Healthcare fueled by AI/ML/DL:
Intelligent and analytical technologies have the potential to rebuild many areas of care provision with computing and empathy. The rate of change will differ tremendously upon the job to be done. While far-future is profoundly different, but pontificating on the short-term is fraught.
AI/ML/DL can make the biggest impact in clinical operations and better-utilizing staff and resources i.e. scheduling, billing, rev cycle management. AI/ML/DL-based systems are changing how health can be managed since it can flag things like which patients need treatments when, and who needs more at-home visits. The concept of the ‘An I doctor’ is over-hyped. HC is not replacing physicians anytime soon. Instead, AI will be assisting and improving the efficiency of providers. Replacing a holistic diagnosis, treatment, and management care pathway with machines is far out over the next decade or so.
To offer an exhaustive list, the following are areas of HC where AI/ML/DL can add value,
1) Real-time case prioritization and triage
2) Assisted or automated diagnosis and prescription
3) Personalized medications and care
4) Patient data analytics
5) Care coordination
1) Diagnosis, treatment, and monitoring of health
2) Diagnosis error prevention
3) Early diagnosis
4) Medical imaging insights
5 Robot-assisted surgery
6) Virtual nursing assistants
Research and development
1) Drug discovery
2) Gene/DNA analytics and possible editing
3) Device and drug comparative effectiveness
1) Health benefits administration
2) Data infrastructure and interoperability
3) Customer acquisition and relationship management
4) Administrative workflows
5) Market research and pricing
Considerations and Issues to address:
As the AI market continues to evolve and new best practices are established, there are challenges and unique considerations for the successful technology adoption. Providers must consider how patient privacy and security will be protected and how to:
i) Effectively process and take advantage of unstructured data
ii) Deal with limited access to high-quality and unbiased data sets
iii) Utilize high-performing and reliable network capabilities
iv) Implement data governance strategies
v) Tackle a lack of talent and develop and adopt new staffing and training strategies
vi) Find a balance between costs and potential benefits
AI Use Cases in Healthcare:
Let us look into AI live use cases across HC value chain covering patient care, medical imaging, and diagnostics, R&D and Health Management
1) Patient care:
- GNS Healthcare: The company uses machine learning to match patients with treatments that prove the most effective for them.
- Oncora Medicals: The software structure, analyze and learn from the data that health systems have to enable them to provide personalized treatment.
- Wellframe: Wellframe flips the script by delivering interactive care programs directly to patients on a mobile device. Its portfolio of clinical modules, developed based on evidence-based care, enables Care Team to provide a personalized experience for any patient.
- Enlitic: Patient triaging solutions scan the incoming cases for multiple clinical findings, determine their priority, and route them to the most appropriate doctor in the network.
- Jvion: The Cognitive Clinical Success Machine precisely and comprehensively foresees risk deliver the recommended actions that improve outcomes.
- Zakipoint Health: The company displays all the relevant healthcare data at a member level on a dashboard to understand risk and cost, provide tailored programs and improve patient engagement
2) Medical imaging and diagnostics:
- SkinVision: SkinVision enables you to find skin cancer early by taking photos of your skin with your phone and get to a doctor at the right time.
- Amara Health Analytics: Amara provides real-time predictive analytics to support clinicians in the early detection of critical disease states.
- Desktop Genetics: Desktop Genetics is an international biotechnology company to help researchers discover and treat the root genetic causes of human disease.
- 4Quant: The company utilizes the latest Big Data and Deep Learning technology to extract meaningful, actionable information from images and videos for experiment design to help pick and choose which components make the most sense for the needs.
- NuMedii: Biopharma company, NuMedii has built a technology, AIDD (Artificial Intelligence for Drug Discovery) that harnesses Big Data and AI to rapidly discover connections between drugs and diseases at a systems level.
4) Health Management:
- MD Analytics: MD analytics is a global provider of health and pharmaceutical marketing research solutions.
- Healint: Company’s product Migraine Buddy has recorded terabytes of data that helps patients, doctors and researchers better understand the real-world causes and effects of neurological disorders.
Roadmap to Implement AI in HC
- AI/ML/DL has the potential to overhaul the processes of researching, purchasing, and implemented IT tools in the healthcare industry. What could be a successful path to implement AI/ML is a puzzle to solve for many HC companies in terms of taking a big-bang approach or build incrementally or taking a path to passively infuse AI in organizations. The following are key components in developing a robust strategy avoiding the pitfalls.
- The first step is identifying the right use case and piloting which demonstrate the right value/improvement
- Realistically assessing where the HC organization currently lies on the maturity curve follows next
- Selecting the right vendors in building the right ecosystem in building and expanding AI capabilities
- Defining the right metrics to gauge the success or failure helps in establishing checks and controls
- A typical deployment involves constantly optimizing the algorithms and iterating to evolve the right solutions
Commoditizing AI/ML in Healthcare:
Let us examine the extent to which AI/ML in healthcare is being commoditized. The combination of EHR ubiquity, increased computational power, the open source movement, and the rise of cloud providers has made training AI/ML models easier. But how can vendors developing AI/ML models and delivering to a client ensures a complete package to make machine learning work toward its intended purpose? So what it takes to deliver accurate models that are easy to implement, targeted to a specific use case, connected to actionable interventions, and surrounded by a machine learning community and support team with extensive, exclusive healthcare experience. Effective machine learning is the product of more than data, features, and algorithms, and is defined by five key differentiators to expedite mass adoption and commoditizing the technology:
- Vendor’s expertise and exclusive focus on healthcare.
- AI/Ml model’s access to extensive data sources.
- AI/Ml model’s ease of implementation.
- AI/Ml model’s interpretability and buy-in.
- AI/Ml model’s conformance with privacy standards.
As health systems and payers continue searching for every opportunity to improve clinical and financial efficiencies to help them deliver better care at a lower cost, AI/ML offers right solutions that hold significant promise for fulfilling HC goals.
AI/ML Impacts on HC. AI and machine learning are already delivering value in HC. The following are high impact areas.
- Specialty care including radiology, pathology, and pharma. For patients, telehealth and remote patient monitoring are two dominant areas.
- Chronic health conditions are expected to benefit most from AI/ML. Cancer, heart disease and diabetes are also commonly seen as big opportunities for overarching healthcare trends such as pop health and precision medicine, so it makes sense that amid the promise of AI and ML is hope the tools can help with those as well
- Lastly, the business-side uses such as resolving operational inefficiencies and optimizing administrative workflows, which are becoming important to HC executives even if they are perhaps less glamorous than patient-facing applications.