Vialto’s views on AI in Healthcare

?In the realm of healthcare, artificial intelligence (AI) and automation are frequently associated with high-dollar, high-stake and high-risk medical applications such as diagnostics and treatment.

These are nevertheless considered "high-hanging fruits" on the value tree—challenging to attain, costly, and entangled with numerous legal and liability concerns [1].

?

Conversely, more accessible opportunities lie within the bureaucratic and procedural aspects of healthcare.

Consider the typical patient experience: from scheduling appointments and filing insurance documents to clinical interactions where numerous metrics are recorded—such as blood pressure, various tests and stats data. Often, the doctor may be more focused on inputting data into an electronic health record rather than engaging with the patient. Subsequently, medical activities must be meticulously coded for billing purposes, a process prone to error due to its reliance on manual input. The resulting invoice is being sent to the insurance company and then to a payer company that actually does the paying and so on, and we haven't talked about a single medicinal procedure yet. Further complicating the system, different teams and functions within healthcare organizations – like clinical teams, pharmacy and procurement – speak their own languages, have their own ways of working and set their own goals and performance measures, which aren’t always aligned.

?

Each of these aspects presents opportunities for quick & low cost optimization & automation through AI.

?

By enhancing these processes, the system not only becomes faster and consumes fewer hours, thus reducing costs, but also becomes less prone to errors, requiring fewer corrections and reducing overall effort and further costs.

?

The most immediate and tangible benefits of AI in healthcare may thus be realized in streamlining these nonclinical processes, leading to quicker and more cost-effective medical services. This, in turn, could lead to improved health outcomes, as patients receive their medications and treatments more promptly.

?

Moreover, the application of AI extends into the clinical domain as well, where bureaucracy is also prevalent. Physicians often need to manage, refer, and translate a multitude of documents, such as notes or biopsy results, into systems that translate these into a wider team of people who are going to be forming a clinical pathway for patients. This process is frequently inefficient and wasteful, especially when it involves manually entering or writing information—an area known for being error-prone among healthcare professionals.

?

Process Intelligence enables today’s healthcare providers to objectively understand how processes actually run, identify improvement opportunities and act upon them. Using industry-leading object-centric process mining (OCPM), the Celonis Process Intelligence Graph (PI Graph) ingests data from all the different systems a healthcare provider uses and creates a process digital twin (an end-to-end, system-agnostic digital representation of how processes run). This digital twin is combined with the unique business context within which the organization operates, including KPI definitions, improvement opportunities and what makes something “good” or “bad” for the organization. The PI Graph illuminates the root causes of process deviations and identifies potential bottlenecks, revealing where value is hiding and provides intelligent recommendations around opportunities to unlock that value.

?

In conclusion, our inquiries and technical implementation projects we carried out with Celonis Process Intelligence [2] and RPA at Vialto, indicate significant potential for bureaucratic improvements in healthcare through AI, which could ultimately enhance patient outcomes far beyond just the diagnostic aspects. As the application of AI and Machine Learning (ML) in healthcare continues to evolve, significant advances are being made not only in diagnostic capabilities but also in streamlining healthcare processes, procedures, and clinical pathways.

?

You can read about some of the most impactful use cases [3] about Vialto Consulting, Celonis and our most relevant references [4] below.

?

----------------------------------------------------

1. what happens if you get that diagnostic wrong? At the moment most of the big and exciting uses for AI are either low-dollar, low-stakes and fault-tolerant applications (helping kids cheat on their homework, generating stock art for bottom-feeding publications) or high-stakes and fault-intolerant applications (self-driving cars, radiology, hiring, etc.). Very few are high-stakes and risk-tolerant applications. While AI decision support is potentially valuable to practitioners (accountants might value an AI tool’s ability to draft a tax return and radiologists might value the AI’s guess about whether an X-ray suggests a cancerous mass). But with AIs’ tendency to “hallucinate” and confabulate, there’s an increasing recognition that these AI judgments require a “human in the loop” to carefully review their judgments. In other words, an AI-supported radiologist should spend exactly the same amount of time considering your X-ray, and then see if the AI agrees with their judgment, and, if not, they should take a closer look. AI should make radiology more expensive, in order to make it more accurate. High-stakes AI refers to the increasing use of AI and machine learning in making life-and-death decisions – in areas such as public health, conservation or justice.

?

2. Vialto Consulting Ltd. (Celonis Gold Partner) conducted the first Process Mining implementation project in S. Korea – An on-prem Celonis implementation for Samsung (SFMI), at Samsung Headquarters in Gangnam, Seoul. Samsung is well known for its AI division and their development in image-based Diagnostics (Samsung CTN ultrasound machines can automatically detect things like cancers or other types of diseases with 99%, very high accuracy, immediately and at a very low costs).

?

3. The integration of AI and ML in non-diagnostic aspects of healthcare holds the promise of more efficient, personalized, and safe patient care. By addressing the bureaucratic inefficiencies and optimizing clinical pathways, AI can significantly contribute to enhanced patient outcomes, reducing costs, and improving the overall efficiency of healthcare delivery systems. As these technologies continue to evolve, their impact on healthcare is expected to grow, fundamentally transforming how care is delivered and managed.

Common Healthcare Business Processes

Standard healthcare business processes such as patient registration, appointment scheduling, billing, and claims management, health information management, inventory management, quality assurance, staff scheduling, and management, patient communication, compliance and regulatory reporting, and financial management are essential for the efficient operation of healthcare organizations.

All these processes can be optimized

A recent Study (https://www.researchgate.net/publication/381018134_Optimizing_Healthcare_Business_Processes_with_Process_Mining_Software_A_Comparative_Analysis) shows that various management practices lik process-based management, Lean Six Sigma, continuous improvement models, cost management, or value-based healthcare, can be used for improving efficiency, reducing waste, and adding significant value to the business. Implementing business process management (BPM) in healthcare settings such as is NSH, or HCA Healtcare, can enhance the standardization of processes, optimization, and ERP transformation, resulting in improved patient satisfaction, workforce conditions, operational efficiency, and financial performance. As IT systems and comprehensive process management started playing crucial roles in supporting primary and secondary care processes, optimizing care delivery, and improving the quality of care for patient, Process Mining has developped into the essential tool for reducing costs, improving processes, and reducing process time in healthcare organizations.

Process mining technique

Process mining, a data analytics approach, has been increasingly applied in healthcare to improve efficiency, reduce costs, and enhance patient satisfaction. It has been used to identify bottlenecks, streamline processes, optimize resource allocation, and automate workflow. Challenges in this application include data quality, algorithm selection, and presentation of results. Process mining has also been used to evaluate healthcare processes using the proposed goal-driven evaluation method. It has been applied to clinical care pathways in primary care, revealing insights and informing service redesigns. A methodology for process mining in healthcare, PM 2 HC, has been developed to provide guidelines for its application

Here are some staggering use cases to consider:

a) Administrative Workflow Assistance - AI and ML are being integrated into the management of day-to-day administrative tasks in healthcare facilities. These technologies can automate scheduling, patient data entry, billing, and claims processing. For example, AI-driven tools like chatbots can handle appointment bookings and reminders autonomously, while ML algorithms can analyze billing data to detect anomalies or potential fraud, thus enhancing the accuracy and efficiency of financial operations.

b) Enhanced Patient Record Management - Electronic Health Records (EHRs) are essential for storing patient data but are often cumbersome due to the volume and complexity of data involved. AI can aid in the efficient organization and retrieval of patient information, ensuring that healthcare providers have easy access to comprehensive patient histories. Natural language processing (NLP) algorithms can interpret and organize unstructured data, such as doctor's notes, into searchable and actionable information.

c) Optimized Resource Allocation - Machine learning models can predict patient inflow in hospitals and clinics, which helps in optimal staffing and resource allocation. This predictive capability can be particularly crucial in managing emergency department wait times and operating room schedules, reducing bottlenecks and improving patient throughput.

d) Remote Patient Monitoring and Engagement - AI-driven platforms can monitor patients remotely through wearable technologies, providing continuous health data to healthcare providers. This enables proactive management of chronic conditions and timely interventions, potentially preventing hospital readmissions and emergency visits. Moreover, AI can personalize communication between visits, improving patient engagement and adherence to treatment plans.

e) Clinical Decision Support - Beyond diagnostics, AI tools assist healthcare providers with decision-making by providing up-to-date medical information from various sources, including clinical guidelines and recent research. These tools can offer treatment recommendations based on patient history and prevailing clinical practices, ensuring that patients receive the most effective care tailored to their specific needs.

f) Streamlining Clinical Pathways - AI algorithms can analyze vast amounts of data to identify the most effective treatments and predict outcomes for specific conditions. This capability can be used to streamline clinical pathways by identifying the most appropriate treatment plans for patients and reducing variability in care delivery. For example, AI can help in designing personalized cancer therapy plans based on genetic information, clinical data, and outcomes from similar cases.

g) Enhancing Patient Safety - AI applications can monitor and analyze patient care activities to identify potential safety issues, such as adverse drug interactions or deviations from standard care protocols. Early detection of these issues enables timely interventions, thereby enhancing patient safety.

?

4. Process mining technology is relatively new but rapidly growing, and its application in healthcare is just beginning to unfold. This is an exciting frontier because it presents a unique opportunity to be at the forefront of innovation within the healthcare industry. While references in healthcare may currently be fewer compared to more established sectors like manufacturing or finance, this is primarily due to the unique complexities and stringent regulations of the healthcare industry which require careful, tailored implementations. Healthcare organizations are starting to recognize the significant benefits of process mining, such as enhanced operational efficiency, reduced costs, and improved patient outcomes through streamlined administrative processes. Early adopters in the healthcare sector are already seeing tangible benefits, setting a precedent for others to follow.

?Here are some organizations that are engaging with process mining early and by leveraging these advanced analytical tools are gaining a competitive edge. This positions them not only to improve their internal processes but also to enhance patient care by reducing administrative burdens and redirecting resources to where they are needed most:

?-?????????? NHS, the publicly funded healthcare system in England and the second largest single-payer healthcare system in the world started utilizing Celonis Process Mining (Process Intelligence) in its hospitals to identify high-impact areas for improvement, opportunities for efficiency gains and cost reductions to be deployed across the system. The success of this initiative underscores the value of employing Process Intelligence to optimize healthcare processes, improve patient outcomes, and enhance overall operational efficiency.

By making targeted changes based on real-world data and insights (all existing data), healthcare providers can achieve significant improvements in operational efficiency and patient care, even with minor adjustments to existing processes.

?-?????????? HCA Healthcare (HCA) [2], one of the world's leading healthcare service providers that operates over 180 hospitals and around 2,400 care sites in the US and UK is driving its organizational change by applying Process Mining to enhance patient experience and journey.

?-?????????? Queensland Health, the public health service of Queensland, Australia, with 16 Hospital and Health Services, beyond improving ist clinical area patiant jurney, interaction with providers and supply chain (procurement - P2P), also uses Process Mining for analysing ground and aero-medical pre-hospital transport processes involving the Queensland Ambulance Service and Retrieval Services Queensland and optimizing transport pathways across the time-critical phase of pre-hospital care for persons involved in road traffic crashes in Queensland.

?-?????????? NothWestern medicine a non-profit healthcare system affiliated with the Northwestern University Feinberg School of Medicine, in Chicago, Illinois are mining the contribution of intensive clinical course to outcome – a modeling strategy allows for the identification of clinical events associated with important prognostic clinical transitions… based on electronic health record data they are developing a modeling strategy that can delineate the clinical trajectories of ICU patients, etc.