OurHealth Leverages Data to Find More Than $150M in Potential Savings

At OurHealth, we focus on improving the health and wellness of the populations we serve through patient-centered primary care with a heavy emphasis on leveraging data and technology. That’s why we were excited to take part in a recent hackathon hosted by Indiana HIMSS and KSM Consulting.

On October 16, 2017, the Indiana Management Performance Hub along with the Indiana Family and Social Services Administration (FSSA) released 25 data sets of aggregated, historical Medicaid claims data from 2012 to 2016 (partial years data). This data covered topics such as top services, mental health, transportation, emergency department utilization and more. This data was released in advance of a one-day hackathon conducted at the Regenstrief Institute. At the hackathon, data scientists were challenged to build visualizations and predictive models leveraging the dataset.

The OurHealth data and analytics team earned first prize in the data visualization challenge, setting out to answer the key question, “How much of the Emergency Department spend might be avoidable?”

Why is this an important question to answer? Healthcare wastes billions of dollars every year and much of this waste happens in the Emergency Department (ED). According to a study of emergency room visits in New York, nine out of 10 visits could have been treated in a different setting. Utilizing the ED in this manner cost $1.3 billion. This fact and the answer to the question led us to dig deeper for trends in this newly released data.

Here’s how we went from a dataset to insight in less than 24 hours

In 2000, New York University developed an Emergency Department Profiling Algorithm which sought to answer the following questions amongst New York’s Medicaid patients, “What proportion of ED cases could be treated in a primary care setting?” and “How much emergent ED use is preventable or avoidable with timely and effective primary care?” This algorithm classifies ED visits into 4 groups including:

• Non-emergent
• Emergent/Primary Care Treatable
• Emergent – ED Care Needed – Preventable/Avoidable
• And Emergent – ED Care Needed – Not Preventable/Avoidable

We sought to understand how these groups may be prevalent within the released FSSA Emergency Department datasets. We mapped the diagnosis codes across these datasets (with some data clean up) to the NYU probabilities that the diagnosis fell within each of the four groups above.  Once mapped we further mapped the diagnosis codes to the Agency for Healthcare Research and Quality (AHRQ) Clinical Classification System Codes (CCS) for both ICD9 and ICD10 allowing us to group the diagnosis codes into clinically relevant groups in our visualization.

Once we had the mapping completed, we began our data exploration. First, we sought to answer the question, “How much spend and claims volume fall into each of these categories?” To do this, we assigned the spend and claims volume for a diagnosis to the category which had greater than 50 percent probability that it was associated with one of the four groups above that group. As shown in our visualization (circled in red), we identified more than $150 million worth of Non-Emergent and potentially Primary Care Treatable visits. We found this very interesting and if even half this was correct based on a deeper analysis of claims, it posed a significant savings opportunity.

We dug deeper…

Next, we asked, “What specialties are driving this spend?” and “What diagnoses are driving the Non-Emergent and potentially Primary Care Treatable visits?”

We found that the top specialties driving this spend included Acute Care, Ambulance, Pharmacy, Emergency Medicine Practitioners and Medical Clinic. All these specialties were defined in the files and likely need broken down to a medical providers NPI in future analysis if deeper data is made available.

To answer the diagnosis question, we took the mapped CCS codes and broke the spend into Non-Emergent and Primary Care Treatable. This allowed us to further segment the trends by age group and try out some cool new Tableau 10.5 functionality of visualizations in tool tips to answer sub-questions.

We identified a few notable patterns:

• In the 0 to 5 years age group, millions of dollars of potentially Primary Care Treatable visits were spent on a diagnosis of Fevers of Unknown Origin, as well as potentially Non-Emergent visits of Other Upper Respiratory Infections.
• Potentially Non-Emergent Abdominal pain visits stood out for the 18 to 32 and 33 to 48 age groups.
• In the 49 to 64 age group, potentially Non-Emergent chest pains made up millions of dollars in spend.
• In the 65+ group, Urinary Tract Infections topped the list.

We won’t go into detail here but in our visual a user can look at a particular facility and see these trends for each entity as well as search by city across Indiana.

Why is all this important?

As I mentioned, OurHealth’s mission is to improve the health and wellness of the populations we serve through our primary care services using technology and data insights. In the case of avoidable ED visits, we ask ourselves questions like, how can we help patients get to the right level of service? And, how could we have prevented that visit? If you think about the typical ED experience, a person checks in and can wait hours to be seen, or spend hours going through the process. This leads to a ton of lost productivity, a frustrating experience on behalf of the patient, and a much higher cost of care.

Take the example of the millions spent on abdominal pain. What if those patients had someone they could call and get an opinion prior to going to the Emergency Department? Would have they gone? Likely not in most cases.

In the patient populations served by OurHealth (those self-insured by their employer), avoiding the ED creates huge savings. In these populations, a trip to the ED would likely cost the patient five to ten times more than what a visit to a primary care provider would cost. And had they visited an OurHealth onsite or near-site clinic, the visit would have likely been free or cost very little.

OurHealth’s primary care clinics also provide a better patient experience by cutting wait times to just minutes and often providing same-day appointments. As a nurse myself, this especially hits home to me.

Our team had fun exploring this data and working in a focused 24-hour timeframe to produce a winning solution for the visualization challenge. We won $1,000, which we plan to donate to Second Helpings, a charitable food pantry in Indianapolis. Good health starts with a great meal and this organization provides just that for those in need in our community. Our donation will help provide more than 900 meals to folks in need. Congratulations and thank you to the data team, Justin Richardson, John Botta and Stan Brown, for their hard work!

 

About the Author:

Brian Norris RN-BC MBA FHIMSS is Vice President of Data and Analytics at OurHealth, a leading provider of employer-based clinic solutions including onsite and near-site primary care clinics. By leveraging Norris and his data team’s insights, OurHealth fulfills its mission is to increase access to primary care providers, improve health outcomes, and reduce healthcare spending for clients through wellness strategy. For more info, visit ourhealth.org or @OurHealthClinic. Follow Brian at @Geek_Nurse.