Artificial Intelligence (AI) is a fascinating powerful tool, it is continuously gaining limelight since its introduction. In this fast paced environment, AI feels like dominating every sector because of its usefulness. It is the beginning and poses challenges for researches in healthcare sector as AI’s ability is yet to be fully explored, learned and used for the greater benefit. In order to utilize AI, the present time demands a new Health Technology Assessment (HTA) framework for evaluations of AI enhanced healthcare. Any new tool/ technology comes with advantages and disadvantages, which are to be carefully studied and some of which are revealed slowly as it is incorporated in to the system. The fear of unknown is obvious but AI can never replace human as it will always need the guidance and human intelligence.
Multi-disciplinary collaboration will go a long way to seize AI’s benefits as HTA researchers may need to upgrade their skills and work together with computer scientists. Traditional HTA methods were fixed but AI’s algorithms and software may need a constant update. In anticipation, by the time HTA policies are released, AI system might be ready for a new update. Additional challenges are expected while conducting economic evaluations for AI systems.
AI enhanced device and procedure involve capital, recurring and indirect cost along with long term savings by early diagnosis, preventive care, reduced unnecessary investigations/ tests and hospitalization, fast recovery, reduce travel cost, efficient workflow. Researchers will need to carry out comprehensive costing analysis to determine the affordability and sustainability.
AI systems can improve healthcare as it is efficient in earlier detection of diseases, reducing medical errors and clinician burden, enabling smooth and fast workflow. One needs to also consider that due to its sensitivity, initially hospital may experience more number of patients who earlier were not getting diagnosed because of missed diagnosis/ human error and this will bear more cost to the health system. If AI assisted decisions go wrong due to algorithm or software error or bugs, who will take the responsibility and bear the repercussion is another concern. Continuous real world evidence generation might be necessary to support software updating of AI system. HTA researches will need to understand whether the decision making process by AI systems can be trusted, validated and generalizable to what extent. Overall consideration, listing and capturing all cost and benefit will be needed to determine value for money (ICER-Incremental Cost effectiveness ratio).
Patient satisfaction and trust are essential to ensure AI is acceptable. By treatment precision, timely diagnosis, remote disease management and monitoring, AI systems suggest improvement in quality of life and more independence specially for sick, old or disabled people. However, patient data privacy, limited digital knowledge, reduced physical empathy can affect accessibility. Overall benefit depends on how one works on finding ways and solutions for challenges. Equity is another consideration as AI can reduce disparities but can also become a barrier. Internet connectivity, digital literacy, infrastructure may be a problem in rural sector. Barriers need to be carefully considered to ensure fairness for disadvantaged population. AI is not be feared but to be used as a tool to enhance health care.
