A Perfect Match

The healthcare industry, which struggles with high spending levels and labor shortages, is in dire need of disruption--and AI has the potential to do just that.   Healthcare is facing many challenges, including increasing demand due to a growing and aging population, unsustainable growth of healthcare costs, physician and nursing shortages, lack on interconnectivity of patient and medical data, an alarmingly high medical error rate, health inequality, the constant need for innovation in pharmaceuticals and medical devices, and the increasing complexity of navigating healthcare systems, especially in the U.S. Given that the healthcare industry is a such a large and increasing contributor to countries' GDPs, with more than 17% in the U.S., AI would have a far-reaching effect, especially in an industry that has such a direct impact on human lives and is seen by many as a basic human right. In this article, we explore how major healthcare subsectors that will likely benefit from AI implement the new technologies and what, if any, credit implications they will face over the near term.

AI increases physicians' treatment capacity.   Growing healthcare demand and the increasing complexity of healthcare systems increased administrative burdens, led to burnouts among physicians and nursing staff, and resulted in a shortage of trained medical personnel. This has significantly reduced treatment capacities and led to longer wait times, higher expenses, and more patient dissatisfaction. Several studies have shown that administrative tasks, such as paperwork regarding medical notes, insurance claims, and prior authorization requests, now account for almost 50% of doctors' and 33% nurses' time. The use of AI, such as GPT-4, would increase medical staff's capacity to treat and interact with patients, as well as reduce burnout and turnover rates within the industry. The continued development of AI tools to help physicians navigate and use patient records could reduce the administrative burden, allow more time for patient care, and improve efficiencies.

AI leads to more efficient operations and reduces costs for providers.   In light of labor and inflationary pressures, AI's potential capability to cut costs and reduce the administrative burden becomes ever more relevant. Providers, such as hospitals, increasingly use AI in corporate and administrative functions, such as scheduling optimization, patient interaction, supply chain efficiency, clinical documentation, and revenue cycle management. Additionally, generative AI can increase retention rates and improve the recruitment of patients, expanding a provider's capacity for high-acuity procedures and lowering usage rates of higher-cost temporary staffing.

Clinical AI-based improvements could lead to changes in business models and reimbursements.   Hospitals and physicians use AI in diagnostics and imaging, patient monitoring, predictive analytics, and collaborative learning and information sharing, with encouraging results when it comes to early detection and treatment outcomes. The use of AI could lead to the proactive, rather than reactive, diagnosis and treatment of diseases and conditions. This could lead to earlier treatments, result in lower costs and healthier patients, and preserve hospital capacity to treat more serious cases.

Multiple use cases

AI has increased its footprint in the healthcare industry in the past few years.   The pace of AI adoption accelerated as its capabilities and uses expanded. Among others, AI can reduce medical errors, enhance early detection capabilities, improve productivity, shorten product development cycles, improve patient access, reduce overall healthcare system costs, and, overall, drastically reshape many healthcare subsectors' existing market structures. However, the pace of adoption differs among the various subsectors, which include insurers, government payors, service providers, pharmaceutical companies, medical device makers, and life sciences companies. We expect AI will disrupt existing market dynamics by shifting the existing power balance, changing risk-reward calculations for investments and acquisitions, and driving consolidation in the industry. Based on a survey conducted by 451 Research, more companies are planning to further focus on implementing AI into their operations.

Table 1

AI could affect ratings

These shifts in existing market dynamics will affect competitive positions and potentially result in rating changes.   We believe the pace of AI adoption in the healthcare industry is accelerating, and with newer, broader-based tools such as GPT-4, the effect is increasing. It is becoming critical that individual companies have an AI strategy and invest to implement, as AI is creating competitive advantages, especially for larger, more established companies, which have sufficient financial capacity to develop or acquire AI capabilities. This has negative rating implications for smaller issuers that cannot leverage the benefits from AI as quickly and could therefore suffer a competitive disadvantage. For smaller companies that implement AI strategies effectively, AI can have a leveling effect, where smaller and more agile companies with superior algorithms could expand rapidly and outpace current industry leaders, particularly in low-margin, highly fragmented areas. We expect the adoption of AI will lead to substantial fixed costs and margin improvements.

AI In Healthcare Services

AI is critical for the healthcare services sector because it can help reduce costs and improve patient outcomes.   For example, acute-care hospitals--key components of the U.S. healthcare services sector that make up a large portion of the issuers we rate in this subsector--tend to be expensive and very labor-intensive, at a time when healthcare labor, due to shortages and employee burnout, is at a premium. AI tools can help offset some of these higher expenses by lowering certain administrative costs while also improving efficiencies for organizations and clinicians. Over time, healthcare service entities and clinicians are also likely to incorporate AI with patient data, technology, and other tools to improve patient care and outcomes.

AI systems could assist clinicians across different patient interactions.   The potential ability of AI to be a supplemental tool and improve physician-patient interactions could improve the quality of healthcare, increase patient satisfaction, and reduce the burden on clinicians. Ambient listening and other AI tools, which help clinicians with patient responses and other patient interactions, could help with clinician burnout, while improving efficiency and quality. AI, such as GPT-4, can also integrate data from different sources, such as patient intake forms and medical records, and reveal insights about patients' health. Clinicians can then properly fill out prior authorization requests and speed up the process.

We expect healthcare services providers' AI adoption rates will vary, depending on technology advancements, data availability, regulations, reimbursements, and clinical research.   Consequently, we expect the effect of AI on rated issuers' credit quality will differ. Healthcare services providers with a track record of investing in clinical and operational transformation will likely be early adopters of AI, for example through internal investments and partnerships. Over time, however, some of these technologies and applications will trickle down to other providers.

Table 2

AI In Health Insurance

We consider health insurers, while still in the early stages of incorporating AI technologies in their corporate strategies and operational infrastructure, are best placed to benefit from AI, at least initially.   They have access to vast healthcare and demographic datasets, including member and claims data. Additionally, health insurers can obtain clinical information because many of them provide care delivery services and have in-house pharmacy benefit management operations. Moreover, the industry has the long-term strategic and financial flexibility to invest in AI, given relatively high barriers to entry and strong cash flows.

Insurers have begun to leverage generative AI to reduce costs and improve risk management and member engagement, with the overall goal of offering higher-quality coverage at lower costs.   From a credit perspective, AI has the potential to solidify a company's competitive position over the short term by enhancing member and provider engagement, increasing retention rates, and simplifying administrative processes. The industry initially focused on AI-enabled customer service interactions and claim management processes--including prior authorization--to cut labor costs, shorten customer wait times, and reduce processing errors. Over the long term, we expect insurers will increasingly use AI to create meaningful competitive advantages in areas such as pricing strategy and product design, medical outcomes and costs, and operating efficiencies. AI can help accelerate the adoption and success of value-based care models by improving predictive analytics, clinical decision-making, and personalized care plans for members.

Health insurers have two key advantages when it comes to AI adoption.   Firstly, their investment horizon is long-term because of their ample capitalization and, secondly, they can test and validate AI-enabled services and tools due to the significant amounts of data they have access to. Smaller health insurers could level the playing field by working collaboratively with peers or forming strategic partnerships with technology companies, such as Google, Microsoft, and Databricks, which focus on the healthcare sector as a key growth area.

Table 3

AI In Medical Devices, Life Sciences, And Diagnostics

AI in the medical devices, life sciences, and diagnostics realm is more widespread than in other healthcare subsectors, given the narrow focus of its application. However, as broader-based AI tools are used, their effect will increase.

Diagnostics and life sciences

AI has the potential to improve lives when used in diagnostics and early detection.   For instance, unsupervised AI models can be trained to uncover hidden patterns in data that the human eye would not be able to spot. In other words, AI is more effective than humans at detecting anomalies and can do so more quickly and at lower costs.

Generative AI is starting to make a mark in diagnostics.   For instance, U.S.-based biomedical cancer research center City of Hope created an oncology-specific large language model that can summarize and answer questions, structure data, and match patients to clinical trials, based on billions of words extracted from clinical, radiology, and pathology notes.

AI integration is relatively advanced in radiology and medical imaging.   Radiology accounted for 87% of approved AI medical algorithms in the U.S. in 2022, according to the U.S. Food and Drug Administration. AI already improves patient care by assisting physicians in identifying conditions more quickly, while predictive analytics are integrated into magnetic resonance imaging technologies. The latter results in better analysis and image diagnostics, more efficient workflows, and improved patient outcomes. AI is also becoming a regular feature in X-ray, computed tomography (CT) scans, and other imaging exams, resulting in improved accuracy and increased efficiency.

AI can help physicians synthesize imaging data with data from other sources.   As such, AI can determine treatments and potentially identify more effective options, based on an individual patient's likely response. GE HealthCare recently announced that its AI models can predict a patient's response, such as an adverse reaction, to immunotherapies with a 70%-80% accuracy. This enhances clinicians' ability to provide precision care.

Given the need to provide more accurate and faster diagnoses, the demand for more automation in laboratories has increased.   AI can drive digitalization in laboratories. The combination of AI with patient data, such as age and gender, and data that is gathered through routine lab tests could inform disease-specific predictive models and support physicians in their diagnoses.

AI-predictive models help clinicians diagnose cancer more quickly and accurately, while supporting the delivery of effective treatments.   AI and machine learning algorithms can automate and optimize many aspects of next-generation sequencing (NGS), which makes DNA and RNA sequencing more efficient. NGS data analysis, which can be complex and time consuming, improves the detection of gene variations by making the process faster, more efficient, and more accurate.

Table 4

Surgery planning and robotics

Robotics have been used in complicated surgeries since the 1980s and are now involved in over 644,000 surgeries in the U.S. annually, according to the U.S. National Institutes of Health.   Several companies aim to implement AI-based tools that analyze past surgeries and can improve the outcomes of future procedures. AI in surgery planning will provide guidance to surgeons, individualize implants based on patients' anatomy, and improve post-operative evaluations.

Table 5

Patient monitoring

Recent technology advancements enabled physicians to monitor patients' health remotely and in real time.   This significantly enhanced their ability to detect and treat problems early, while increasing the amount of data available for analysis. Remote patient monitoring supports outpatient services and can free up limited and expensive hospital beds. The integration of AI can not only enhance patient care through continuous data collection and analysis but also supports early intervention. Despite the benefits and potential of AI-based remote patient monitoring, however, the technology is far from widespread adoption. In 2022, a survey by the Medical Group Management Association showed that just 25% of polled practices used remote patient monitoring. This highlights a significant sales growth potential for monitoring devices, including wearable devices and implants.

AI algorithms can analyze vast amounts of patient data--including vital signs, such as heart rate or blood pressure--in real time.   The data are collected by wearable devices or sensors, which can detect deviations and flag any deteriorations that otherwise might have gone unnoticed--particularly if the deviations for an individual are within the normal ranges for the broader population. The early detection of patterns and anomalies can help healthcare providers intervene promptly and improve treatments.

Table 6

AI In Pharma And Biotech

AI can enhance pharma companies' drug discovery, increase their success rate, accelerate development and approval processes, and reduce overall development costs.   Patients could benefit from health improvements, resulting from new drugs, and potentially lower drug prices, while pharma companies could increase the number of approved products and improve their research and development spending. The largest biopharma companies currently require more than 10 years and spend more than $2 billion to bring a new drug to the market, according to the most recent Deloitte study published in 2024. AI, combined with a substantial amount of qualitative data, can enhance molecule selection by improving insights, predicting protein-drug interactions, and optimizing existing compounds. For example, Moderna used AI to release the first clinical-grade batch of its Spikevax COVID-19 vaccine 65 days after the virus sequencing. AI-driven tools help pharmaceutical researchers "fail intelligently," an evolution of the older "fail fast, fail often" approach. Additionally, they improve cost and time allocations.

AI and machine learning could lead to new successful novel compounds, which would accelerate the discovery of new drugs.   Supervised machine learning could be trained to automate medical image analysis and diagnosis, which enables healthcare professionals to dedicate more time to more significant tasks. Beyond that, AI could help predict regulatory queries, thus cutting the regulatory approval times by months and reducing repetitive back and forth between companies and regulators.

AI could lead to better target selection and faster development times.   However, drug development failure rates remain high and development costs continue increasing. Against this background, the use of AI could result in major improvements over the near to medium term. Moreover, advances in cell and gene therapy and in mRNA-based treatments, such as personalized vaccines, can increase the use of AI in identifying treatment plans with the highest likelihood of success.

AI will affect all companies in the pharma drug development chain, from contract research organizations (CROs) and contract development and manufacturing organizations (CDMOs) to big pharma companies.   Leading clinical research organizations are actively adopting various AI technologies, whether by developing them in-house or through partnerships, to streamline various aspects of the clinical trial process, such as patient recruitment, data analysis, pattern recognition, and identification of potential adverse events. For example, by analyzing patients' data--including genetics--past medical histories, and other relevant factors, AI can make clinical trials more efficient and reduce the time and cost it takes to bring a drug to the market. CROs' and CDMOs' AI advancements can help smaller biopharma companies take advantage of AI, without them having to invest heavily in the technology. The widespread adoption of AI by big pharma companies creates risks to the continued outsourcing of noncore services, particularly in clinical research and drug development and manufacturing.

As is the case in other healthcare subsectors, AI in the pharma industry particularly benefits larger companies that have better access to data and financial resources than smaller companies.   Even so, we think AI will create a level playing field as open-source AI, which is provided by companies such as Merck, enables smaller companies to compete through partnerships and outsourcing. Generative AI, which enables pattern recognition and helps develop new molecules, has led to several new biotech startups and potential new competitors.

Table 7

AI Raises More Questions Than It Answers

The medium- and long-term effects of AI on healthcare organizations and, ultimately, credit quality remain uncertain. For example, allegations of automated claim denials in the insurance industry give a foretaste of the significant challenges issuers could face when AI becomes an important component of healthcare decisions. We also note that untransparent algorithms prevent quick resolutions, delay widespread adoption, and will likely result in legal challenges for early adopters. Additionally, intellectual property challenges regarding the use of AI in product development, including its use in training AI models, are likely to rise as the commercial effects of AI become more meaningful.

Related Research

• Artificial Intelligence Powering Synthetic Biology: The Fundamentals, June 25, 2024