The healthcare trade has lengthy been a hotbed for AI innovation, pushing the envelope on prospects. The trade has lengthy suffered a expertise crunch, an issue that AI might assist resolve. Nonetheless, utilizing synthetic intelligence and machine studying on this subject comes with a bunch of challenges.
First, the medical information of customers should be preserved at any value. Medical information is among the most delicate types of personally identifiable data and requires the very best safety and confidentiality doable. Second, AI options can’t be allowed to make choices on their very own with regards to the standing of sufferers.
However these roadblocks, automation by way of AI can vastly enhance the standard of life for medical professionals. Analytics India Journal reached out to Aneesh Nair, CIO & co-founder of MyHealthcare, a speciality well being tech firm launched in 2017, to study extra about how AI is being deployed within the medical subject.
AIM: Medical information is among the most vital areas to implement information privateness and safety. What measures does MyHealthcare take to guard customers’ information?
Nair: As a cloud-native, and mobile-first healthcare utility supplier, we leverage new-age cloud modules to ship the gold normal in safety for our clients. To guard affected person information, we use Information-at-Relaxation and Information-in-Movement Encryption, which is among the simplest methods to guard it from unauthorised entry. This makes it way more troublesome for attackers to entry or steal the info.
Our ecosystem has additionally applied strict entry management measures as one other vital step in defending affected person information. That is carried out utilizing a wide range of strategies, resembling isolation, authentication and authorization, to make sure that solely approved customers can entry the info.
Conducting common safety audits is one other vital measure that we use to make sure the privateness and safety of affected person information. We additionally make sure that all software program and safety patches are put in promptly to guard towards the most recent threats. By making a number of copies of affected person information and storing them in several areas, we guarantee restoration of the info in case it will get misplaced or corrupted. We additionally present common coaching to our consumer’s workers on safety and encourage them to report any suspicious exercise.
AIM: What are a few of the greatest optimisations you will have seen deploying AI purposes in your utility?
Nair: MyHealthcare makes use of AI to assist automate many routine duties resembling information entry and evaluation. This could save time and assets, permitting healthcare suppliers to deal with extra vital duties and ship care extra effectively. AI helps us analyze a affected person’s medical historical past and different information to create personalised remedy plans. This may assist docs tailor remedy protocols to the precise wants of every affected person, main to higher outcomes and a extra satisfying expertise for the affected person.
We will additionally analyze giant quantities of knowledge with the assistance of AI to establish traits and make predictions about future well being outcomes. This may help our healthcare suppliers establish potential issues early and take steps to stop them, bettering general well being and lowering prices. Our ecosystem additionally makes use of AI to research medical pictures and different information to establish patterns and make extra correct prognosis.
By implementing Speech to Textual content characteristic in our IPD & OPD EMR purposes, we’re lowering the docs’ burden by offering them the speech choice to enter all of the affected person particulars like chief complaints, historical past of current sickness, evaluate of methods, present drugs and allergy symptoms. Equally, to higher serve the sufferers, we now have an actual time analytics engine applied to enhance operational effectivity throughout our platforms. Additional to get the 360-degree view of sufferers, we current the affected person information available for the docs by way of a affected person journey mapping.
The system mentions the preliminary signs, assessments suggested, prognosis carried out, drugs and observe up particulars, all of that are made out there on a single display for the docs. Total, the advantages of utilizing AI in our purposes are quite a few and have a constructive affect on affected person care and outcomes.
AIM: Are you able to delve deeper into the specifics of how you will have deployed AI in your enterprise ecosystem?
Nair: We now have efficiently deployed AI-assisted tuberculosis detection modules for the Radiology division, the place a pc vision-based utility educated on 1000’s of X-ray studies, works on affected person x-rays and identifies any abnormalities within the lungs. The algorithm additionally highlights them by way of warmth maps and gives the likelihood rating for the docs to research additional.
Such methods establish the lung nodules that even an skilled radiologist would possibly miss together with his bare eye. Equally, the system is ready to precisely establish abnormalities in several areas and even behind the bones. This helps docs make higher choices in regards to the remedy, resulting in improved affected person outcomes.
Along with TB, we now have additionally developed an utility to detect Parkinson’s illness, a neurodegenerative illness, which causes sufferers to undergo from motion issues, making it troublesome for them to handle their frequent hospital visits.
Our cellular app makes it doable for sufferers sitting within the consolation of their very own properties to ship their strolling movies to the AI engine. Right here, we apply human pose estimation strategies to extract the physique coordinates and compute the variability utilizing a educated deep studying mannequin. This then gives the likelihood rating for the docs to resolve on the subsequent plan of action.
AI is used extensively to research digital well being data to establish traits and make predictions about future well being outcomes. This helps docs get a longitudinal view of affected person data and establish potential issues early and take pre-emptive steps to stop them. We additionally use NLP algorithms in our purposes to research affected person notes and different textual content information to extract data and achieve insights. Additional, on the radiology aspect to automate the report era, we now have deployed our specifically educated radiology speech engine.
This resolution understands the complicated radiology terminology dictated by the radiologist and precisely converts the speech into textual content, thus bettering the effectivity of the radiologist. At the moment, we’re engaged on constructing a complete scientific choice assist system which gives references and recommendations to docs for the given signs.
AIM: There are additionally many dangers arising from utilizing AI in healthcare. Have you ever confronted any?
Nair: Whereas utilizing synthetic intelligence in our purposes brings many advantages, we’re conscious that there are additionally potential dangers related to unsupervised AI. The best way we use AI (Augmented Intelligence) is to reinforce the capabilities of docs and healthcare professionals and by no means use it as a instrument for changing docs.
We at all times guarantee there’s a human in loop for all decision-making processes the place AI is concerned. Additional by design, we make sure the ethics and equity consideration are a part of our machine studying mannequin, in order that we keep away from any potential errors and biases in direction of the decision-making course of. By fastidiously managing the potential dangers and taking steps to mitigate them, we make sure that AI is deployed by way of our purposes methods and is used safely and successfully.
