Growing Impact of Generative AI on Healthcare – Asrar Qureshi’s Blog Post #902

Growing Impact of Generative AI on Healthcare – Asrar Qureshi’s Blog Post #902

Dear Colleagues!  This is Asrar Qureshi’s Blog Post #902 for Pharma Veterans. Pharma Veterans  aims to share knowledge and wisdom from Veterans for the benefit of Community at large. Pharma Veterans Blog is published by Asrar Qureshi on  WordPress, the top blog site. Please email to asrar@asrarqureshi.com for publishing your contributions here.

Credit: This is Engineering

Credit: Michelangelo Buonarotti

Credit: Google Deepmind

It is true that we are seriously behind in development and the gap between our country and the developed world is widening with passing time. However, in the digital space, we have generally been able to stay close to major developments. The irony is that the ills of technology reach us immediately, while the benefits take much longer and may or may not arrive fully.

An important factor is that we prefer to use technology which comes free of cost, because we hate to pay for things we like to use. As far as we can exploit free goods, we do it, but when it comes to paid ones, we back off. The biggest example is Windows operating system. The entire corporates are using unlicensed versions of OS and MS Office apps. I understand that all technology developers and sellers do one or both the things: they keep a basic, free version so that more and more people may get enticed and buy paid version; and/or look the other way when pirated versions are used even on mass scale because it adds to their reach and popularity. If people in our countries and so many other countries like ours could not use pirated Windows OS, it would never become a household name. It is not difficult for Microsoft to block pirated versions, but they look the other way.

Artificial Intelligence – AI – has literally catapulted the whole world. Generative Pre-Trained – GPT – came as ChatGPT from open AI. It was free in the beginning for a short time and then became paid. Later, Google came up with Bard and Microsoft introduced Copilot. These are three generative models developed by three major companies. The AI- based apps are already running in hundreds, and more are coming every day.

The onslaught of generative AI is so fast and powerful that it is becoming difficult to keep pace. There is no industry or area of work or life which is not getting impacted by AI. A negative impact is that job losses have already started happening. Google has announced cutting down 12,000 jobs due to AI. In Japan, 60% jobs are likely to change or shifted to AI; 46% jobs in USA are predicted to change entirely. If people wish to keep working, they will have to learn new skills – reskilling – or enhance their existing skills – upskilling. 

AI is a much faster change, if not bigger also, than the introduction of computers, which took many years to get into general use. The development took years from Disk Operating System – DOS – to Windows, and then to its multiple editions. Today, the softwares are developed very quickly and are updated at an even greater speed.

Boston Consulting Group is a global consulting firm that partners with leaders in business and society to tackle their most important challenges and capture their greatest opportunities. BCG was the pioneer in business strategy when it was founded in 1963. Presently, BCG has 30,000 employees, over 100 offices in over 50 countries, $11.7 billion in annual revenue, and works with most of the largest corporation across all major industries.

Few months ago, they had predicted that the generative AI will transform healthcare sooner than we may think; this is already being proven right. True, that their focus is on the US and European markets, but the whole world shall be impacted. We must also remember that a large number of Pakistanis live in the US, Europe, Australia etc. and they individually may bring some technologies to their own people here while our systems are not upgraded yet.

I shall discuss some key points and relate these to our market and situation.

There is no doubt that Generative AI holds the potential to change things dramatically, increase efficiency, improve quality of care, and create value for healthcare organizations. There are risks and uncertainties also. ChatGPT, Bard, Copilot put a disclaimer that AI can give false results, and it does, except for some straightforward queries. It will improve over time as it learns more, but relying entirely on AI for all cases is not advisable. The greatest value of AI is in data extraction and organization, and for repetitive tasks where human interaction may not be necessary. BCG recommends that leaders must plot a path to capitalize on the technology – starting today. I am taking major input from their report for the information shared below.

Healthcare Providers

Several generative AI providers are developing solutions—from diagnosis to care provision to patient monitoring—to help providers improve clinical outcomes. Others are working to improve resource utilization by both clinical and administrative staff.

Paige.AI, a digital pathology company, is integrating generative AI into its products to improve the accuracy and efficiency of prostate cancer detection. It was the first company to receive FDA approval for AI use in digital pathology and is looking to integrate the resulting information into patient electronic health records along with other clinical data.

On the administrative front, Doximity, Abridge, and DeepScribe are exploring applications that automate processes such as documentation, claims handling, preauthorization and appeals, patient onboarding, and scheduling. DeepScribe, which offers AI scribing services, has been able to decrease the amount of time providers spend on administrative tasks by three hours each day, and Abridge’s ambient AI scribing products are now in use at more than 140 provider locations in the University of Kansas Health System.

Pharmaceutical Companies

Generative AI is accelerating drug discovery, improving clinical-trial planning and execution, and leading to more precision medicine therapies.

Generative AI allowed Insilico Medicine to go from novel-target discovery to preclinical candidate in just 18 months, spending only $2.6 million. The company’s idiopathic pulmonary fibrosis drug recently received the agency’s Orphan Drug Designation after completing the preclinical phase in 30 months, much faster than average for a new treatment.

The biotech company Exscientia is using generative AI to analyze patient tissue and employ functional precision oncology to improve patient outcomes. NVIDIA is offering a set of generative AI cloud services that enable customization of AI foundation models to accelerate drug discovery and research in genomics, chemistry, biology, and molecular dynamics. The services provide pretrained models and enable researchers to fine-tune generative AI applications on their own proprietary data. The offering has been adopted by drug discovery startups such as Evozyne and Insilico Medicine, as well as by incumbents such as Amgen.

Medical Technology

Generative AI could help companies create more personalized and patient-centered devices—incorporating software that allows for preventive maintenance and repairs, for example.

The UK’s National Centre for Additive Manufacturing is applying generative AI to optimize the design of medical devices such as prosthetics and implants, tailoring them to the needs of individual patients. And medtech company Implicity is using the technology to incorporate remote monitoring in pacemakers and implantable defibrillators.

In brain health, DiagnaMed recently announced the development of a platform leveraging generative AI to analyze electroencephalography signals in order to predict and monitor brain aging and provide insights and tools in the diagnosis, prevention, or mitigation of cognitive decline in patients with mental health and neurodegenerative disorders.

Public-Health Agencies

Public-health agencies, other health organizations, and government ministries could leverage generative AI to improve resource planning and allocation, anticipate public-health needs and interventions, and execute programs more effectively.

BioNTech recently acquired InstaDeep in order to develop an early-warning system for new COVID-19 variants. Structural modeling of the SARS-CoV-2 protein combined with InstaDeep’s generative AI capabilities allows the system to proactively alert researchers, vaccine developers, health authorities, and policymakers.

Going forward, generative AI-powered tools could be used to monitor public health and allocate resources. In the US, Medicaid could potentially leverage the technology to better manage allocations based on health data and forecasted need. The FDA could use it when reviewing the safety and efficacy of drugs, and generative AI could help public-health groups like Doctors Without Borders predict outbreaks and mobilize resources to minimize impact.

Managing Potential Risks

Although generative AI technology is promising, some near-term caution is warranted. There are several inherent risks that providers must address before broad adoption in health care can occur.

• Biased Outputs. Generative AI results can reflect inherent biases in the underlying data. In response, generative AI companies need to assign experts to review the data and results and correct for bias through oversampling and other statistical techniques.

• False Results. Because the models are still evolving, they can sometimes generate results that are simply wrong (a phenomenon in AI known as hallucinating). Providers will need to make their models more transparent and emphasize the need for human review of outputs.

• Patient Privacy. Patient health data is sensitive and needs to be handled with extreme care. Companies with generative AI solutions should clarify data ownership with partners, strengthen cybersecurity, and look beyond existing data to the development of synthetic data.

• Opaque Results. A challenge with generative AI is that it operates as a black box, making it unsettling to some users. To build trust and increase adoption rates, health care organizations will need to explain how a given algorithm works and how a specific set of data leads to a prognosis.

• Misuse or Overreliance. Patients may rely too heavily on information from generative AI. Hospitals, clinicians, and payers should clarify how specific solutions should be used, with clear messaging that AI-generated insights are recommendations rather than mandates.

Sum Up

Sooner than later, generative AI shall be flowing in our landscape also. Given our relative handicaps such as lack of expertise, urge for expediency, tendency to ignore implicit problems, and aversion to invest into upgrades, training, and development, we run a much higher risk of going astray. We must understand that once we handover ourselves to technology, it will be well-nigh impossible to take it back.

Concluded.

Reference:

https://www.bcg.com/publications/2023/how-generative-ai-is-transforming-health-care-sooner-than-expected?utm_campaign=none&utm_content=202401&utm_description=top10&utm_geo=global&utm_medium=email&utm_source=esp&utm_topic=top10_in_2023&utm_usertoken=CRM_ec8a706ddd1afd128c0d41808e57173ddaa23495&mkt_tok=Nzk5LUlPQi04ODMAAAGQnPkD2bHSF2_0ayNEzLMATXEk_fb90MOrtd-l9NLSnX6KwnyWPVz74MN22AVbtRee7kHNNZoDE1CkZA0j_9BvLFo8T46isR7UK27VJWXC4xuX7Q