Free EMR Newsletter Want to receive the latest news on EMR, Meaningful Use, ARRA and Healthcare IT sent straight to your email? Join thousands of healthcare pros who subscribe to EMR and EHR for FREE!

Self-Learning Analytics and Making Analytics Useful

Posted on April 2, 2018 I Written By

John Lynn is the Founder of the blog network which currently consists of 10 blogs containing over 8000 articles with John having written over 4000 of the articles himself. These EMR and Healthcare IT related articles have been viewed over 16 million times. John also manages Healthcare IT Central and Healthcare IT Today, the leading career Health IT job board and blog. John is co-founder of and John is highly involved in social media, and in addition to his blogs can also be found on Twitter: @techguy and @ehrandhit and LinkedIn.

One of the shocks to me at HIMSS 2018 was that there wasn’t nearly as much discussion around healthcare analytics as I thought there would be. I thought for sure we’d see an explosion of proven analytics that healthcare organizations could start to take advantage of. Maybe I just missed it, but I certainly didn’t see anything all that new.

It’s too bad because that’s one of the huge opportunities I see for healthcare. I was looking through some old notes from conferences and saw a note where I wrote: “What you do with the data is the competitive differentiator, not the data.

Certainly, you need access to the data to be successful, but there are a lot of organizations out there which have access to health data and they’re not making any sort of dent. Many of the now defunct HIEs had access to the data, but they didn’t know what to do with all that data. I’m still on the search for more analytics which are useful.

One other idea I found in my notes was the concept of a self-learning analytic. Related to this was the discussion we had about black box analytics in a recent #HITsm Twitter chat. I don’t think they have to be the same, but I do think that the key to successful healthcare analytics is going to require some component of self-learning.

The concept is simple. The analytic should look at its past recommendations and then based on the results of past recommendations, the analytic should adjust future recommendations. Notice that I still call it recommendations which I think is still the right approach for most analytics. This approach to constantly learning and evolving analytics is why it’s so hard to regulate healthcare analytics. It’s hard to regulate moving targets and a self-learning analytic needs to be moving to be most effective.

This is possibly why we haven’t seen an explosion of healthcare analytics. It’s hard to get them right and to prove their effectiveness. Plus, they need to continually evolve and improve. That’s the opposite of what researchers want to hear.

This is why the future of healthcare analytics is going to require deep collaboration between healthcare analytics vendors and provider organizations. It’s not a black box that you can buy and implement. At least not yet.

What’s been your experience with healthcare analytics? Where are you seeing success? We’d love to hear your thoughts in the comments.

The Sexiest Data in Health IT: Datapalooza 2017

Posted on May 15, 2017 I Written By

Healthcare as a Human Right. Physician Suicide Loss Survivor. Janae writes about Artificial Intelligence, Virtual Reality, Data Analytics, Engagement and Investing in Healthcare. twitter: @coherencemed

The data at this conference was the Best Data. The Biggest Data. No one has better data than this conference.

The sexiest data in all of healthIT was highlighted in Washington DC at Datapalooza April 27-28, 2017.  One of the main themes was how to deal with social determinants of health and the value of that data.  Sachin H. Jain, MD of Caremore Health reminded us that “If a patient doesn’t have food at home waiting for them they won’t get better” social data needs to be in the equation. Some of the chatter on the subject of healthcare reform has been criticism that providing mandatory coverage hasn’t always been paired with knowledge of the area. If a patient qualifies for Medicaid and has a lower paying job how can they afford to miss work and get care for their health issues?
Rural areas also have access issues. Patient “Charles” works full time during the week and qualifies for Medicaid. He can’t afford to miss a lot of work but needs a half a day to get treatments which affect his ability to work. There is no public transportation in his town to the hospital in a city an hour and a half away. Charles can’t afford the gas or unpaid time off work for his treatment.

Urban patient “Haley” returns to her local ER department more than once a week with Asthma attacks.  Her treatments are failing because she lives in an apartment with mold in the walls. As Craig Kartchner from the Intermountain Healthcare team responded to the #datapalooza  hashtag online- These can be the most difficult things to change.

The 2016 report to Congress addresses the difficulty of the intersection between social factors and providing quality healthcare in terms of Social Determinants of Health:

“If beneficiaries with social risk factors have worse health outcomes because the providers they see provide low quality care, value based purchasing could be a powerful tool to drive improvements in care and reduce health disparities. However, if beneficiaries with social risk factors have worse health outcomes because of elements beyond the quality of care provided, such as the social risk factors themselves, value based payment models could do just the opposite. If providers have limited ability to influence health outcomes for beneficiaries with social risk factors, they may become reluctant to care for beneficiaries with social risk factors, out of fear of incurring penalties due to factors they have limited ability to influence.”

Innovaccer just launched a free tool to help care teams track and monitor Medicare advantage plans. I went to their website and looked at my county and found data about the strengths in Salt Lake where I’m located. They included:

  • Low prevalence of smoking
  • Low Unemployed Percentage
  • Low prevalence of physically inactive adults

Challenges for my area?

  • Low graduation rate
  • High average of daily Air pollution
  • High income inequality
  • High Violent crime rate per 100,000 population

Salt Lake actually has some really bad inversion problems during the winter months and some days the particulate matter in the air creates problems for respiratory problems. During the 2016-2017 winter there were 18 days of red air quality and 28 days of yellow air quality. A smart solution for addressing social determinants of health that negatively impact patients in this area could be addressing decreasing air pollution through increased public transportation. Healthcare systems will see an increase in cost of care during those times and long term population health challenges can emerge. You can look at your county after you enter your email address on their site. This kind of social data visualization can give high level insights into the social factors your population faces.

One of the themes of HealthDataPalooza was how to use system change to navigate the intersection between taking care of patients and not finding way to exclude groups. During his panel discussion of predictive analytics, Craig Monson the medical director for analytics and reporting discussed how “data analytics is the shiny new toy of healthcare.”    In addition to winning the unofficial datapalooza award for the most quotes and one liners – Craig presented the Clinical Risk Prediction Initiative (CRISPI).  This is a multi variable logistic regression model with data from the Atrius health data warehouse. His questions for systems to remember in their data analysis selection are “Who is the population you are serving? What is the outcome you need? What is the intervention you should implement?”

Warning- Craig reminds us that in a world of increasing sexy artificial intelligence coding a lot of the value analysis can be done with regression. Based on that statement alone I think he can be trusted. I still need to see his data.

CRISPI analyzed the relative utility of certain types of data, and didn’t have a large jump in utility when adding Social Determinant Data. This data was one of the most popular data sets during Datapalooza discussions but the reality of making actionable insights into system improvement? Craig’s analysis said it was lacking. Does this mean social determinant data isn’t significant or that it needs to be handled with a combination of traditional modeling and other methods?  Craig’s assertion seemed to fly in the face of the hot new trend of Social Determinants of Health data from the surface.

Do we have too much data or the wrong use of the data? Most of the companies investing into this space used data sources outside the traditional definition to help create solutions with social determinate of health and Patient outcomes. They differed in how they analyzed social determinant data. Traditional data sources for the social determinants of health are well defined within the public health research.  The conditions in which you work and live impact your health.

Datapalooza had some of the greatest minds in data analytics and speakers addressed gaps in data usefulness. Knowing that a certain large county wide population has a problem with air quality might not be enough to improve patient outcomes. There is need for analysis of traditional data sources in this realm and how they can get meaningful impact for patients and communities. Healthcare innovators need to look at different data sources.  Nick Dawson, Executive director of Johns-Hopkins Sibley Innovation Hub responded to the conversation about food at home with the data about Washington DC.  “DC like many cities has open public data on food scarcity. But it’s not part of a clinical record. The two datasets never touch.” Data about food scarcity can help hospital systems collaborate with SNAP and Government as well as local food programs. Dawson leads an innovation lab at Johns Hopkins Sibley where managers, directors, VPs and C Suite leaders are responsible for working with 4 innovation projects each year.

Audun Utengen, the Co Founder of Symplur said “There’s so much gold in the social media data if you choose to see it.” Social data available online helps providers meet patients where they are and collect valuable data.  Social media data is another source to collect data about patient preferences and interactions for reaching healthcare populations providers are trying to serve. With so much data available sorting through relevant and helpful data provides a new challenge for healthcare systems and providers.

New Data sources can be paired with a consultative model for improving the intersection of accountable care and lack of access due to social factors. We have more sophisticated analytic tools than ever for providing high value care in the intersection between provider responsibility and social collaboration. This proactive collaboration needs to occur on local and national levels.  “It’s the social determinants of health and the behavioral aspects that we need to fund and will change healthcare” we were reminded. Finding local community programs that have success and helping develop a strategy for approaching Social Determinants of Health is on the mind of healthIT professionals.

A number of companies examine data from sources such as social media and internet usage or behavioral data to design improvements for social determinants of health outcomes.   They seek to bridge the gaps mentioned by Dawson. Data sets exist that could help build programs for social determinants of health.  Mandi Bishop started Lifely Insights centered around building custom community plans with behavioral insights into social determinant data. Health in all Policies is a government initiative supporting increased structure and guidelines in these areas. They support local and State initiatives with a focus on prevention.

I’m looking forward to seeing how the data landscape evolves this year. Government Challenges such as the Healthy Behavior Data Challenge launched at Datapalooza will help fund great improvements. All the data people will get together and determine meaningful data sets for building programs addressing the social determinants of health. They will have visualization tools with Tableau. They will find ways to get food to patients at home so those patients will get better. Programs will find a way to get care to rural patients with financial difficulty and build safe housing.

From a healthcare delivery perspective the idea of collaborating about data models can help improve community health and decrease provider and payer cost. The social determinants of health can cost healthcare organizations more money than data modeling and proactive community collaboration.

Great regressions, saving money and improving outcomes?

That is Datapalooza.

Randomized Clinical Trial Validates BaseHealth’s Predictive Analytics

Posted on March 11, 2016 I Written By

Andy Oram is an editor at O'Reilly Media, a highly respected book publisher and technology information provider. An employee of the company since 1992, Andy currently specializes in open source, software engineering, and health IT, but his editorial output has ranged from a legal guide covering intellectual property to a graphic novel about teenage hackers. His articles have appeared often on EMR & EHR and other blogs in the health IT space. Andy also writes often for O'Reilly's Radar site ( and other publications on policy issues related to the Internet and on trends affecting technical innovation and its effects on society. Print publications where his work has appeared include The Economist, Communications of the ACM, Copyright World, the Journal of Information Technology & Politics, Vanguardia Dossier, and Internet Law and Business. Conferences where he has presented talks include O'Reilly's Open Source Convention, FISL (Brazil), FOSDEM, and DebConf.

One of the pressing concerns in health care is the validity of medical and health apps. Because health is a 24-hour-a-day, 365-day-a-year concern, people can theoretically overcome many of their health problems by employing apps that track, measure, report, and encourage them in good behavior. But which ones work? Doctors are understandably reluctant to recommend apps–and insurers to cover them–without validation.

So I’ve been looking at the scattered app developers who have managed to find the time and money for randomized clinical studies. One recent article covered two studies showing the value of a platform that provided the basis for Twine Health. Today I’ll look at BaseHealth, whose service and API I covered last year.

BaseHealth’s risk assessment platform is used by doctors and health coaches to create customized patient health plans. According to CEO Prakash Menon, “Five to seven people out of 1,000, for instance, will develop Type II diabetes each year. Our service allows a provider to focus on those five to seven.” The study that forms the basis for my article describes BaseHealth’s service as “based on an individual’s comprehensive information, including lifestyle, personal information, and family history; genetic information (genotyping or full genome sequencing data), if provided, is included for cumulative assessment.” (p. 1) BaseHealth has trouble integrating EHR data, because transport protocols have been standardized but semantics (what field is used to record each bit of information) have not.

BaseHealth analytics are based on clinical studies whose validity seems secure: they check, for instance, whether the studies are reproducible, whether their sample sizes are adequate, whether the proper statistical techniques were used, etc. To determine each patient’s risk, BaseHealth takes into account factors that the patient can’t control (such as family history) as well as factors that he can. These are all familiar: cholesterol, BMI, smoking, physical activity, etc.

Let’s turn to the study that I read for this article. The basic question the study tries to answer is, “How well does BaseHealth predict that a particular patient might develop a particular health condition?” This is not really feasible for a study, however, because the risk factors leading to diabetes or lung cancer can take decades to develop. So instead, the study’s authors took a shortcut: they asked interviewers to take family histories and other data that the authors called “life information” without telling the interviewers what conditions the patients had. Then they ran the BaseHealth analytics and compared results to the patients actual, current conditions based on their medical histories. They examined the success of risk assignment for three conditions: coronary artery disease (CAD), Type 2 diabetes (T2), and hypertension (HTN).

The patients chosen for the study had high degrees of illness: “43% of the patients had an established diagnosis of CAD, 22% with a diagnosis of T2D and 70% with a diagnosis of HTN.” BaseHealth identified even more patients as being at risk: 74.6% for CAD, 66.7% for T2D, and 77% for HTN. It makes sense that the BaseHealth predictions were greater than actual incidence of the diseases, because BaseHealth is warning of potential future disease as well.

BaseHealth assigned each patient to a percentile chance of getting the disease. For instance, some patients were considered 50-75% likely to develop CAD.

The study used 99 patients, 12 of whom had to be dropped from the study. Although a larger sample would be better, results were still impressive.

The study found a “robust correlation” between BaseHealth’s predictions and the patients’ medical histories. The higher the risk, the more BaseHealth was likely to match the actual medical history. Most important, BaseHealth had no false negatives. If it said a patient’s risk of developing a disease was less than 5%, the patient didn’t have the disease. This is important because you don’t want a filter to leave out any at-risk patients.

I have a number of questions about the article: how patients break down by age, race, and other demographics, for instance. There was also an intervention phase in the study: some patients took successful measures to reduce their risk factors. But the relationship of this intervention to BaseHealth, however, was not explored in the study.

Although not as good as a longitudinal study with a large patient base, the BaseHealth study should be useful to doctors and insurers. It shows that clinical research of apps is feasible. Menon says that a second study is underway with a larger group of subjects, looking at risk of stroke, breast cancer, colorectal cancer, and gout, in addition to the three diseases from the first study. A comparison of the two studies will be interesting.

Where Will We See Analytics in Ambulatory Medicine?

Posted on April 9, 2015 I Written By

John Lynn is the Founder of the blog network which currently consists of 10 blogs containing over 8000 articles with John having written over 4000 of the articles himself. These EMR and Healthcare IT related articles have been viewed over 16 million times. John also manages Healthcare IT Central and Healthcare IT Today, the leading career Health IT job board and blog. John is co-founder of and John is highly involved in social media, and in addition to his blogs can also be found on Twitter: @techguy and @ehrandhit and LinkedIn.

As I prepare to head to the mecca of healthcare IT conferences, I’m getting inundated with pitches. Much like last year, analytics is still a really popular topic. It seems like every healthcare IT vendor has some analytics offering. Many abuse the term analytics (which is fine by me) and that term has come to mean analyzing your health data in order to provide value.

As I think about analytics, I wonder how much of it will apply to the small physician practice. I should do this survey, but I bet if I asked 5 physicians working in small group or solo practices about their analytics strategy they’d all give me blank stares. Small physician practices don’t have an analytics strategy. They’re not looking for new ways to leverage analytics to improve their practice. That’s not how a small practice thinks.

So, does analytics have a place in small ambulatory medicine?

The short answer is that it absolutely does. However, I think that it will be delivered in two forms: packaged or purchased by the borg.

In the packaged approach, analytics will be part of a small practice’s EHR system. Much like a doctor doesn’t have a mobile EHR strategy (it just comes with the EHR), they won’t have an analytics strategy either. They’ll just take the analytics solutions that come with the EHR.

In some ways, the reporting capabilities in an EHR have been doing this forever. However, very few organizations have been able to use these reports effectively. The next trend in EHR analytics will be to push the data to the user when and where they need it as opposed to having to pull a report. Plus, the EHR analytics will start trying to provide some insights into the reports as opposed to just displaying raw data.

One key for ambulatory EHR vendors is that they won’t likely be able to build all the EHR analytics functions that a doctor will want and need. This is why it’s so important that EHR vendors embrace the open API approach to working with outside companies. Many of these third party software companies will provide EHR analytics on top of the EHR software.

In the purchased by the borg scenario, the small practice will get purchased by the borg (You know…the major hospital system in the area). This is happening all over the place. In fact, many small practices cite the reason for selling out to the local hospital is that they don’t think they’ll be able to keep up with the technology requirements. One of those major requirements will be around analytics. We’ll see how far it goes, but I think many small practices are scared they won’t be able to keep up.

Ok, there’s one other scenario as well. The local hospital system or possibly even a local ACO will purchase a package of analytics software (ie. purchased by the bog) and then you’ll tap into them in order to get the benefits of a healthcare analytics solution. We see this already starting to happen. I’ve heard mixed results from around the country. No doctor really likes this situation since it ties them so deeply with the local hospital, but they usually can’t think of a better option.

That’s my take on how analytics will make its way to ambulatory practices. Of course, most large hospital systems also own a large number of ambulatory practices as well. So, some of the analytics will trickle down to ambulatory in those systems as well. I just wonder how much value ambulatory doctors will get from the hospital analytics vendors that are chosen. I can already hear the ambulatory doctors complaining about the analytics reports that don’t work for them because they’re so hospital focused.

Where are you seeing analytics in the ambulatory setting?