There has recently been a lot of talk around explainable AI (XAI). How can we open up the ‘black box’ around AI to better understand what decisions it is making, what data it is using to make those decisions and how can we have resources get more confident in ’s ability to make decisions?

One area where (XAI) holds a lot of promise is in health care. The question becomes, is the promise of (XAI) going to follow a similar hype as ()? has been around for over 50 years, with at least 3 time periods where it was going to take over the world, in the 1960’s, in the 1990’s, and now. Will it take 5 years, 10 years, or longer for explainable AI (XAI) to be adopted?

Health care has to change and explainable AI (XAI) might just be the push the ecosystem needs to transform itself.

The areas around pathology, radiology and dermatology have all seen advancements in AI. There have been studies that show great strides, with AI models; where AI is better at detecting melanomas than dermatologist. Just because a model is better at detecting something, doesn’t that mean that it is ready to make decisions autonomously.

These are business and challenges that need to be reviewed and understood on a case by case basis. Not every model that gets created will need additional rigor about how it came to its decision. You need to fully understand a system before you can allow it to make decisions without human involvement if that is the goal.

This is where the breakdown starts to happen. Technical teams love experimenting with different technologies and generally look to fully automate their models. Business and owners aren’t looking for full automation of models because they won’t blindly trust a model. They don’t have confidence in how the model is working and transparency is absolute. The regulatory constraints that healthcare has to operate within, provide an opportunity for explainable AI.

This is where business, and technical stakeholders have to all be on the same page. Business and owners have to be involved in working with their technical resources to define what parts of the organization they are comfortable with systems making decisions autonomously. What areas should systems be and verified by stakeholders? Do business owners, owners and technical owners understand what the risks are around the systems that are being created?

The healthcare ecosystem could be greatly advanced with  (XAI). With many strides that academic institutions like the University of Toronto are making and technology companies like Google continuing to do research around (XAI), how can healthcare leverage these advancements to transform itself? Technology, Business and owners will all need to start by working in smaller and less complex areas to test their hypothesis. Business and owners will need to push their technical teams from the beginning to design models that can be explained. Business owners will need to start asking more questions around the , regulatory, and financial risks of the models.

The hope of (XAI) is that future models will be able to share the features that it used to make the decision. There are two open source projects 1) Local Interpretable Model-agnostic Explanations (LIME) which looks to help explain what AI models are doing and 2) SHAP (SHapley Additive exPlanations) is a unified approach to explain the output of any machine learning model. Both of these projects have repositories on GitHub that can be utilized by technical teams to start creating explainable AI (XAI) models.

Health care has to change and explainable AI (XAI) might just be the push the ecosystem needs to transform itself.

The post Health Care and the Promise of Explainable Artificial Intelligence (XAI) appeared first on HealthTech Magazines.

By Molly Menton, Director, Clinical Delivery, Evolent Health

Technology has changed our lives in every imaginable way. Every industry, every sector, and every consumer has been irrevocably changed by the digital revolution. 50 years ago, it would have been only in the Jetsons cartoon that people would make purchases without paper money, go shopping without going to the mall, and call a cab without using a telephone. But here we are; we’re living our space-age dreams from childhood (I’m still waiting for my flying car…) and we’re continually finding ways to redefine our known experiences. There is one industry that has lagged. One of our most powerful industries yet one that is often a lumbering beast with minimal agility: healthcare.

Of course, no one is suggesting that the healthcare industry is still stuck in the days of bloodletting and lobotomies, we’ve managed to adopt many technological advances that have improved patient outcomes, patient safety, and clinical delivery models. We’ve (reluctantly) accepted that EMRs are here to stay and that digital imaging systems really are far better than looking at x-ray films in a crowded hallway. Yet somehow the healthcare industry has been slow to embrace the idea of telemedicine as a modality that can be both high-tech and high-touch.

High-tech healthcare doesn’t have to sacrifice high-touch relationships

There is often a collective gasp when I tell people that the future of telemedicine does not lie in better video capabilities but instead in the virtual space, without cameras. True innovation for telemedicine will mean fewer “face to face” meetings with providers for common medical concerns (think: UTI, Conjunctivitis, URI, etc.) and far more asynchronous interviews in which a condition can be diagnosed and treatment can be prescribed using adaptive interviews and branching logic. The technology will aid in clinical decision making, especially in urgent care and primary care settings. Of course, I’m not suggesting that patients with serious, rare, or life-threatening conditions would benefit from a “virtual visit” but a patient that gets a UTI once a year and simply needs a routine course of antibiotics would benefit greatly. She gets her diagnosis, prescription, and be on her way within minutes without having to leave her home or make an appointment.

The technology is available and as healthcare becomes less authoritarian and more consumer-driven, we must accept that patients will expect and seek out options that allow for quick, uncomplicated diagnosis and treatment for routine medical issues. By embracing adaptive interviewing technology in the virtual space as a complement to current clinical practice versus a supplement, we are signaling to patients that we understand their needs and we respect their time. Gone are the days when patients will keep the same provider for 30 years; patients are now more focused than ever on efficiency, cost, and convenience. I’ve worked with several health systems who have struggled with patient retention and loyalty and the million-dollar question was always “how do we get them to come in?” The answer lies in providing the patients with enough convenience options for non-acute, routine conditions that they are willing to use the same system when they have serious medical needs. Consider the retail industry; many brick and mortar stores are dying on the vine while virtual retailers are blossoming. Why? Because virtual retailers filled a need for customers and in turn, those customers now buy nearly everything online (think: Amazon). According to Pew Research, 79 percent of Americans make purchases online, up from 22 percent just 18 years ago. It doesn’t mean that customers have deviated from brand loyalty, they just buy their products virtually. Imagine if in the next 20 years nearly 80 percent of patients received care online? Is your health system ready for that level of demand?

It is time to accept the fact that the next generation of patients are savvy and although they want relationships, they want relationships on their own terms. By offering user-friendly, cost-effective virtual care options, the health system is demonstrating an understanding that high-tech options meet a specific need for patients. And by meeting that need, you are in turn demonstrating that you care about their needs thus creating a relationship (even if the patient has never set foot in your building) that will build brand loyalty in the future.

High-tech healthcare doesn’t have to sacrifice high-touch relationships. Both things can co-exist and provide the patient with meaningful experiences that allow them to be equal partners in their health care. By accepting that medicine will not always be a face-to-face experience and leveraging technology to provide the patient with virtual care, we are moving with the times instead of fighting against the future. Think about it: had Blockbuster saw the writing on the wall and bought Netflix when they had the chance, things would have turned out very differently for our now nostalgic video store. As healthcare leaders, it is up to us to push the envelope and ensure that traditional medicine welcomes telemedicine warmly so we can continue to take care of our most important charge, our patients.

Note: The opinions expressed in this article are those of the author and do not necessarily reflect those of Evolent Health.

The post Getting High-Tech to Remain High-Touch appeared first on HealthTech Magazines.

Medical records are thought to contain some of the most sensitive and in-demand data. These records include information the patient provides, medical history, examination results, diagnoses, and treatment plans–all of which can instantly alter a provider’s patient care plan if made available. Yet, in today’s technologically advanced hospital environment, providers are still facing obstacles when retrieving patient medical records.

The way we manage medical records, whether advantageous or not, is transforming within Health Information Management (HIM) at a rapid pace. Evolving from traditional copy service through disclosure management and release of information (ROI), the technology enabled future of ROI is at the foundation of clinical revenue integrity.

As the healthcare industry grows more specialized and complex, challenges that come with managing health information increase in turn. The complicated realm of state and federal laws governing medical records means that compliance requires ongoing research to avoid accidentally stepping outside the bounds of the law. Inconsistent and, often at times, antiquated record keeping systems dispersed across multiple providers are hard to manage, and actually obtaining relevant data from these records can take time.

Every time a patient receives health care, a record is maintained of observations, medical interventions, and treatment outcomes. While medical records supply providers with relevant, real world data, the process of abstracting those records is riddled with potential pitfalls that can lead to delays and ineffective data retrieval.

When retrieval of data is delayed, it not only impacts real patients but also the provider bottom line, as the need for clinical information is increasingly part of the broader revenue cycle. Despite these obstacles, though, the future of ROI can be bright if providers effectively implement technology and leverage the process of clinical revenue integrity. A major advantage of technology-enabled ROI are the additional insights and analytics offered to the clinical revenue integrity process. ROI has always been thought of as beneficial mostly for HIM departments, but it truly sits at the convergence of clinical data and financial performance.

The more advanced ROI technology allows for constant patient feedback for specific points of improvement

Hazards can easily be maneuvered around using standardization technology, allowing health records to more easily produce actionable insights to use in the operational performance of ROI workflows as well as unique insights into patient and population-centric health data, leading to improved patient quality care outcomes. Once they have organized the multidisciplinary histories of care, providers can use them to learn about progression of chronic disease, more rapidly identify best practices for plans of care, and extract information that’s nearly impossible to quantify without technology. Health systems can identify and track network leakage across patient groups as well as develop programs to support high-risk revenue streams.

Providers can also maintain a better understanding of where requests are coming from and what for, allowing them to track volume trends and time-sensitive requests. This can be critical, as time-sensitive health plan based requests continue to grow. Healthcare providers are being strained by growth in these critical health plan requests, routinely 30 percent year-over-year, with some experiencing even higher demand.

Patients themselves also appreciate the ease that technology brings to securely and safely manage their own medical record. Given medical information may be the most important personal information to a patient – real-time, self-service tracking means they’re always confident their request is being addressed. With detailed reporting available, requesters can avoid having to follow up with phone calls and creating additional requests. Electronic request and delivery also allow patients and other requesters to harness technology, saving them from having to deal with manual requests, cumbersome fax inquiries, hold times and call backs to providers, or making a special trip for an in-person request. And, when special handling or authorization is needed, the more advanced ROI technology allows for constant patient feedback for specific points of improvement.

No matter what kind of requests are being dealt with, managing compliance and health information laws is complicated and ever changing. Technology is helping navigate these challenges and given the right tools, appropriate security levels, encryption, and continued assurance that users handling data have authorized access to it can be accomplished.

Changes at this scale require significant time and energy, not to mention the additional effort it takes to earn stakeholder buy-in. The road to full digitization can be long, but ultimately, it brings major quality improvements for providers and patients. According to a study by the Commonwealth Fund, administrative costs accounted for 25 percent–or more than $200 billion–of total hospital spending in the United States. Providers have the opportunity to reshape several of these high-value processes by integrating a scalable ROI technology platform into their clinical revenue integrity strategy, and the change needs to begin sooner, rather than later.

The post How Technology Enabled ROI is Transforming Clinical Revenue Integrity appeared first on HealthTech Magazines.

None of us is lacking in data. It’s coming at us from every angle, all day long. In fact, 30 percent of the world’s stored data is generated by the healthcare industry, according to a 2012 Ponemon Institute survey. But it’s not doing much good if we can’t put it to use. The challenge for those of us in health care, whether we’re providing care or managing reimbursements, is finding practical ways to turn all those undiscovered insights into workflow changes that can help us improve quality while saving money.

I first worked on an enterprise data warehouse at a health system in the mid-’90s. Back then, I thought it was only a matter of time before everyone – clinicians, administrators, actuaries and technologists – would have better information at their fingertips, allowing them to make more informed decisions. More than 20 years later, we have made progress in health care, but we still lag other industries in finding smart ways to bring together siloed information and infuse insights into our workflow.

If you’re frustrated by your organization’s lack of speed or effectiveness at making data actionable, I’ve observed a few key characteristics among those who are consistently getting it right:

1. Look at the tools, but not only the tools. Whenever I attend a big trade show like HIMSS, I’ll walk by hundreds of exhibitor booths, and at least half say they can provide “analytics.” In many cases, they are selling packaged analytics tools – many of which are not designed for the complexities of health care. As necessary as those tools are, they’re only a part of the equation. If your IT department is buying business intelligence tools and not considering how the insights they generate will be used throughout the enterprise, then you’re not getting the full picture. Focus on developing a data-driven culture, where the data science is embedded within and throughout the business.
2. Remember people still matter. Data and insights are only as useful as the people who are acting upon them. Here’s a common complaint I hear from C-Suite IT leaders: “We have so much information, but it’s not telling me anything I didn’t already know, and no one is doing anything with it.” If the information isn’t being shared on an enterprise level with those who are empowered to be changemakers, then it’s no surprise the data is just sitting there. Without people to interpret results and implement change, it’s not adding value. Relentlessly recruit, train, and retain talented employees – not just data scientists, but also people throughout the business who can interpret the data. That old business adage about the importance of people, process and technology is still true today. Scattered information can help create a cohesive care model if it’s linked and organized in a usable way.
3. Focus on the questions you’re not asking. There are industry leaders who are using more advanced techniques like advanced predictive modeling, machine learning and other applications of artificial intelligence. These tools differ from traditional analytics because you no longer need to know the question to get an answer – which has major potential impact for care delivery organizations and health plans. The difficult part is finding the skilled data scientists to optimize these capabilities for your organization’s needs.

We all know that hospital readmissions are upsetting to patients and drive unnecessary costs. To tackle issues like readmissions, our team of 26,000 multi-disciplinary professionals focuses on infusing healthcare intelligence into the solutions we offer. The healthcare intelligence comes from a unique combination of data, analytics and applied expertise, and we call this OptumIQ. This approach allows our clients to uncover data patterns that help identify people who are most in danger of readmission. According to an Optum analysis, hospitals and health systems often look to obvious sources of readmission yet miss 39 to 62 percent of the burden of illness. Having access to integrated electronic health record and adjudicated claims data as well as the tools and talent to extract answers without specific questions can help an organization tailor their outreach to the specific needs of a subset of patients. This data analysis can lead to systemic changes like instituting interventional programs where nurses visit patients’ homes to check on their medications, take vital statistics and assess socially related health factors. It’s interesting to note these nursing programs are often underwritten by health plans, not health systems because they have shown to be such a key factor in improving quality and lowering costs.

4. Check your bias. You know who your changemakers are. But if the insights they need are buried deep in the IT department then business and clinical change can’t happen. Your organization must operate with a bias toward action, ensuring the integration and empowerment of key players who can create change in clinical, financial and operational sectors. One way to foster this bias and avoid drowning in data is to start by identifying top use cases, prioritizing where analytics can provide the greatest business value. Determine the specific data sources and data practices needed to deliver on those priorities.As you’re trying to foster one type of bias, you may want to curb another. If you’ve been operating with a “do-it-yourself” bias and haven’t seen the results you want, you might benefit from other organizations’ best practices. Is it time to move toward more professionally managed analytics, working with external consultants or technologists to help advance how you use data and advanced technology? Only you can answer that question, but it’s an important consideration for the future. Just be sure to find a partner who sees the full picture and applies technology from a perspective that encompasses every sector of the industry.
5. Demand an ROI for everything, especially technology. If you’re embedding new technology or making investments, especially one with advanced analytics, don’t forget to establish clear and measurable goals. If a technology-enabled program isn’t aimed at providing a tenfold return on your investment, you shouldn’t consider it. You’ll need to measure frequently, adding new actions to the mix when you begin to see progress against your first set of metrics. AI should always have an ROI.
6. Look ahead. As you capitalize on current information sources, plan for the data you’ll need going forward. Some important information you’ll use in the future will likely be socio-demographic data, like social determinants of health, and consumer preferences and behavior – so get ready today.

Following these six tenets may not always be easy, but they could help you turn current pain points into future gains – informing better decisions and discovering new solutions to long-standing challenges. If we can learn how to harness the power of data and insights as a tool for action and build those insights into the very fabric of the enterprise, we can find transformative ways to solve today’s most challenging healthcare problems.

The post Drowning in a Sea of Data? Six Ways to Translate Insight into Action appeared first on HealthTech Magazines.

I have an Aunt, who is 86 years old and provides aide services 4- 5 hours a day to a 94-year-old woman living in her neighborhood. The 94-year-old, who I will call Ethel, is what my Aunt describes as “feeble”; she walks around the apartment with a cane, needs assistance with meal preparation and other activities of daily living. However, Ethel is alert and oriented, or as my Aunt describes it “totally with it”.

Ethel has no children but is in touch with a niece, who lives a distance away but calls frequently and visits monthly. Ethel has communicated to my Aunt that she is financially secure and has not hired “a girl” because she does not really need one and prefers to pay my Aunt “to visit” instead.

Ethel’s niece, in order to give herself peace of mind, insisted that her Aunt accept an “alert pendant”. Ethel’s niece explained that it was very easy to use, and if there was a problem Ethel could hit the button and help would come. Since her niece was paying, Ethel did not want to see the niece’s money wasted so she religiously wears the alert pendant, never removing it.

If you are rolling out technology/Telehealth commit to having several modalities

One night, Ethel gets up to go to the bathroom and falls. It is 2 am, and Ethel cannot get up. Ethel crawls from her bathroom into her living room and asks “Alexa” to call both my Aunt and her niece. When my Aunt arrived and helped Ethel up from the floor she asked Ethel why did not hit the alert button around her neck. Ethel replied that the button would result in a big commotion, and she knew that “Alexa”, which Ethel ordered for herself on her computer, would not make a big deal out of the incident.

Technology and elder care still struggle with synergy, and Ethel’s story resonates with me as part of our hurdle. I had been a Home Care Nurse for over 25 years and had implemented different modalities of Telehealth with the Medicare population, the Medicaid population as well as private pay. In that time frame there, I learned few lessons that I am more than happy to share.

Lesson #1:  There is not one modality that will connect with all your elders.  We know that there is a multitude of technologies to help you achieve your health goals: For example, Apps, wearable devices, and exercise webinars. Yet for some reason, at a certain age, we seem to think one modality will work for everyone. Once you hit 75, everyone should get an “alert pendant” for example. That modality works very well for some, not for the Ethels of the world. If you are rolling out technology/Telehealth commit to having several modalities. Some people will be very receptive to Medication Dispensers, others to Interactive Monitors, while rest to Applications. Trying to oversimplify will limit adaption.

Lesson #2: People, and this includes older people, do not always do things because it is good for them. Otherwise, we would all be exercising a minimum of 30 minutes a day. Telling an older person to use a technology because it will avoid a hospitalization or will keep them healthy is not always the right motivation. It is imperative to find out what the older person’s goals are in order to impact on utilization. Some people will say “I would like a reminder for my medications”, others will say “I would like a call if I am not feeling well”. Determining the elder’s motivation will have a significant impact on utilization.

Lesson #3: No one likes to have their decision-making ability limited, and older people are particularly sensitive to that danger. Ethel’s niece was right on target that Ethel needed to be connected to avoid being stranded in an emergency. But 911 and an ER trip was clearly not Ethel’s goal. If Ethel had not had Alexa there is a high probability she would have waited for my Aunt to arrive at her designated time versus hitting the alert button. Now, alert pendants can be utilized to function just like “Alexa”, but that needs to be established in advance. Older people, like all of us, want to make sure they control their own destiny. If the technology is perceived as “reporting” and then “resulting in undesirable consequences”, the older person will not use it. Utilization will increase if the older person is assured that an alert will trigger a call to them, to discuss next steps. I had a Telehealth patient who shared that the unit would have to go if her daughter had access to her data. The patient already knew she had to lose weight but having the daughter call after each weighs in was driving the patient insane.

Lesson #4: Avoid data overload, for both you and your patients. There are amazing technologies today: sensors that alert you if someone has not opened their refrigerator or moved off the couch all day, wearable devices that will send all sorts of biometric data, interactive devices that can identify a multitude of early decompensation symptoms. The risk is not to overload the user and your own backend with data. Just because you can get the data does not mean you want it. Identify what your global goal is and choose your data points carefully. The average Telehealth user will spend 8-10 minutes using the device. That is it, so be cognizant of the time required when you are designing your program. Older people have lives too and do not want to “feel sick” by having to use a unit every day for an extended period of time. Then you have to decide what your team will focus on as a “red alert”, something that requires follow up. If you “Red Alert” everything, you will wind up responding to nothing, and the first missed crisis will result in a non-user ever after.

Designing our technology programs with the goals of the older person in mind is an important first step.

The post Need of Telemedicine: Know the Healthseeker appeared first on HealthTech Magazines.