AI revolution in the field of healthcare
A doctor being counterquestioned by the patient armed with newly acquired knowledge regarding his symptoms through numerous online websites is already a frequent occurrence nowadays. Add to that the apps, which monitor your vital parameters at all times and the scope for further improvements; wherein, the test results can be fed to your app and the diagnosis indicated by the app would usually be pretty accurate. When Artificial Intelligence, Robotics and Big Data analysis is added to this, it just keeps getting bigger and better. Like most of the conventional industry, the healthcare industry is going to see a big disruption soon. Should the doctors be alarmed, may be yes, may be no; should they be excited Yes!
The advances made in the field suggest that AI should be embraced with open arms in the healthcare sector as it would augment the medical practitioners in not only penetrating these services to larger percentage of society, but also utilizing the time more effectively on each patient and arrive at a more accurate diagnosis. The data collected and learnings inferenced can further improve the medical as well as training processes.
The changes in disease profiles in our country and increase in chronic diseases demand (increase in diseases due to pollution is yet to show its real size) that state and central governments collaborate with researchers and healthcare giants to implement suitable programs that improve healthcare and AI can be a force multiplier in this regard. The effect of AI on major problems being encountered in India in healthcare sector will be discussed one by one:
1. Neglect of Rural Population. India faces acute lack of healthcare facilities in rural areas in terms of both quality and quantity. As per information in open domain, the rural population which comprises 75% of total populace have access to only 31.5% hospitals and 16% beds. Only 37% of people were able to access IP facilities within a 5 km distance, and 68% were able to access out-patient facilities due to limitation in number of doctors and lack of infrastructure. Unwillingness of skilled healthcare officials to work in rural areas only compounds the issue. Minimum tenures of officials have been made compulsory in some states; however, the same is grossly inadequate. According to the rural health statistics of the Government of India (2015), about 10.4% of the sanctioned posts of auxiliary nurse midwives are vacant, which rises to 40.7% of the posts of male health workers. 27% of doctor posts at PHCs were vacant, which is more than a quarter of the sanctioned posts. AI based solutions can use this limited infrastructure and manpower much more effectively. The present and predictive analysis of number of patients is one such tool through which the District Administration can keep a tab on actual situation on ground and can even predict the outbreak of an epidemic. The smart use of paramedicals may also be employed. Use of lightweight equipment for measuring ECG, blood pressure, heart rate, auscultation, oxygen saturation and the temperature of a patient which can transmit data wirelessly anywhere is already a reality and is already making big impact through low cost solutions to rural healthcare. ‘A3RMT’ is one such startup which assisted in treatment of more than 56,000 patients in 450 locations in India and saved more than 2,000 lives through emergency intervention.
2. Only Cure Based. Though the old saying preaches that ‘Prevention is better than Cure’, the present trend in India remains cure based. Though the government’s thrust on health with awareness of Yoga and general health is in the right direction, it’s too little too late. With continuous monitoring of health parameters on individuals using AI, the march towards preventive healthcare can become a reality. Also, self medication is very much prevalent in India. To equip personnel to verify the veracity of such remedies, a startup Myupchar’ has been created. Other such apps can help the users to avoid wrong or non-effective diagnosis, thereby not losing precious time which may prove crucial in certain critical ailments. Also, awareness about health can also be increased by using various tools and apps. Fitness bands have also increased the overall awareness.
3. Low Budget on Healthcare. The present allocation of about 1% of GDP to healthcare budget is way below what is needed. Though the government plans to increase this to 2.5% by 2025, it is still way below the required mark of 5 to 6%. The entry of AI into healthcare also means that the Government healthcare schemes would have much larger coverage with the same budget. The lower and effective medical help would mean lowering of costs of medical insurance thereby making it more affordable. At present 76% of Indians do not have any health insurance. This in turn should bring in more people to buy medical insurance and providing healthcare protection to larger sections of society. This would increase the overall load on the healthcare infrastructure of the country and thus cheaper medical education to churn out more doctors and nurses with effective use of AI and robotics would become a necessity. As the numbers increase the costs should slide down further.
4. Shortage of Healthcare Officials. A 2011 study estimated that India has roughly 20 health workers per 10,000 population, with allopathic doctors comprising 31% of the workforce, nurses and midwives 30%, pharmacists 11%, AYUSH practitioners 9%, and others 9%. The resource and knowledge sharing in between healthcare organisations can become online and instant depending on present usage of the resources. If the resource can be more effectively (or profitably) utilised by another organisation through virtual allocation by AI, the overall working efficiency would be higher. This is especially important in a country like ours where the infrastructure and number of healthcare specialists are far lesser than the need. Simple example of the same is sharing of ambulances.
5. Lack of Medical Research. The international pharmaceutical companies have for long neglected research in medicines for tropical diseases due to low profitability and the Indian companies have been limited in success in this area due to prohibitive costs of R&D. With the latest amendment in the rules for CSR by the government, the money can now be utilised for R&D and this would mean that the pharmaceutical giants would be pumping in large amounts of money into research in India. This would mean significant increase in the clinical trials in the country. The AI tools can make the clinical trials more effective and result oriented and would be able to indicate the side effects on the subject much earlier than otherwise.
6. Alternative forms of Medicine. AI tools can be utilised to monitor the results of medical practices like Allopathy, Ayurveda, Unani and Homeopathy. The data on successful diagnosis through these could be compared for various diseases and most effective form of medicine for each type of disease can be computed for further analysis.
7. High Cost of Healthcare. The high cost of healthcare is prohibitive for majority of Indian population and thus proper healthcare facilities remain out of their reach. The factors listed above will bring down the overall costs of healthcare.
8. Awareness. A study in urban Haryana found that only 11.3% of the adolescent girls studied knew correctly about key reproductive health issues. Since the same is a difficult topic to cover for parents as well as teachers, AI based training modules can be used to teach certain sensitive issues.
9. Accountability. With better data logging and AI based approach, the responsibility can be determined for various actions. Last year, a dengue patient was billed around Rs.16 lakh by Fortis Hospital, Gurugram for 15 days treatment in its ICU despite the death of the patient. Inspite of having the best of the doctors and latest technology, cases of medical negligence are reported frequently. In another case, an alive baby was declared dead by Max Hospital, Shalimar. With better use of AI tools, accountability in such cases can be fixed. However, for this, complete legal framework will have to be prepared to ensure that only those are punished who are proven guilty of intentional malpractice beyond doubt. Also, the data once stored should be protected to ensure that no tampering can be done even at back end. The privacy of data also needs to be taken into account.
10. Home Care. Avenues of patients being treated at home also need to be explored. This would ease the load on limited hospital infrastructure and reduce the chances of secondary infection to the patient. This also eases the life of attendants of patients, usually the close relatives. Start-ups like ‘PORTEA’ are working in this field and providing such avenues with mush lesser costs.
The doctors should also be excited as by intelligent use of AI, the major problems faced by them in India could be eased in a limited manner:
1. Understaffing. The doctors are overworked and often complain of fatigue. The problem is compounded due to poor infrastructure and support facilities. A human is after all a human and human fatigue of doctors in hospitals can be brought down with the use of AI tools effectively. The number of specialist doctors and nursing staff on duty can be brought down with AI machines acting as assistants to the duty staff. The AI machine with complete history of each patient and earlier diagnosis by the specialist can be of great help to the duty staff. This would also mean that time spent on each patient by the doctor reduces significantly thus giving doctors and staff to utilise this time effectively in upskilling themselves. The monitoring of patients with critical illness and assistance provided to them can be greatly enhanced. The use of AI tools in hospital administration would bring up the overall efficiency and hence reduce time spent on avoidable tasks.
2. Troublesome Patient and Relatives. With better sampling of data in case of each patient and data being available for analysis in future, the number of cases of malpractices would significantly come down. This would vastly improve the trust factor between the doctor and the patient, which would eventually bring down the cases in which the relatives manhandle the doctor and damage the hospital property. ‘Credihealth’ is one of the medical support startup that gives guidance to the patient regarding finding the right doctor to guiding them in admission to discharge process based on disease to be treated and even give cost estimates for every process. The policy and rules would have to be amended accordingly to cover the legal aspects of medical cases with the availability of AI and big data. All this would also require a
framework where the data is protected and reproduced as required without any chances of tampering at back end.
3. Expectations of Instant Relief. Use of AI tools could help convince the patients better regarding the diagnosis administered and time to recover. A predictive graph of such ailments and dosage would greatly dispense fears amongst patients and their relatives.
4. High Medicinal Costs and Charges by Hospitals. Doctors have often complained that whilst patients bear huge medicinal costs, the consultation costs in India remain very low. With the use of AI tools, the exact dosage for patients can be worked out, which will bring down the medicinal costs borne by the patients. Further weaning off of a patient from medicines could be predicted. Also, as the working of hospitals and organisations could be automated with use of AI, the total operating costs of hospitals could be significantly lowered. This would reduce the total cost burden on a patient and thus higher consultation costs of doctors could be facilitated. The option of choosing between a non-emergency plan and emergency plan of medical cover is also being provided through smart solutions. ‘Afforplan’ is a startup that gives plans where one could plan, save and pay for non-emergency planned procedures like pregnancy.
5. Non-compliance by Patients. In instances where the patients do not comply to the instructions of doctors, the blame still comes to the doctor. With better recording of parameters and use of AI tools, the monitoring of patients can improve significantly, thereby reducing the blame on doctors.
6. Prohibitively High Training Costs. The training costs for doctors especially for specialisation are prohibitively high. Presently average cost of completing MBBS from a private college is approximately 25 Lakh and for MD it is almost three times or more. These training costs can be brought down significantly with the use of AI-based training and college administration tools. The automotive and other industries have already started moving towards AI-based training making it more personalized and focussed.
The net throughput to every student in such cases is also higher resulting in better skill of every trainee. Reduction in training costs should also be utilised to educate the young trainees to the financial aspects of the healthcare (both the public and the private sectors) in low-income countries result in gaps between providers’ knowledge and the care provided, this is known as “knowdo gaps” and is a serious concern in delivery of proper healthcare services. Start-ups like ‘Timble Tech’ are working in the field of training using AI tools, the same company is presently also working with AIIMS on mapping the symptoms and response to certain kinds of diseases. a delivery system so as to bring in behavioral change towards cost-consciousness in the implementation of governmental schemes. Also, low knowledge by the providers
7. Poor Working Conditions. Majority of the government infrastructure, especially rural lack good infrastructure, proper management, dedicated staff and many other things which are required to provide reasonable and appropriate healthcare. The AI based tools can help this on multiple accounts. The tools can help in better reporting and monitoring of the issues. Also, the staff responsible for administration can be better trained with special emphasis on implications of poor hygiene and infrastructure and finally in better accountability. In 2017, 300 infants were reported to have died in Baba Raghav Das Medical College, Gorakhpur due to shortage of oxygen supply, indicating poor management. There is little evidence to indicate that the lessons learnt have been implemented or rather lessons have even been learnt.
The fear of AI taking away all the jobs should be dispensed with as the machines cannot think, not yet. All they can do is learn from previous results. Thus, large sample size is a pre-requisite for better estimate of probability and hence better inferences by AI operated machines; however, the results differ with different algorithm used on the same data sample. This difference in results based on different algorithms is called the algorithm bias. This issue is compounded by the problem that the healthcare data gathered through various tools may be inaccurate and the same is mostly not counterchecked. The AI operated machine once fed with such skewed interpretations can create havoc if left unsupervised. Thus, for a long time to come, these biases would ensure that the final decision making remains out of the control of machines. The healthcare industry thus needs to upskill themselves to be able to utilise the AI tools more effectively. Even with the use of AI, more health officials are required to be able to provide adequate medical coverage to the citizens of this country. The legal framework to cover aspects of AI based health coverage and training thus form the most important tasks ahead.
The decision making on use of AI in healthcare sector should be based on penetration of quality healthcare facilities rather than profitability. As the affluence in the populace increases, the profits would eventually rise too. Though AI is not a ‘one solution to all problems’ remedy, it can facilitate improving the conditions in the healthcare sector significantly. Each AI based solution will have to be built carefully and incrementally to ensure low investment. The study of functioning of each organisation could reveal the areas where the AI based tools should be employed the first; however, it is felt the hospital administration and training are the areas where its demand is going to be highest. Integrating of every individual’s data from smart watches/ fitness bands could come in next followed by intelligent resource sharing amongst various organisations.
Further advancements could be decided on the Return on Investments on the ones quoted above. There are numerous start-ups in India working in the field of AI being incubated under the ‘Make in India’ program and should be able to partly meet demands of healthcare sector to move towards AI. The Government and the healthcare sector need to factor the opportunity costs of loosing global business by not moving towards AI at the right time. The medical tourism in India is a significant contributor of revenue in this sector. Improved facilities and care could significantly increase the medical tourism in the coming years.
Composed by Gaurav Sharma
