NextHealth Technologies

JP Morgan conference san francisco

Insights from the 36th JP Morgan Healthcare Conference

Healthcare executives need to know what works to make better, faster decisions.

Companies such as Amazon, Boeing, Microsoft, Netflix and Capital One run tens of thousands of experiments each year to pinpoint which products and programs work. They have to because they can’t afford to waste time and money on investments that don’t deliver revenue or cost savings. Collectively they’ve uncovered hundreds of millions of dollars of value by making decisions faster about what works and what’s a waste of time. And with Amazon’s recent entry into the healthcare space, health plans have no choice but to build effective test and learn capabilities and strong cultures of experimentation. Data-driven disruption just came from something looming on the horizon to knock on the front door.

Can you confidently and quickly isolate which of your programs are driving impact so you can optimize them ?

That was one of the key questions on the minds of healthcare executives at the 36th Annual JP Morgan Healthcare Conference that took place January 8-11, 2018 in San Francisco, CA.

NextHealth Technologies attended the conference and co-hosted a dinner with Deloitte and Norwest Venture Partners for national and regional plan executives to discuss affordability and how plans can tackle the unsustainable cost environment.

Two key themes emerged from the conference and dinner event:

1. Urgent need to know what works

Attendees at the dinner agreed there is an urgent need to better know what is working when it comes to reducing costs and engaging members. Yet plan executives and investors polled at the conference felt that health plans in general are struggling to prove ROI of all their investments. Which ones are worth the effort and which are a waste of time?

Plans are trying desperately to halt the tsunami of rising costs and in the rush to try many things at once, may actually be driving up costs by investing in programs that don’t work. The net result is an environment that fails to generate the critical insights about which investments work, which investments are a waste of precious resources and why.

By running tests to prove what works and what doesn’t you’ll widen your band of confidence to make better decisions, faster.

2. Differing levels of analytics maturity

When it comes to data and analytics – a key focus for healthcare executives in 2018 – there were widely diverse levels of maturity.  Some attendees and plan executives are working on data cleanliness and integrity while others have already invested “into the nine figures” in toolsets and programs aimed at tackling the affordability challenge.

Even if they don’t have perfect data today, plans agree there’s a pressing need to start using the data they currently have to move towards a test and learn culture and focus on knowing what works. The stakes are too high to not know what works.

Start today to make ROI and measurement part of your culture while you continue to take concrete steps toward cleaner data and personalized analytics.

NextHealth’s Perspective

  • Unlock transformative value by making better decisions faster

The current environment of trying everything in hopes that something works is unsustainable. The data and analytics horsepower currently exists so that healthplans can affordably prove which interventions and programs drive down costs within a matter of weeks. Solutions such as NextHealth easily augment and amplify current analytics programs to isolate and increase what works faster.

  • Create a culture of analytics

Adopt a posture of measure, know and optimize. Plan executives need to measure their efforts to know what works and then optimize their investments to do more of what they know works. It all starts by creating a culture of experimentation and analytics. It doesn’t matter where you are on the maturity curve – what matters is that you start testing and start learning. “By running tests to prove what works and what doesn’t,” said NextHealth CEO Eric Grossman, “you’ll widen your band of confidence to make better decisions, faster.”

It doesn’t matter where you are on the maturity curve – what matters is that you start testing and start learning.

  • Start wherever you are

Attendees at the conference and the roundtable dinner represented multiple points along the spectrum of data and analytics maturity. Some had already invested millions of dollars in programs and tests around consumer engagement. Some were still focused on cleaning up their data warehouses to ensure they had good data for personalization. Start where you are and move toward formalizing tests, learn from the results to know what works, do more of what works and scale those programs for cost saving across the entire business.

  • Optimize for what you know works

The crux of the opportunity is less about data quality debates and more about discussions of how to build a culture that is willing to make decisions and optimize based on what works. A true measure of a culture of innovation is the ability to drive value from what works and being unafraid to turn off what doesn’t and try again.

Once you have good data to measure what works and know what is effective then it’s time to optimize for what works. By now many people are familiar with the Facebook/ Snapchat rivalry where Facebook copied the popular Stories feature of Snapchat soon after parent company Snap went public. The net result was and continues to be devastating competitive pressure for Snap.

What many people don’t realize is that Facebook began it’s assault on Snap in France with a small test. The company rolled out a test of the Stories feature, proved that it worked and then started to scale to the larger Facebook platform. But even then it wasn’t a sure thing. It took multiple (failed) iterations to figure out the right combination of features and placement before the feature won widespread adoption.

The company didn’t stop running tests and trying new combinations until it found the one that worked and could prove it with test data.

NextHealth provides the scientific rigor and analytical insight to help plan executives know what works to reduce costs at scale.

See how the NextHealth analytics solution reduced avoidable emergency room visits by 25%.


Finding the Signal in Difficult Data

Customizing Algorithms to Drive Better Outcomes

By Dr. Douglas Popken,
SVP of Analytics

Doug-Popken-e1455057483612

Changing consumer behavior is complicated. Many different factors can impact human health-related behavior, both systematically (weather, changing regulations, general economic conditions, cultural influences, etc), and through individually based random variations (health status, mood, prior-held beliefs, financial circumstances, etc). Healthcare marketing teams craft messages targeted at consumers to try to influence consumer behavior, but even when they see impacts, they often cannot pinpoint whether it was their messages or perhaps another factor that caused the change. It then becomes difficult, if not impossible, to know how to efficiently allocate intervention resources.

At NextHealth, we designed our automated platform and methodology to address this critical issue. NextLift, a major component of the platform, includes a module that calculates both the lift and the statistical significance of each campaign or “nudge” impact on consumer behavior.  By using randomized controlled trials (RCT’s) we can isolate whether the change is due to the nudge itself or some other systematic factor.  By considering significance, we can also judge whether the measured lift is simply a reflection of random variations in the data.

In statistics, one way to distinguish an underlying effect from random variation is to use a measure called “p-value”.  P-value is a probability – the probability of observing a lift value more extreme than what was observed if the “null hypothesis” (that the true lift is equal to zero) were true.  P-value provides a scientific basis for claiming causality.  To claim success, the p-value needs to be small; at NextHealth, we use as a baseline statistical standard that if p <= .05, then the observed lift can be attributed to the nudge.

Lift is measured via a statistical “comparison of means” test, where the mean KPI (Key Performance Indicator) for the trial group is compared to that of the control group.  The standard statistical approach for determining the significance (and confidence intervals) of a comparison of means test is some form of “two-sample t-test”.  The drawback to unmodified use of these traditional techniques is that they require assumptions about the nature of the underlying data, especially, that the data follows a normal distribution.  However, the medical utilization and cost data typically encountered with our clients is often highly skewed with both a long right-hand tail and many zero values.  On the other hand, the two-sample t-test is known to be highly robust to non-normality if the data sets are large enough and/or the data is not too severely non-normal.  In practice, it is difficult to know when these conditions have been met.  For these reasons, NHT now uses a modern, robust methodology known as the bootstrap technique (Efron and Tibshirani, 1993) for determining the p-value and confidence intervals for the lift value.

The bootstrap method is based on repeated resampling to simulate comparison outcomes, with samples drawn from an empirical distribution of the observed trial and control data.  Its key advantage is that it requires no assumptions about normality and is therefore highly appropriated for skewed medical data.  It reduces the motivation to transform the data before analysis to make it less non-normal (e.g log transformations or trimmed mean approaches), allowing for direct comparison of the true means of the two groups.  To compute p, we use bootstrap samples of the t-statistic (each of which is computed as described below) for comparison of the population means of the two groups.  A variance stabilization technique is automatically applied to achieve the highest accuracy.  To compute confidence intervals on the lift, we use a bootstrap distribution of mean lift values with bias and skewness corrections known as BCa.  The specific techniques we use are described in greater detail in Barber and Thompson (2000).  See also Efron (1987).

To compute the t-statistic for each bootstrap sample, NextHealth uses a modification of a well-known statistical test for comparing the means of two populations with unequal variances, “Welch’s t-test” (see http://itl.nist.gov/div898/handbook/eda/section3/eda353.htm  or https://en.wikipedia.org/wiki/Welch%27s_t_test).  Other methodologies are available, but research has shown that Welch’s t-test generally provides the most accurate results for the type of data we typically work with (Fagerland and Sandvik, 2009).  The modification NextHealth has made to the standard Welch’s t-test is to weight the observations in each group by the duration of the observation period (in years).  One immediate advantage is that the duration weighted mean is equivalent to the population mean KPI per member year (total KPI value/total member years), which is the most relevant statistic for our analyses.  Another motivation for weighting is that our underlying observations are annualized, but the observations themselves have variable observation periods.  Without weighting, a non-zero observation resulting from a short observation period can become relatively large when annualized, thereby causing the dataset to have unusually high variances, particularly in the early stages of a campaign/program.  A similar weighting approach is described in Bland and Kerry (2008) for a t-test that assumed a single pooled variance for the samples.  To compute the weighted means and weighted variance of each group, we rely on equations described at https://en.wikipedia.org/wiki/Weighted_arithmetic_mean, in the subsection, “Weighted Sample Variance…Reliability Weights”.   The weighted means and weighted variances computed from these equations then replace the unweighted mean and variance parameters described within the Welch’s t-test.

Measuring program success in the face of skewed outcome data with a high degree of random variation is difficult.  NextHealth has employed a combination of the best statistical methodologies available to achieve the highest degree of accuracy.

References

  • Barber, J.A. and Thompson, S.G.  Analysis of cost data in randomized trials: an application of the non-parametric bootstrap.  Statistics in Medicine, 19, 3219-3236.
  • Bland, J.M. and S. Kerry.  2008.  Weighted comparison of means.  BMJ, 316, 129.
  • Efron, B.  1987.  Better bootstrap confidence intervals (with comments).  Journal of the American Statistical Association, 82(397), 171-200.
  • Efron, B. and Tibshirani, R.J.  1993.  An Introduction to the Bootstrap.  Chapman and Hall, New York
  • Fagerland, M.W. and L. Sandvik.  2009.  Performance of five two-sample location tests for skewed distributions with unequal variance.  Contemporary Clinical Trials, 30, 490-496.


Product Analyst

Careers

Job Title:

Product Analyst

CAREERS

Description

NextHealth Technologies enables health plans to reduce medical and administrative costs with a secure, scalable prescriptive analytics and consumer engagement platform.  We are seeking a  Product Owner to be a member of an agile product team and serve as the voice of the customer. This role reports to the SVP of Product and is responsible for working with internal and external teams to define and prioritize the product backlog so that the solution effectively addresses program priorities while maintaining technical integrity. The product owner also attends most relevant product management meetings, planning, and backlog/vision refinement sessions to align project execution with product vision.

Responsibilities

  • Backlog Refinement. With input from the product manager and other stakeholders, the product owner has the primary responsibility to build, prune, and maintain the team backlog. Backlog items are prioritized based on user value, and time and other team dependencies which are determined in Release Planning.
  •  Iteration Planning. The product owner reviews and re-prioritizes the backlog as part of the preparatory work for Iteration Planning, including coordination of content dependencies with other product owners. During the iteration planning meeting, the product owner is the main source for user story detail and priorities and has the responsibility to accept the final iteration plan.
  • Feature Design. The product owner is responsible for collaborating with internal teams to scope and design new features. This involves understanding technical tradeoffs, mapping dependencies, incorporating user feedback, and producing artifacts such as wireframes and sequence diagrams.
  • Writing User Stories & Acceptance Criteria. The product owner must be able to distill high-level features into small, incremental pieces of development work in the form of user stories with detailed acceptance criteria. The product owner is responsible for providing additional detail and clarification to stories as needed, to keep the development process running smoothly.
  • Accepting stories into the baseline. The product owner is the only team member who can accept stories into the baseline. This includes validation that the story meets acceptance criteria and that each has the appropriate, persistent acceptance tests, and otherwise meets its definition of done. In so doing, the product owner also fulfills a quality assurance function, focusing primarily on fitness for use.
  • Participating in Team Demo and Retro. As an integral member of the team, and the one responsible for requirements, the product owner has an important role in the sprint showcase/product demo, reviewing and accepting stories in the baseline, and in the iteration retrospective, where the teams gather to improve their processes.

Qualifications

Required 

  • Agile/Scrum experience
  • 2 or more years of experience delivering commercial software products
  • Knowledge of market conditions to develop new ideas based on industry experience and contact with customers and prospects
  • Inquisitive and innovative mindset with a demonstrated ability to recognize opportunities to create distinctive value that are not evident to most others
  • Demonstrated ability to create and manage complex development projects in a highly dynamic environment
  • Demonstrated ability to pull together diverse teams to solve problems and pursue opportunities
  • Demonstrated ability to provide leadership and influence in a decentralized organizational structure
  • Strong interpersonal skills; demonstrated ability to build trust and strong relationships
  • Strong conceptual, analytical and strategic thinking skills
  • Strong communication and presentation skills
  • Knowledgeable in technology
  • Ability to manage several projects concurrently
  • Focuses on the details of solving each business problem/need correctly

Preferred  

  • Bachelor’s degree in Computer Science, Business or related field
  •  Previous experience with healthcare or analytics products
  •  Agile certification (Scrum Master or CSPO)
  • UX design experience including interaction design, information architecture, and usability testing
  • Experience delivering products using Application Life Cycle Management (ALM) with an Agile Process


Test and Learn Your Way to Medical Cost Savings

To be truly innovative you need to create a culture of experimentation.

Have you ever bought anything from Amazon zShops? Probably not.

How about Amazon Auctions? Not likely.

You are not alone because these two massive experiments were quickly scuttled by the retailing giant 15 years ago on the road to hitting their home run with Amazon Marketplace where nearly 50% of current units sold on Amazon.com are from third-party sellers. zShops and Auctions were not failures; rather, the company had to test these two concepts in order to learn how to successfully execute the Marketplace offering. A test and learn culture is critical for innovation not just in retail but in every industry, including healthcare.  

“To invent you have to experiment, and if you know in advance that it’s going to work, it’s not an experiment,” says Amazon’s CEO Jeff Bezos in his 2015 shareholder letter. “Most large organizations embrace the idea of invention, but are not willing to suffer the string of failed experiments necessary to get there.”

“Most large organizations embrace the idea of invention, but are not willing to suffer the string of failed experiments necessary to get there.”  – Amazon CEO Jeff Bezos

NextHealth CEO Eric Grossman recently wrote about why it’s important that healthcare leaders transform their organizations into test and learn cultures in his recap of the 2017 Oliver Wyman Health Innovation Summit. There are important lessons in experimentation and cost savings for healthcare industry executives from other industries including retail and financial services.

In a recent discussion during the “Finance Disrupted” event by The Economist, Capital One co-founder and QED Investors Managing Partner Nigel Morris detailed how the competitive advantage for companies today lies not just in having lots of data to run experiments but rather how companies leverage that data for insights by building a test and learn culture. Morris explained it best in his outline of Capital One’s strategy:

“The core idea was that the credit card business is not really the traditional lending business, it’s not really banking at all. What it really is is the leveraging of information in order to be able to put the right product in the right customer’s hands at the right time. The way you did that was in two or three ways. One was amassing huge amounts of data at the customer level and then using experimental design and testing different product and market combinations in order to optimize value to the consumer and net present value to the entity.”

Sound familiar?

You could replace “credit card business” in the above quote with “healthcare industry” and be accurate in describing the opportunity for health plans to build test and learn cultures that better understand and align the right products for the right members at the right time. Capital One now runs over 80,000 tests per year to pinpoint the ideal combination of product, customer, and timing.

For Amazon, Capital One, and a growing number of health plans, when it comes to reducing costs, growing revenues, and improving outcomes, the insights from the tests and how (and how quickly) they are applied are what matter most. Through thousands of experiments, Amazon learned that customers are already using its artificial intelligence framework for early disease detection that saves lives and lowers costs. Through testing and learning, CapitalOne found that people who complete an application in all capital letters pose a higher credit risk (i.e. higher costs). 

Here is how Jeff Bezos described a test and learn approach for Prime Now that was put in place in a matter of months:

“Prime Now offers members one-hour delivery on an important subset of selection, and was launched only 111 days after it was dreamed up. In that time, a small team built a customer-facing app, secured a location for an urban warehouse, determined which 25,000 items to sell, got those items stocked, recruited and onboarded new staff, tested, iterated, designed new software for internal use – both a warehouse management system and a driver-facing app – and launched in time for the holidays. Today, just 15 months after that first city launch, Prime Now is serving members in more than 30 cities around the world.”

How does NextHealth help its health plan customers test, learn and optimize outcomes quickly?

NextHealth’s Measurement and Optimization platform automates health plans’ ability to quickly test, measure and optimize any clinical program. Key benefits include:

  • Faster, data-backed insights allow teams to get to better business decisions quickly and with greater consensus

  • Standardized and consistent analytics methodologies and program set-up help elevate the organization to a shared conversation based on experimentation, building a culture of measurement

  • Automating the ability to test, learn and optimize allows talented analytics and healthcare economics resources to do more with less and operate at their highest capacity

The platform has enabled clients to measure existing programs in as little as 60 days, learning which programs are effective and which are not, gleaning administrative cost savings quickly. New programs, such as reducing avoidable ER utilization, EPDST, HEDIS improvements, and others, can be added to the platform as well, delivering medical cost savings using machine learning to optimize outcomes by assigning members to the programs that are most likely to work for them. NextHealth’s managed services offering, included in every engagement, ensures that clients have access to the learnings gleaned from our experience, furthering speed to value.

Is your organization reducing medical and administrative costs through a test and learn culture?

Our platform accelerates data-driven decisions and enables a culture of measurement. See how.


Know What Works - Make Data-Driven Decisions To Drive Down Medical Costs

Insights from the Oliver Wyman Health Innovation Summit 2017

Eric Grossman, CEO of NextHealth Technologies, recently attended the Oliver Wyman Health Innovation Summit, Nov 6-8 in Dallas, TX. Titled “Industry Interrupted: Delivering on the Promise of Change”. The event convened healthcare industry leaders to discuss disruption in healthcare. We recently sat down with Eric to highlight his thoughts on the event.

What were your three main takeaways from the event?

1. Create a test and learn culture

Being able to test and learn – at scale – is an underpinning of competing in an industry ripe for disruption. Nigel Morris from Capital One provided a great example how taking an experimental approach (i.e. a test and learn mindset) can position a company to not just weather disruption but come out ahead as a result.

Capital One conducted 46,000 test-and-learn experiments in the year 2000 to see which combinations of campaigns, products, messages and offers resonated most with consumers. The company experimented to find out what worked for each micro-segment and created an infrastructure to optimize what was working. They captured the data exhaust on everything they did so they could figure out what worked. They kept failing to find what was successful. Thomas Edison put it well when he noted: “Many of life’s failures are people who did not realize how close they were to success when they gave up.”

2. Support the supply chain

Health plans need to figure out how to derive value and align incentives in the supply chain. The physicians are saying “We have engagement fatigue. You’re calling us. You’re reaching out to us. But we only have so much time in a day. You have to help us prioritize where we can get the biggest impacts in our days.” The way we do that is through the use of analytics, and in particular, both machine learning and prescriptive analytics. Enabling physicians to be more efficient and maximize reimbursement is heavily dependent on analytics.

Prevent engagement fatigue by empowering physicians with better insights on how to maximize their return on time and focus. That will require developing this test and learn mindset.

3. Engage the connected consumer

It’s time to acknowledge and engage the connected consumer, especially the data exhaust from their digital, always-on life. Leveraging this data exhaust in an analytics environment built around the connected consumer is vital. As consumers take on more responsibility for cost, there is now a need for plans to do things other than just ingest claims and generate payments. They need to become a trusted advisor for consumers as they navigate the complexities of managing their own care and cost decisions. Health plans are trying and eager to adjust to this new reality, but to be effective, they need to have a faster, more reliable way to make data-driven decisions and an infrastructure that will help them innovate more effectively.

What’s the “landscape of disruption” look like for healthcare in 2018?

The threat of disruption is real for the healthcare industry. Look at how Amazon is getting into the pharmacy business and paid for the recent acquisition of Whole Foods with a day’s worth of capital market gains. Look at Uber, Square, etc.

Necessity is the mother of invention. We’ve got the right alignment of incentives now. The industry has to innovate because it’s threatened by disruption. At the same time, plans have the data and the membership. They have large enough populations to move the needle and also have the capital to invest in these areas to have real impact, build trusted relationships with their members, and avoid losing market share.

What’s different about the current environment?

Healthcare is unsustainable at its current cost. We (the industry) have been saying that for years so that’s not necessarily groundbreaking, but if you look at the geopolitical dynamics of healthcare you can see that things really are different now. Look at the difficulty Washington has had enacting change despite the enormous costs at stake. The unit costs are unsustainable and companies just can’t afford it anymore. The fact that commercial insurance completely subsidizes Medicare and Medicaid is just not sustainable. In addition, now that consumers are shouldering more of the risk/cost burden, they need help in learning how to better manage that responsibility. That shift creates either real opportunity or potential downfall, depending on how the industry reacts.

How do organizations create a test and learn culture?

First, without strong executive sponsorship and governance to drive a culture of measurement into the organization, the initiative will be doomed. The second priority is to drive standardization and processes around measurement that don’t rely on politics or influence, but rather rely on data and insights as the currency for decision-making. Lastly, creating a “test, learn, and optimize” environment relies on building comfort with a culture of failing a lot more and continuing to understand that failure should be seen as a means to an end instead of just an end.

How are you going to innovate if you don’t have an infrastructure that allows you to test and learn quickly? Remember, analytics is a no-regret investment.

Our platform accelerates data-driven decisions and enables a culture of measurement. See how.


NextHealth Technologies to Host Live Webinar on Measuring Program Effectiveness and Optimizing ROI

nexthealth_technologies_logo_clearspace 300


Wednesday, October 25, 2017 — NextHealth Technologies, a prescriptive analytics and consumer engagement platform that reduces medical costs for health plans, announced that it will host an exclusive webinar on Tuesday, October 31, from 3:00 PM to 4:00 PM ET. The webinar will be hosted by Anne Marie Aponte, SVP of Operations, and Dan Masciopinto, SVP of Product, and will focus on how health plans can drive attributable administrative and medical cost savings through program measurement and optimization.

“The ability to determine what is working for whom is a considerable challenge for health plans,” said Aponte. “At NextHealth, we built our platform to target specific member clusters and scientifically measure which efforts are working, how well, and for which members.” Added Masciopinto, “Plans can significantly amplify ROI by systematically optimizing resources and focusing on the programs that work and turning off the ones that don’t.”

The webinar will cover solutions to problems that health plans struggle with on a daily basis, including:

  • Measuring which interventions work for specific populations
  • Optimizing around channel effectiveness to reach certain members with the interventions that resonate
  • Allocating resources to the programs that are scientifically proven to have impact

For more information about the webinar or to register, visit the NextHealth website.

About NextHealth Technologies

NextHealth Technologies empowers people to make the best healthcare decisions. Our prescriptive analytics and consumer engagement platform has helped innovative health plans such as UnitedHealthcare, BlueCross BlueShield of Tennessee, and Florida Blue significantly reduce administrative and medical costs through better targeting, measurement, and optimization of member interventions. Scalable to multiple use cases and lines of business, plans know at a glance who to target, which programs work, which don’t, and precisely how well – all from a single dashboard. Outcomes are measured within 60 days from deployment and offer in-year ROI. NextHealth offers an optional managed services contract and puts its fees at risk based on delivered outcomes. For more information, visit nexthealthtechnologies.com.

Awards and Accolades

NextHealth was named in three Gartner industry reports: “Hype Cycle for U.S. Healthcare Payers, 2017” (in the Healthcare Consumer Insights as a Service category), “Hype Cycle for Healthcare Providers, 2017” (in the Healthcare Consumer Persuasion Analytics category) and “Hype Cycle for Consumer Engagement for Health and Wellness, 2017” (in the Healthcare Consumer Insights as a Service category). Frost & Sullivan also named NextHealth in its 2017 Global Healthcare Data Analytics Companies-to-Action report.

For media inquiries:

NextHealth Technologies
Melissa O’Connor
VP of Marketing
moconnor@nexthealthtechnologies.com


Richard Thaler’s Nobel Prize – Can Behavioral Economics Drive Down Healthcare Costs?

Behavioral economics is critical to understanding non-rational behavior. How can plans apply these concepts to reduce medical costs?

Earlier this month Richard H. Thaler was awarded the Nobel Prize in Economic Sciences for his contributions to behavioral economics. Thaler’s research has been a seminal step towards understanding the impact of non-rational decision making that produces a theoretically worse outcome for the individual decision maker. In layman’s terms: people frequently make decisions that are not in their own best interest – economic or otherwise. 

“This type of thinking – understanding humans as non-rational decision makers – is critical in addressing healthcare both from the viewpoint of economics and designing for better outcomes.”

The insights of non-rational behavior as viewed through an economic lens has been significant for other research and development in the areas of psychology and consumer behavior. There is broad application to many fields that anticipate good behavior based on self-interest but see individuals struggle to meet expected compliance, behavior change, choice or self-care objectives. This applies to wellness, self-care and behavioral compliance with care plans across a range of areas (from Physical Therapy to Pharmaceutical Prescriptions to Discharge Plans to Utilization Choices).

Translating Thaler’s work to healthcare

Thaler’s research moved economics beyond pure logic and into the realm of irrational action and attempts to understand why we are irrational actors.  This is critically important as his work has shed an important light on individual examples of non-rational behaviors, as well as broader trends that can be identified (or shaped) with society more broadly.  In the healthcare field, his work, and the work of other psychologists and consumer behavior researchers is contributing to change in the design paradigm for how we think and act on healthcare initiatives at the individual level (care plan, utilization) and for large health populations (wellness, disease management). This type of thinking – understanding humans as non-rational decision makers –  is critical in addressing healthcare both from the viewpoint of economics (can digital therapies be more effective than drugs in some cases?), and designing for better outcomes (what message is most effective to achieve a positive health/social outcome?).

How NextHealth uses nudges to change behavior and improve outcomes

NextHealth is translating many behavior design concepts to healthcare in areas initially advanced by Thaler, and later his peers in related fields of consumer behavior and psychology. Examples of behavioral nudges used by NextHealth clients include:

  1. Selection of frames (e.g., gain frame or loss frame for potential plan benefits; social frame emphasizing the choices of relevant others)
  2. Use of key words (e.g., free benefits)
  3. Appeals to mental accounting
  4. Active choices that encourage recipients of nudge messages to make decisions while they have relevant message information in front of them

Below is an example from an outbound call script using a selection of these behavioral nudges:

“This free program is designed to help you get answers and advice about <your/your child’s>health and where to get medical help quickly and safely with just a phone call, and can be used immediately.”

Read More: NextHealth helps client reduces avoidable ER visits by 25% by implementing nudges and targeted outreach.

Additionally, NextHealth is increasing its impact by elevating the rigor around two areas abutting program design; identification of impactable target populations and validation of program efficacy.

Extending Behavioral Economics with better targeting

Certain factors in healthcare invoke significant challenges for designing programs that impact behavior. Critical aspects of behavior design assume the ability to trigger consumer action based on a consumer context or opportunity; however, the timing and acuity of many health issues can be unpredictable, which conflicts with the best state- of-the-art intervention approaches.

“NextHealth uses machine learning based on health data to find signals that indicate a particular group is a highly impactable target and economically viable to address.”

Not all consumers behave irrationally, and some people are unaware of, or have dissonance to messages and information that creates the appearance of irrational behavior.  Therefore, the ability to identify populations that would significantly benefit from a program designed to change their behavior is critical. NextHealth uses machine learning based on health data to find signals that indicate a particular group is a highly impactable target and economically viable to address.

Thaler’s research sustains the notion that people can be induced to better decisions by re-framing a process, outcome or action to off-set what otherwise might be interpreted as irrational cognition. Refining targets using machine learning enhances our ability to profile population segments and dial-in more specific behavior design elements. The specificity of the consumer targets also ensures that we can define the correct cohort to generate a Random Clinical Trial in concert with any behavior design to validate the impact of the program.

Understanding who to address, how to frame and shape decision-making, and optimizing validated results are necessary to drive and impact behavior design for healthcare.

Although Thaler is an economist, and much of his work is focused on behavioral economics, his key findings serve to open many fields to the idea that we are not always making rational decisions. In healthcare this extends to the food we eat, the activities we choose or the choice to discontinue medication compliance as well as many other short and long-term life-impacting decisions.  

Today there is an open field of opportunity to contribute advancements in ethical behavior design that can shape economics, health and society in positive ways. We can thank Richard Thaler and researchers in the field of consumer behavior for enlightening us and moving us toward more sophisticated design strategies that take into account both rational and non-rational characteristics as we address health decision making going forward.

Targeting impactable members is critical to changing behavior and reducing medical costs. See how easy it can be.


Improve the Bottom Line by Reducing Low-Cost, Low-Value Services

Low-cost, low-value, high-volume health services contribute to 65% of unnecessary medical costs – here’s what health plans can do about it.

Each year, the U.S. spends nearly $1 trillion on unnecessary medical expenses. [1] A study published by the Harvard Business Review estimated that roughly 40% of unnecessary medical spending can be avoided by addressing the clinical waste, administrative complexity, excessive prices, fraud and abuse experienced in the system. Professional guidelines, personal judgment and health plan design make these types of waste difficult to address.

Is it possible to tackle low hanging fruit in the waste equation to chip away at the egregious cost for unnecessary medical care?

Absolutely.

Waste in U.S. Healthcare
Source: Harvard Business Review

Focus on low-value, low-cost services to start

A recent study published by Health Affairs highlights how health plans can reduce unnecessary costs by curbing low-value services. [2] Looking at the Virginia All Payer claims database for 2014, roughly $586M in unnecessary medical costs were attributed to low-value services across two categories:

  1. Low-value, low-cost services typically valued at or under $538
  2. Low-value, high-cost services typically valued over $538

Surprisingly, low-value, low-cost services made up 65% of the total unnecessary medical costs in the data set examined. [3]

Examples of low-value, low-cost services include:

  • Baseline lab tests for low-risk patients having low-risk surgery: Studies have shown that a good history, physical exam, followed by a review of a patient’s chart are sufficient for low-risk patients who are headed to get a low-risk surgery. [3, 4, 5]
  • Stress cardiac and other cardiac imaging in low-risk, asymptomatic patients: An article in the Cleveland Clinic Journal of Medicine indicated that low-risk patients that undergo unnecessary cardiac stress tests may be exposed to more risk through additional follow-up testing. [3, 4, 6]
  • Annual EKGs or other cardiac screening for low-risk asymptomatic patients: The American Academy of Family Physicians cited risks of false positives that often lead to unnecessary invasive procedures, overtreatment, and misdiagnosis for annual EKGs for low-risk patients. [3, 4, 7]

Why reducing low-cost, low-value services is difficult

Some of the challenges stem from the lack of health education and awareness that directly impacts people’s decision making abilities. Specifically, not many consumers make research-based decisions when it comes to their health. In fact, many consumers rely on their doctor’s recommendation to make key health decisions according to a McKinsey survey. [8] Therefore, if providers recommend a low-cost, low value service to a patient, they are more likely to comply – contributing to the $1 trillion in unnecessary medical expenses.

How to reduce costs

The first step focuses on targeting. Who in the plan membership is anticipated to get a low-cost, low-value service? Health plans can intervene by providing members with educational materials to help them understand the costs and benefits of getting these services so they can make informed decisions.

NextHealth enables just the type of targeting that helps health plans identify the right members to engage. Using claims and provider data, NextHealth identifies members who are predicted to get a low-cost, low-value service and intervene with a nudge that incorporates education and behavioral science theories so members make better health choices. The process is simple and can be a powerful way to tackle the low hanging fruit when it comes to reducing the $1 trillion in unnecessary medical expenses.

Contributed by Thomas Tran, Engagement Manager, NextHealth Technologies

Sources:

[1] How the U.S. Can Reduce Waste in Health Care Spending by $1 Trillion

[2] Low-Cost, High-Volume Health Services Contribute the Most to Unnecessary Health Spending

[3] Study: Unnecessary health spending fueled by low-cost, low-value services

[4] Low-Cost, Low-Value Healthcare Services Ripe for Reaping

[5] Perioperative Testing

[6] Is cardiac stress testing appropriate in asymptomatic adults at low risk?

[7] Annual EKGs for Low-Risk Patients

[8] Debunking common myths about healthcare consumerism

Targeting impactable members is critical to changing behavior and reducing medical costs. See how easy it can be.


Machine Learning in Healthcare: From Data to Prediction

Machine learning for more accurate predictions and more precise targeting

Predicting consumer behavior can be a complex process, especially in a healthcare setting. Fortunately, the abundance of data sources (such as claims data and clinical statistics), provide ample opportunity to generate meaningful insights. The advancement of machine learning algorithms has opened up even more opportunities to get ahead of complex problems and to predict future behavior.

At NextHealth, we have incorporated machine learning into our platform. For example, the platform can accurately predict emergency room visits by identifying health risks and utilization patterns within identified member populations. Backed by these predictions, the platform then targets at-risk and impactable member clusters, assigns them to the most effective intervention campaigns, and measures what works for whom.

“By integrating machine learning into our platform, we can transform health data into actionable intel and impactful results.”

Better Targeting of Populations

Once a client’s raw data has been standardized and processed, NextHealth builds the algorithms to maximize predictive power. Our data scientists mainly use tree-based learning algorithms due to their regression and classification capabilities, along with their scalability into both linear and nonlinear relationships.

In order to target high risk populations that offer the greatest opportunity for impacting a particular use case, we use Decision Trees (a flowchart-like structured algorithm) due to the intuitive visualization design and interpretation capabilities. As the decision trees are processed, we split the features based on the largest information gain. The output gives us the difference between the impurity of the parent nodes and the sum of the impurity of the child nodes. This splitting process is repeated at each child node until it reaches to the very bottom leaves. In order to avoid overfitting – a very deep tree with lots of nodes – we use a “prune” technique to limit the maximum depth of the tree.

An example from the platform: In this case, decision trees help us predict the population clusters that have the highest ER utilization in comparison to the total population and offer the highest potential for use case impact

Member-Level Risk Prediction

At the member level, we run our risk prediction by using Random Forest and Gradient Boosting.

Random Forest is an ensemble of decision trees and applies a Bagging technique to tree learners. Random Forest takes random samples of the training set (with replacement) and grows a decision tree from each sample. After many trees are formed, the prediction will be aggregated and decided on majority vote. The ensemble method is believed to be robust to noise.

Gradient Boosting is an additive model for including weak learners using a gradient descent procedure – an iterative method that takes steps proportional to the negative of the gradient of the function to find a local minimum of a function. Decision trees are used as the weak learner in Gradient Boosting. Trees are implemented one at a time, and a gradient descent procedure is used to minimize the loss when adding trees.

While Random Forest reduces variance, Gradient Boosting reduces bias. We use a stacking technique to combine the two algorithms for improved accuracy.

The Result: Better Predictions, Improved Targeting

At NextHealth, we deliver measurable outcomes for our health plan customers by 1) better targeting their most impactable members, 2) deploying the personalized and persistent interventions most likely to impact behavior, and 3) measuring and optimizing what works for whom. The machine learning methods we employ drive more reliable behavioral predictions and more accurate targeting of at-risk member populations. Ultimately, these processes enable health plans to get ahead of costly behaviors and deliver the most impactful interventions to the members who are most likely to be receptive. By integrating machine learning into our platform, our customers transform big data into actionable intel and impactful results.

By Cathy Zdravevski, Data Scientist, NextHealth Technologies

Machine learning helps us get ahead of many problems. Explore other use cases.


Reduced Risk, Scalable Value: The Build-Operate-Transfer Model

When addressing complex problems, the right deployment model can be just as important as the solution.

In a world of finite resources, limited time, and competing priorities, health plans must be extra diligent when investing in new solutions. Identifying the best solution, whether it’s new technology or a streamlined business process, is merely the first step. How the investment is deployed can differentiate failure from success. Organizations must deploy solutions in a way that mitigates short-term risk, does not drain organizational bandwidth, allots time to prove ROI, and sets up the organization to successfully scale solutions long-term.

Build-Operate-Transfer deployments reduce short-term investment, prove ROI, and empower teams for long-term, scalable growth.

Implementing NextHealth’s platform via our Build-Operate-Transfer (“BOT”) methodology delivers scalable operational and financial benefit, while achieving those multi-pronged objectives – making our solution a rare find in today’s world. While NextHealth customizes the BOT deployment timeline and approach for each client, the image below illustrates how this phased approach unfolds and translates into scalable value.

Process & Timeline

Months 0 – 3: Build

NextHealth’s Managed services team leads the implementation process, including key activities such as providing project management support for data exchange processes and establishing a core governance framework and operating rhythm. In conjunction with the client, NextHealth will also begin preparing the regional deployment of new programs designed to reduce the costs associated with a pre-determined use case, such as avoidable ER utilization.

Months 3 – 18: Operate

During this critical window, NextHealth leads program execution, including direct member engagement across all channels – including telephonic, digital, and print. As results unfold, NextHealth partners closely with the client to identify how the platform fits into the health plan’s workflows to amplify existing capabilities and/or fill critical gaps in the end-to-end orchestration of cost reduction programs.  End users are identified within the client organization and trained to serve as subject matter experts.

Months 18 & Beyond: Transfer

As the initial deployment of the solution delivers tangible, causal results, NextHealth supports the client in scaling the program across a broader population to amplify cost savings. This proof point also serves as the launch pad for the identification of additional use cases, leveraging the 30+ off-the-shelf KPIs built into the platform or custom-built measures developed in the platform by the end user. During this phase, member engagement activities are transferred to the client, enabling closely knit integration and improved operational efficiencies.

Advantages

The BOT deployment methodology is advantageous for customers in three primary ways, leading to increased net paid savings overtime:

  1. Risk Management: BOT enables testing of the NextHealth solution while minimizing both upfront resource requirements and financial risk. NextHealth does the upfront legwork associated with implementation and bears risk for results during the initial deployment.
  2. Optimization: BOT enables the opportunity for the client and NextHealth to partner together to identify the optimal way to integrate the platform into the health plan’s existing operations to ensure long term success.
  3. Scalability: BOT capitalizes on the scalability of the platform, leading to increased savings over time via the expansion of successful programs and deployment and testing of new use cases.

By Anne Marie Aponte, SVP of Operations, NextHealth Technologies

The NextHealth platform scales to many use cases. Explore the problems we solve.