GenAI is tailor made for the challenges that healthcare faces: CitiusTech

CitiusTech is a leading provider of consulting and digital technology to healthcare and life sciences companies. It is strategic partners to the world’s leading payer, provider, MedTech and life sciences companies to drive innovation, business transformation and industry-wide convergence.

Sridhar Turaga, SVP, Data and Analytics Practices, CitiusTech, tells us more. Excerpts from an interview:

DQ: Can you provide insights into the current state of GenAI adoption in the healthcare industry, and its potential impact on the overall healthcare ecosystem?

Sridhar Turaga: Traditionally, healthcare has been slower and more conservative in adopting new technologies, compared to other industries due to a highly regulated environment, safety-first mindset and fragmented nature of the industry. With Gen AI, almost every healthcare organization wants to experiment and explore.

Firstly, that is because Gen AI is particularly suited for the handling of unstructured data such as text, documents, and images. Data in healthcare is 70% unstructured. For instance, 70% of hospitals in the US still use faxes to communicate. So, in some ways, GenAI is tailor made for the challenges that healthcare faces.  

Secondly, Covid has thrown up huge cost challenges across the world and has exposed many inefficiencies. Several healthcare organizations are still reeling from that impact, and Gen AI offers many ways to reduce costs. Thirdly, Gen AI can reduce the burden that physicians and care teams struggle with due to all the documentation needed in healthcare.

DQ: Can you share specific examples or use cases in healthcare where GenAI is already making a significant difference? How is it augmenting human capabilities and enhancing clinical decision-making?

Sridhar Turaga: Many hospitals and EMR systems are already implementing summarization of clinical notes. In fact, some have linked voice recognition of patient - physician conversation and are auto creating clinical notes for the physician - a huge relief to their administrative burdens.

Health plans call center clients are exploring how Gen AI can create agent assist solutions – as even a simple benefits verification call is a huge task due to the fragmented nature of information. This use case alone has the potential to unlock 20-30% of call center costs for health plans. There are multiple opportunities to search, retrieve and summarize research information, clinical guidelines and patient education information using Gen AI.

One thing I would caution against is its application to clinical decision-making. Gen AI lacks the rigor and verticalization to play in that area as of now.

DQ: What role does GenAI play in addressing key healthcare challenges, such as cost reduction and improving overall healthcare efficiency?

Sridhar Turaga: The biggest opportunity is cost reduction in areas like member support, prior authorization etc. The second opportunity is reducing the administrative burden of physicians and care teams. The third biggest application is hyper-personalization for patients and members.

DQ: What challenges or considerations should healthcare organizations keep in mind when implementing GenAI solutions?

Sridhar Turaga: The first challenge is cost-at-scale. PoCs and prototypes are easy. Once you scale up, the cost of usage can shoot up quite rapidly. The transaction costs of using LLMs are very high. So, picking the right use cases, determining the RoI, and designing for scale is key.

Second challenge is quality -- how does one measure the responses and results are accurate, relevant, creative, and readable? How can one go beyond subjective measures, into more scientific ways of measure thinks like bias and human-like nature of results. Third challenge is hallucination. How does one measure and control for hallucination? The other side of the creativity coin.

Then, there are a few cultural challenges to overcome. Given the pace at which Gen AI is evolving, one needs to question how they can inculcate an experimental mindset, as there are no definitive answers. Then, there are challenges specific to healthcare. Many of the LLMs are horizontal and don’t understand the healthcare context. For example, the word ‘coverage’ or ‘value’ has a meaning in healthcare that is very different from how we use them otherwise.

DQ: What are some of the most promising future trends and innovations in GenAI that will continue to shape the healthcare landscape? How does CitiusTech plan to stay at the forefront of these developments?

Sridhar Turaga: CitiusTech has created a multi-functional task force reporting directly into the CEO. This allows us to bring AI, Consulting, Engineering, Data experts together to create a path forward. We also have created a nuanced healthcare POV and solutions showing customers what Gen AI can do for them. In fact, some of these are not just recognized as accelerators, but have also now been selected by top cloud providers as a joint offering to their clients.

We are also investing heavily in training all our CTzens on the technology itself, business solutions and in prompt engineering. We are betting big on open source LLMs models and vertical LLMs. We have also built one of healthcare industry’s first quality monitoring framework and are about to launch a Gen AI design studio.

In addition to all business solutions, we are actively investing on engineering productivity with Gen AI across areas like auto code generation, documentation, migration, and QA. Our vision is to be the largest Gen AI-powered healthcare technology services company.