Signed in as:
filler@godaddy.com
Global challenges
Global health in the post-pandemic world
The pressures applied to health and healthcare during the COVID-19 pandemic exposed a world unprepared to respond to existential global health challenges. The universal need for robust digital solutions to bridge the clinical, epidemiological and public health gaps in managing infectious pandemics was apparent to all. The emergent onset also exposed global health systems lacking in fundamental resilience and most of the world’s population neglected or underserved regarding their fundamental healthcare needs. Patient centered care is about treating the person, not the disease. A patient should be involved in decision making about his or her own health. To accomplish this, it is important that understandable information is available, on both the disease as well as the treatment options. Social determinants of health, the sources of health disparities, chronic disease profiles and emerging comorbidities, need to be better understood. The underlying translational mechanisms must be retooled to respond to existing, future health challenges, and emergencies, and to direct adequate resources to improve global health. Enabling cooperation with data-connected innovation is key. AI can affect and add value in many areas of patient centered healthcare and has an important role in enabling such a future of precision and sustainability.
Clinical robustness
A study was conducted by Johns Hopkins University School of Medicine with other institutions in the US on the clinical robustness of digital health companies. The study scored their clinical robustness using data on clinical trials applications registered on ClinicalTrials.Gov and FDA filings[i]. Only 9% of the companies and start-ups that were sampled had a robustness score of 8 to 10 plus. An overwhelming 70% of the companies had 0 to 3 robustness scores. We must address the fundamental issue with the introduction into clinical practice of AI beyond the current research applications.
Translation & Scaling
The Council of Europe Steering Committee for Human Rights in Biomedicine & Health says that AI deployment in care remains nascent; clinical efficacy demonstrations are lacking in relation to research; and performance generalization from trials to clinical practice is still largely unproven.
Carbon footprint
The Stanford Institute for Human-centered Artificial Intelligence (AI) considers the carbon footprint of AI an emerging policy issue (AI Index Report 2022), with a disproportionate amount of computing power used to produce relatively limited results. Data efficiencies are too low to enable innovation scaling. This is known as the Pareto principle in AI and ML, where 80% of the effort and cost are spent on curating data rather than developing AI. This is particularly pronounced in the health space and essentially translates to a 20% efficiency for investments in AI projects and Machine Learning (ML), a critical component of digital health development for clinical applications; specifically for the discovery of knowledge in various data pools, to enable precision medicine and to reduce the carbon footprint of healthcare.
Translational medicine
The U.S. National Institutes of Health reports that 80-90% of research projects fail before they ever get tested in humans and for every drug that gains FDA approval, more than 1000 drugs developed fail. Almost 50% of all experimental drugs fail in Phase III trials. Hence, moving new drug candidates from preclinical research into human studies and the approval process occurs in only approximately 0.1% of drug candidates.
Social innovation & transformation
Scalable and sustainable digital health interventions and health technology innovations are far and few between, and despite the many published reviews examining evidence for effectiveness, cost-effectiveness, patient perceptions and effects on patient outcomes, several ongoing concerns remain. An often-overlooked aspect is the social value created by the technology innovation. Social innovation can be long-lasting changes through scaling and adoption, and include the organization and functions of health systems, governance transformations, innovation in care models and the re-organization of care processes. Social innovation impacts social values by growing and shaping a trusting and trusted relationship with the health system. The lack of longitudinal real-world evidence necessary to conduct resource/carbon- and time-efficient clinical trials hampers a robust learning healthcare systems.
LEARN MORE:
van Niekerk, L., Manderson, L. & Balabanova, D. The application of social innovation in healthcare: a scoping review. Infect Dis Poverty 10, 26 (2021).
Whyle E, Olivier J. Social values and health systems in health policy and systems research: a mixed-method systematic review and evidence map. Health Policy Plan. 2020 Jul 1;35(6):735-751.
Stern AD, Brönneke J, Debatin JF, et al. Advancing digital health applications: priorities for innovation in real-world evidence generation. Lancet Digit Health. 2022 Mar;4(3):e200-e206.
Rudrapatna VA, Butte AJ. Opportunities and challenges in using real-world data for health care. J Clin Invest. 2020 Feb 3;130(2):565-574.
Electronic Health Records
The cost and complexity of EHR development is prohibitive in LMICs and in many other countries, with the U.S.A. perhaps being the only exception of large-scale deployment and use, with many unintended consequences observed. Many problems associated remain with EHRs using AI for innovation evaluation and scaling. EHRs do however offer a substantial source of data that could be combined with telehealth and digital phenotyping applications and AI tools for data mining and public health research.
LEARN MORE:
Colicchio TK, Cimino JJ, Del Fiol G. Unintended Consequences of Nationwide Electronic Health Record Adoption: Challenges and Opportunities in the Post-Meaningful Use Era. J Med Internet Res. 2019 Jun 3;21(6):e13313.
Anatol-Fiete Näher, Carina N Vorisek, Sophie A I Klopfenstein, et al. Secondary data for global health digitalization. The Lancet Digital Health. February 2023.
Data ecosystems
The key challenges leading to health care innovation failure of the translational process from bench to bedside, relate to:
1. Poor quality of initial hypotheses to design the product before conducting tests, e.g. from previous reverse translations.
2. Lack of effective biomarkers.
3. The lack of integration of data inspite of various technological approaches;
4. Fragmented data coming from different fields – academic, research, clinicians, policy makers, bioinformatics and biostatistics.
5. Irreproducible and unreliable input data.
6. Lack of mechanisms to maintain validity and clinical coherence in data from clinical practice to research and public health uses, extending beyond matters of interoperability.
The same limitations impact the translational process in all health innovation, including the development of software as a medical device and clinical AI.
The limitations in the data ecosystem have a significant impact on demonstrating the cost-effectiveness of clinical AI, which further impacts the availability of trusted and accurate, structured data for clinical and policy decisions. The availability of real-world evidence undermines the assessment of the AI tools and of new models of care, including treatment approaches and drug development. This leads to an innovation predicament in health and health care, where the pace of technology-creates risks and complexities increases faster than the pace at which technology addresses these risks, making the current digital health innovation approach and affected health systems unsustainable. Risks and complexity include over or under-diagnosis, medicalisation of social issues, bias and exclusion, treatments with poor health outcomes, and healthcare provider burnout, to name a few.
LEARN MORE:
Seyhan, A.A., 2019. Lost in translation: the valley of death across preclinical and clinical divide–identification of problems and overcoming obstacles. Translational Medicine Communications, 4(1), pp.1-19.
Sabroe, I., Dockrell, D.H., Vogel, S.N., Renshaw, S.A., Whyte, M.K. and Dower, S.K., 2007. Identifying and hurdling obstacles to translational research. Nature Reviews Immunology, 7(1), pp.77-82.
Wagner, P.D. and Srivastava, S., 2012. New paradigms in translational science research in cancer biomarkers. Translational Research, 159(4), pp.343-353.
Sabroe, I., Dockrell, D.H., Vogel, S.N., Renshaw, S.A., Whyte, M.K. and Dower, S.K., 2007. Identifying and hurdling obstacles to translational research. Nature Reviews Immunology, 7(1), pp.77-82.
Ioannidis, J.P., 2016. Why most clinical research is not useful. PLoS medicine, 13(6), p.e1002049.
Ioannidis, J.P., 2005. Why most published research findings are false. PLoS medicine, 2(8), p.e124.