The Growing Role of Artificial Intelligence in Ambulatory Health Care

 

Introduction

Yesterday, I wrote about the use of Robotic Process Automation (RPA) in health care. Today, I want to discuss the growing role of artificial intelligence in ambulatory health care.

Artificial Intelligence (AI) is reshaping nearly every aspect of modern health care, and ambulatory care settings such as outpatient clinics, physician offices, and same-day surgery centers are no exception. Ambulatory health care has traditionally relied on high patient volume, efficient workflows, and rapid decision-making. With AI emerging as a key enabler of smarter operations, organizations are finding new ways to enhance efficiency, improve diagnostic accuracy, and strengthen patient engagement outside of hospital walls.

What AI Means for Ambulatory Health Care

AI in ambulatory care involves the use of machine learning, natural language processing, and predictive analytics to automate tasks, assist in clinical decision-making, and deliver personalized patient experiences. Unlike inpatient settings, ambulatory care focuses on long-term, short-term and preventive services where time and data coordination are critical. AI systems can quickly analyze patient data, detect patterns, and offer actionable insights that help clinicians make faster and more informed decisions.

Key Use Cases of AI in Ambulatory Care

1. Predictive Scheduling and Resource Management

AI algorithms can analyze historical patient flow, appointment types, and seasonal trends to optimize scheduling and reduce no-show rates. Predictive analytics helps allocate staff and resources more effectively, improving both operational efficiency and patient satisfaction.

2. Clinical Decision Support

AI-powered systems can assist clinicians by analyzing patient data, lab results, and medical history to recommend potential diagnoses or treatment options. In an outpatient setting, this can speed up consultations and ensure early detection of chronic conditions such as diabetes, hypertension, or COPD.

3. Patient Engagement and Virtual Assistance

AI chatbots and voice assistants can manage routine communication such as appointment reminders, pre-visit screenings, and post-visit follow-ups. This frees up administrative staff while improving accessibility for patients who prefer digital interaction.

4. Documentation and Coding Automation

Natural language processing (NLP) tools can transcribe clinical notes, extract relevant information, and automatically assign billing codes. This reduces manual documentation, minimizes errors, and allows clinicians to spend more time with patients.

Ambient listening technology such as that provided by Nuance DAX Copilot and Abridge can listen in on a patient encounter and transcribe the pertinent information into the note section of the patient’s chart in the electronic health record (EHR) system, completing 80-85% of the encounter note for the provider to later review, edit and sign. This frees the provider up to have a more positive interaction with the patient without having to be buried in a laptop taking notes. It also makes the encounter more efficient, allowing the patient to be able to discuss more of their issues with the provider leading to better overall outcomes.

5. Predictive Analytics for Identifying At-Risk Patients

AI algorithms can be applied data stored in an EHR system using a risk-scoring model looking at such data as medical history, social determinants of health, lab results and trended vital stats, treatment plans, medications and urgent care or emergency department and/or inpatient admissions to identify patients who may be at risk of being admitted or readmitted to the hospital. Identifying these patients allows the provider to reach out and potentially intervene to prevent an admission or readmission which could make a huge difference in the health of the patient. Waiting until a condition worsens to the point where the patient has to be admitted to the hospital could have adverse effects on the future health of the patient. Not to mention the huge risk of potential hospital-borne infections, especially for those patients who are immune-compromised. Additionally, these avoiding these preventable admissions or readmissions free up health care dollars and resources that can applied elsewhere to further the health of the community.’

This is rather new use that is still developing and will continue to evolve as more and more data is used to train and refine the models and experts continue to calibrate them. This could also lead to breakthroughs in identifying new protocols and further developing and applying existing ones.

6. Population Health and Risk Stratification

AI can aggregate and analyze data across patient populations to identify those at higher risk of hospitalization or disease progression. Ambulatory clinics can use these insights to implement targeted preventive interventions and reduce avoidable admissions.

Protecting Patient Health Information (PHI) in an AI Environment

As AI systems handle sensitive health data, maintaining privacy and compliance with regulations such as HIPAA is essential. The use of AI in ambulatory care introduces specific security challenges, including the protection of EHRs and the safe handling of data used for model training.

To mitigate these risks, organizations should adopt the following safeguards:

 

1. Data De-Identification: Remove personally identifiable information before using data to train AI models.
   
   
2. Encryption and Access Controls: Encrypt all PHI at rest and in transit and use role-based access to limit data exposure.
   
   
3. Audit Trails and Monitoring: Maintain detailed logs of who accesses data and when to ensure accountability.
   
   
4. Vendor Due Diligence: Evaluate AI vendors for compliance with HIPAA, HITRUST, and other relevant standards.
   
   
5. Human Oversight: Keep clinicians and compliance officers involved to review AI recommendations and identify potential biases or anomalies.
   
   
6. Local Storage and Execution of Models: Wherever and whenever possible, store and run AI models locally. While this requires more compute resources and deeper knowledge to train the models, it ensures that PHI doesn’t leave the organization through the use of a shared model. As models become more specialized to particular tasks and industries, they are also becoming much smaller. Something the size of ChatGPT is not required to run the use cases that we just discussed as models like ChatGPT, Gemini, Claude, etc. are created to handle a wide-open array of questions and requests which require an enormous breadth of data, while the uses we just discussed only require specialized data, such as patient data, charges, etc. making them substantially smaller in size. In fact, some models have been demonstrated that can be run on just a laptop and accomplish amazing tasks! Some vendors may also train and deliver models for on-premises use that can be delivered and retrained at regular intervals.
   
   
7. Create an Organizational Policy Covering the Use of AI: Just as most organizations have policies covering security, having an AI policy that governs the use of AI in the organization is key. Additionally, these policies should be written in such a way that they can be evolved and updated as the technology and use of it grows.

Security and privacy must be treated as foundational elements of responsible AI use in health care.

The Future of AI in Ambulatory Health Care

The future of AI in ambulatory care is deeply connected to interoperability and precision medicine. As electronic health records, wearable devices, and remote monitoring systems continue to generate vast amounts of data, AI will play a central role in integrating this information and making it clinically useful.

In the coming years, we can expect advances in:

• Personalized treatment recommendations that adapt to individual patient profiles.
  
  
• AI-Driven Preventive Screening Tools that analyze biometric, genomic, and lifestyle data to identify early indicators of disease before symptoms appear.
  
  
• Automated Clinical Workflow Optimization that continuously monitors clinic operations, identifies process bottlenecks, and recommends real-time adjustments to improve throughput and efficiency.
  
  
• Emotion and Sentiment Analysis for Patient Interaction that enables AI to assess tone and engagement during patient communications, helping clinicians identify patients at risk of anxiety, depression, or disengagement from care.
  
  
• AI-driven triage systems that guide patients to the appropriate care setting before they arrive at a clinic.
  
  
• Synthetic Data for Research and Model Training that allows developers and researchers to create realistic, privacy-safe datasets for testing and innovation without exposing real patient information.
  
  

The most effective AI systems will act as augmented intelligence, enhancing rather than replacing human clinicians. When designed responsibly, AI can empower health care teams to deliver more efficient, data-driven, and compassionate care.

Conclusion

AI is transforming ambulatory health care by reducing administrative burdens and allowing clinicians to focus on what matters most: patient well-being. By combining automation with strong data governance and security practices, health care organizations can confidently adopt AI while maintaining patient trust and privacy. As technology continues to advance, AI will help make outpatient care faster, more accurate, and more human-centered.

Revolutionizing Health Care with Robotic Process Automation (RPA)

 

Introduction

The health care industry faces constant pressure to accomplish more with fewer resources. Hospitals and clinics must handle increasing patient volumes, growing data demands, and complex regulations, often with limited staff and budgets. To meet these challenges, many organizations are turning to Robotic Process Automation (RPA). RPA uses software “robots” to automate repetitive, rule-based tasks, allowing clinicians and administrative staff to focus more on patient care and less on paperwork.

What is RPA in Health Care?

RPA involves the use of software bots that mimic human actions to complete digital tasks across multiple systems and applications. These bots can log into systems, enter or extract data, and perform calculations without human intervention. In a hospital or clinic, this means that bots can process insurance claims, manage patient records, and schedule appointments, all while reducing manual effort and human error. The goal being to take low level processes that don’t require human judgement off of employees to free them up for much higher value work that does require human judgement and interaction. These higher value tasks could be patient care, system troubleshooting in IT, deeper data analysis and forecasting in Finance and Operations, etc.

Key Use Cases of RPA in Health Care

1. Patient Data Management

RPA can automatically collect, verify, and update patient information from various systems, ensuring that electronic health records (EHRs) remain accurate and current. This helps reduce data entry errors and administrative delays while freeing up clinicians to spend more time with patients.

2. Claims Processing and Billing

Insurance claims often involve complex, time-consuming processes that are prone to errors. RPA bots can check claims for accuracy, submit them to payers, and follow up on rejections. Automating these steps speeds up revenue cycles and minimizes costly mistakes.

3. Appointment Scheduling and Reminders

Managing appointments requires coordination between patients, staff, and multiple systems. RPA can simplify this by integrating scheduling tools, sending automated reminders, and rescheduling appointments when necessary. This improves efficiency and reduces missed appointments.

4. Regulatory Compliance and Reporting

Health care organizations must adhere to strict regulations such as HIPAA and CMS requirements. RPA can assist by tracking access to sensitive data, generating compliance reports, and maintaining detailed audit logs to ensure accountability.

5. Supply Chain and Inventory Management

Hospitals depend on a steady supply of medications, instruments, and protective equipment. RPA bots can monitor inventory levels, generate purchase orders, and alert staff when supplies are low. This prevents shortages and reduces waste.

Benefits of RPA in Health Care

• Improved Accuracy: Automation reduces manual data entry errors and inconsistencies.
• Increased Efficiency: Tasks that once took hours can be completed in minutes.
• Enhanced Patient Experience: Staff can dedicate more time to patient interaction and care.
• Scalability: RPA systems can handle larger workloads without additional staffing.
• Cost Reduction: Automating repetitive processes lowers labor and operational costs.

Challenges and Considerations

Although RPA provides many benefits, health care organizations must address certain challenges such as system compatibility, data security, and employee adoption. A successful implementation requires careful selection of processes for automation, strong collaboration between IT and clinical teams, and consistent performance monitoring to ensure the bots function properly.

The Future of RPA in Health Care

The next evolution of RPA combines automation with artificial intelligence (AI) and machine learning (ML), creating what is known as Intelligent Automation. These advanced systems can not only complete routine tasks but also make informed decisions, detect irregularities, and predict outcomes. For example, intelligent bots could prioritize urgent patient requests or assist clinicians in making diagnostic decisions. This integration of automation and intelligence promises a more proactive and adaptive health care environment.

Obviously, any introduction of AI into a health care context requires careful consideration of HIPAA privacy and security regulations and the protection of protected health information (PHI). As you evaluate Intelligent Automation solutions, protection of PHI has to be of paramount importance. In many cases, this may involve localized implementation of AI processes where large language models (LLM) are hosted and managed on-premises or in a cloud implementation controlled by the organization to ensure that PHI are not leaked out into broader implementations of the LLM that could lead to leakage outside the organization and a breach event. This is where organizational policies regarding the use and governance of AI are crucial.

Conclusion

RPA is transforming the health care industry by reducing administrative burdens and allowing staff to focus on what matters most: patient well-being. As automation technology continues to evolve, it will play an even greater role in improving safety, efficiency, and overall quality of care. For health care organizations ready to innovate, RPA represents a major step toward a smarter and more sustainable future.

Exploring the Wonders of Neural Networks and Large Language Models

 

 

Photo by Google DeepMind on Unsplash

Hey there, curious minds! Today is the second in my series on Generative AI. Picking up where we left off last time, today we're diving into the fascinating world of neural networks and large language models. Don't worry if these sound like big words – we'll break them down into bite-sized pieces so you can understand them like a pro!

What are Neural Networks?

Imagine you are solving a puzzle, and to do that, you have many friends helping you. Each friend is good at solving a specific part of the puzzle. Similarly, a neural network is like a team of friends that work together to solve a problem. But instead of a puzzle, they handle tasks that computers find challenging, like recognizing objects in images or understanding human language.

The neural network is made up of artificial brain cells called "neurons." Each neuron takes some input, thinks about it, and then gives an output. They work together in layers, with each layer adding more complexity to the problem-solving process. The information flows through the layers, just like passing a message along from one friend to another until they solve the puzzle together.

For example, when you show a neural network a picture of a dog, the first layer might notice basic shapes, like circles and lines. Then the next layer combines these shapes to identify parts of the dog, like its nose or ears. Finally, the last layer puts everything together and says, "Hey, this is a dog!"

What about Large Language Models?

Alright, let's shift gears a bit and talk about large language models. Have you ever asked a smart assistant a question or typed something into a search engine, like, "What is the tallest mountain in the world?" Well, behind the scenes, there might be a large language model working hard to give you the answer! Large language models are like super-smart robots that can understand and generate human-like language. They can read stories, answer questions, write essays and even have conversations!

Language models are trained on a vast amount of text from books, articles, and websites. They learn patterns and relationships between words, just like you learn new things by reading lots of books. With this knowledge, they can predict what words come next in a sentence or even complete a whole paragraph by themselves!

When you ask a language model a question, it uses its vast memory to find the most likely answer based on what it has learned from all the text it has read. It's like having access to the world's biggest library, and the model can fetch the most relevant information for you!

How They Work Together

Now, the magic happens when we combine Neural Networks and Large Language Models. Imagine a giant puzzle, one that is so big that it would take a team of people a very long time to solve. But with a neural network and a language model working together, it's like having a super team of puzzle solvers!

The neural network can understand the puzzle's structure and divide it into smaller parts, just like your friends in the puzzle example. Then, the language model comes in, using its vast knowledge to help solve each part. It's like your friends asking for advice from the smartest people they know!

Just imagine having your very own super helper who can assist you with anything you need, like solving math problems, explaining tricky topics, or even telling you fun facts!

Applications of These Technologies

You might be wondering how we use these incredible technologies. Well, they power many things around us! When you use voice assistants like Siri or Alexa, they rely on language models to understand and respond to your questions. When you watch videos on platforms like YouTube, neural networks help recommend videos you might like based on what you've watched before.

Final Thoughts

Neural networks and large language models are incredible creations that use the power of computers to make our lives better and easier. They can do amazing things that would take humans a very long time to accomplish. They are kind of like a dream team of problem solvers and knowledge gatherers that make our digital world more exciting, helpful and smarter. As we continue through this series, we will dive deeper into how you can use the applications backed by these models in your business and personal life and even look at how you can integrate these into your programs.

 

Generative AI, What Is All the Hype About Anyway?

 

www.freepik.com

Over the past few months, I have been delving into this subject called Generative AI. Unless you have been trapped on a desert island somewhere without any communication with the outside world, it is pretty difficult to miss all the talk and hype around this subject. As someone who makes his living off of technology and, sometimes bleeding edge technology, I felt that it would be wise for me to start learning a little bit about this. What I didn’t expect was the rabbit hole that I would be going into as I began to learn about this. This post is the first in a series of posts about what I have learned in an attempt to bring you, my readers, up to speed on this subject without having to go through the hundreds of blog posts and papers that I have. My goal with this series is to enlighten on the topic of Generative AI and hopefully, give you a jumpstart into this space and teach you about the components of Generative AI; show you how magic works to provide you the answers you receive when you ask ChatGPT, Claude 2 or Bard a question or give it a task; delve a little bit into the current implementations of Generative AI that out there; and give you some insight into how to use these tools through something called prompt engineering.

What is Generative AI?

So, let’s get started by defining what this thing is that we call Generative AI. In a nutshell, Generative AI is a lot like having a genius friend who can consume and remember any piece of information that has been fed to it. In turn, it can use this knowledge to invent new and exciting things like art, music, writing stories and even help scientists and experts solve problems. Okay, so that is pretty high level, generic definition that could have come from the marketing department. So, let’s take a little deeper look into Generative AI and talk a little about how it works.

Where does all of this knowledge come from?

Behind a Generative AI application there are complex machine learning models that are trained to understand the large datasets that are fed to it. A machine learning model is basically a program that allows computers to learn without having to be explicitly programmed. In the days before AI (or before most of us knew about it anyway), computers had to be given explicit instructions via a program for everything you asked it to do. For example, when you click on an icon to launch a program, there is a set of instructions or code that tells the computer what it is supposed to do, i.e., go to the exact path where the executable file associated with the icon that you just click is stored and run it. Everything you ask your computer or a specific application to do requires explicit code that the computer runs to respond to your request. Machine learning allows computers to be able to take things a step further and allow computers the ability to reason or, more accurately, predict what it should do when you ask something of it. Now, for the foreseeable future, this won’t impact the way your computer works. For the next few years, it will still need explicit code in programs to tell it how to respond to a certain user action.

What machine learning and Generative AI is starting to give us today is systems that can be used as the basis to solve complex problems; create automated systems, such as chatbots, that can answer questions and provide you contextual information, freeing up humans to do more high value work; do things like create websites or small applications based on a few sentence description that the user inputs; create photo-realistic artwork based on the user entering a few words describing what they want; create outlines for ad campaigns, papers or even write stories or ad copy; analyze patterns of behavior to catch criminals; and even create formulas for new and exciting pharmaceutical medications. And this is just the start. We in are very early days in this process even though researchers, scientists and data engineers have been working on this for decades. In the long-term, we can expect that AI will be used to offload even more mundane or low-level tasks, automate monitoring of systems and, of course, bring us widespread adoption of the long promised self-driving vehicle.

Now that we have some understanding of what machine learning is, let’s talk a little bit about how these models are developed and how they work.

Machine learning basics

Sidebar: The term machine learning was actually coined way back in the 1950’s, which gives you an idea of just how long this whole process of developing so-called artificial intelligence has taken. What’s even more amazing is that wasn’t even the beginning of the groundwork for artificial intelligence. The roots of AI actually go all the way back to the early 1900’s. The biggest movement in AI didn’t really start to happen, though, until the 1950’s.

Machine learning is basically the process of inputting large amounts of data into machines (we’ll refer to these as computers going forward) and having those computers analyze that data for patterns which they can then turn around and use to make predictions when it is given a task or question relating to the data it has been given. This isn’t a one and done process. It is actually a very iterative process that usually requires several different phases and methods of machine learning. Initially, it might be use something called supervised machine learning which is where the computer receives labeled data such as pictures that are labeled as cat or dog or other things that have been classified or labeled by humans. This allows the computer to begin to recognize things. If you feed enough different labeled pictures of cats and dogs, it will start to be able to recognize or predict based on that information whether a picture that you show it is either a cat or dog. Obviously, this is a very time-consuming endeavor, but is really necessary to teach the model how to recognize basic things.

Unsupervised machine learning is where computers begin to be able to recognize patterns from unlabeled data. This could be anything such as data about your customers, marketing, stock prices, etc. Unsupervised machine learning then uses a technique known as clustering to segment this data to try to discern patterns. With clustering, the computer uses rules to determine the probability that a particular item should belong to a certain group or cluster.

Reinforcement learning is basically the feedback mechanism to tell the model whether the choice that it has made is correct or not. Again, this requires many iterations and quite a bit of human intervention along the way. Eventually, the model has enough data that it can begin to start evaluating its decisions and learn from its mistakes. This can also involve additional models that can act as a checkpoint for the model and provide feedback.

These models are all very complex and use probability and statistics to make their predictions and evaluations. One of the things to keep in mind with these models is that they are using data created by humans. As such, that data carries the biases of its human creators and can cause the models to behave in unintended ways, therefore, these models have to be continually monitored and tuned by humans to try to correct for these potential inherent biases.

We have just scratched the surface of what is involved in machine learning as this is an entire unto itself, but we laid a good foundation for you to begin to understand what is behind the Generative AI application that you may be using. In the next post, we will talk about neural networks and large language models.