses data based on the previous cases, patients, research patients, etc., and gives you medical preventive or treatment suggestions.
How would it work?
- User imports data (full control of the analyzed data; could choose to analyze only a dentist history, for example)
- Software analyses data locally and translates it into AI-calculated coefficients that define the patient's state, change during time, and described the drug-symptoms relations (units of data). These locally calculated coefficients (units) are sent to the cloud where they get compared to the other units, analyzed and conclusions are drawn. Feedback data is sent back to the local computer and suggestions are given to the patient. --> no privacy issues since data never leave local storage
Additional information
- The coefficients (or units) could be protected with end-to-end encryption to double-up the protection.
- The algorithm should suggest the patients to upload more data if it's not sure or some data is missing (based on the other people's data or "minimal data for conclusion" guidelines).
- The software suggestions should be available to the doctor, which would approve the treatment when real documentation is presented.
- If the user wants, it could choose to sell some of its data to pharmaceutical companies. If not, the data would forever stay locally stored.
Benefits
People would get a chance to play a role in medical diagnosis and therapy decision making.
Please leave the feedback on this idea
The problem halting this is free healthcare
Summary: If you exclude advertisements (selling individual's health data), the service should cost a lot in order to cover all the expenses. Not a viable product in free healthcare ecosystems.
The idea is great and seems like the future of medicine. However, a few years ago, I read about Medicus.ai and Sema4, companies that intended to do the same thing by implementing AI-driven technologies to existing systems. Although I think this is great, I am not sure that your proposed way is the best way to do it. I am keener on believing that private medical labs and clinics will be much better at doing this. Why?
Individual approach to every patient is the advantage and benefit of both, the patients and the company. Patient health data are among the most profitable markets worldwide. Imagine your browsing history and all the sneakers you googled and the money people earn by selling them to the advertising companies. Now imagine a company having data on your health and providing this information to pharma companies that can control and sell you whatever they want, and you need it so badly! They would pay a lot for this data, so it's a great asset you would build by implementing this model. However, it would significantly increase the costs of the medical service since all this data needs to be stored somewhere and processed by machines doing heavy-duty work. Not to mention breach protection, cooling, and maintenance of the software and databases. Not everybody is ready for this. Some countries, like mine, have "free" healthcare for everybody, meaning that you pay a monthly fee and can use basic healthcare services if needed. Implementing your model would significantly raise the prices, since a huge change of system should be done. That's why private labs, where people already pay a lot for medical services, will probably make it work sooner.
Please leave the feedback on this idea
Contrived _voice2 years ago
you're right there. but if private clinics and labs do it first then the goal will always be monetary gain first and then healthcare next. It's a grounded approach but it doesn't make healthcare accessible and affordable for everyone.
You could build such a program for the express purpose of harvesting patient data and selling it to pharma companies and it could make medicine and healthcare more accessible to some extent but I don't see people willingly giving up their medical history for it
Please leave the feedback on this idea
jnikola2 years ago
Contrived _voice Yes. What I wanted to point out is that the future of medicine where help will be given in the right moments by AI-tracking and prediction is probably going to come from the private sector first. But to make it accessible, we need to apply the same principle to all the hospitals. So what we need to think of is a way how AI can be introduced on the level of each individual's data, without big privacy issues.
Please leave the feedback on this idea
Contrived _voice2 years ago
J. Nikola, and that right there is the catch. If you could find a way to reassure people that their privacy is secure, a healthcare AI could be implemented for both public and private practices
Please leave the feedback on this idea
Add your creative contribution
General comments
jnikola2 years ago
https://aiforgood.itu.int/meet-your-virtual-avatar-the-future-of-personalized-healthcare/
Please leave the feedback on this idea
ByDarko Savic
AI-generated articles comprised only of quotations from scientific papers and based on what you want to convey
3 years ago
1
ByDarjan Hren
Influence emotions through visual meditation
3 years ago
1
BySubash Chapagain
A distributed network of sensors to warn people about allergens
2 years ago
3
ByDarko Savic
Hand disinfecting pocket inserts for kids
3 years ago
2
ByDarko Savic
Using big data to match compatible people and form great friendships
2 years ago
ByDaami A.
What If You Had Your Own Personal Stylist?
3 years ago
6
ByDarko Savic
Personal omnipresent content recommendation algorithm
3 years ago
3
Please leave the feedback on this idea