AI Is Here to Stay: Adopting AI for Cancer Service Line Strategy
Editor’s note: Senior Consultant Kiran Yemul contributed to this post.
Artificial intelligence–based solutions have quickly moved from being a novelty to providing novel solutions to complex health care challenges. And while these solutions may only be scratching the surface of the value they will deliver over time, this exponential improvement has positioned artificial intelligence (AI) to disrupt nearly every aspect of the clinical enterprise. In fact, at this very moment, AI is not only already being utilized, but it is also improving clinical care across service lines.
Let’s take a look at how the cancer service line is utilizing AI applications today.
Cancer: AI Has Multifaceted Impacts, From Screening to Treatment
Of all the areas where AI can provide value across the care continuum, oncology is already well ahead of the curve, applying AI to challenges faced by organizations nationwide. In radiology, AI is being integrated into imaging solutions to enable earlier, more precise cancer detection, including Google’s LYmph Node Assistant (LYNA) program, which made waves with 99% accuracy in detecting advanced breast cancer in pathology images. Similarly, a large study of AI in low-dose CT lung cancer screening images found that AI detected 11% fewer false positives and 5% fewer false negatives than study radiologists. Improved cancer detection will have a significant impact on patient outcomes, treatment selection and costs.
Furthermore, improved cancer detection has also led to the use of AI-based risk prediction tools to facilitate targeted treatment regimens. In a recent study, AI predicted which premalignant breast cancers were more likely to progress to invasive breast cancers, while a similar application of AI assessed which lung cancer patients would likely benefit from chemotherapy based on digital pathology. Such tools could potentially lower costs by reducing unnecessary service utilization (eg, fewer radiation therapy fractions equals less need for genetic panels).
Finally, AI is being explored as a potential asset in determining when to initiate end-of-life care conversations, which can be particularly challenging for many oncologists. Recent research from Penn Medicine investigated the use of AI to predict short-term mortality risk and to encourage physician initiation of advanced care planning conversations. This type of approach could generate significant savings through reduced spending on maintenance therapies such as noncurative chemotherapy.
Service Line AI Adoption Recommendations
Overall, organizations must evaluate existing operations and practices before introducing AI and understand that to realize the maximum benefits of any emerging technology, a significant care redesign effort is necessary. It is essential to create proactive, supportive and consistent care pathways that clinicians can use to guide interventions when AI identifies patients who could benefit from preventive or immediate health care.
When it comes to services lines specifically, however, Sg2 recommends:
- Avoid unintended consequences of AI adoption. In oncology, this may take the form of additional costs if efficiencies are not realized as a result of lack of support in care redesign efforts.
- Align your AI utilization to service line–level priorities within your organization. For cancer care, improve the diagnostic pathway and broaden appropriate research trials.
- Deploy a portfolio approach to your AI strategy. There will be some AI solutions that are more administratively oriented, clinically driven or operationally impactful. Additionally, various AI solutions have variable financial impacts. Structure your AI strategy to ensure you are balancing AI solutions that drive revenue or cost savings with those that may be an expense to the system but are critically important to delivering efficient, effective clinical care.
As AI has grown in sophistication, it has quickly become one of the most widely discussed, analyzed and invested-in technologies in the health care industry. It has also created a very challenging and complex environment for health care leaders who are attempting to determine when to invest in AI and how to develop their organizations’ AI capabilities.
Reach out to Sg2 for help with your AI strategy, and stay tuned for more from us on service lines and AI.
Tags: AI, cancer, cancer detection, Care redesign, clinical care, end-of-life care, oncology, risk prediction tools, treatment