The Future of Targeted Healthcare

Healthcare’s appointment-level data has the potential to do a lot of good. A row of an appointment-level dataset usually includes provider detail, name and npi, patient detail, and appointment detail, length of time and exact type, down to the specific procedure code. When collected at the region, state, or national level, this dataset becomes all-powerful.

Leakage analytics is a popular way in which the appointment-level dataset is used. The dataset is analyzed over a certain timeframe, following a patient’s movements. For a two-month window an Oregon-based patient may see:

  • An E.R. doctor for a broken wrist at A Healthcare for $2,000, day 0
  • A Wrist Specialist a week later at B Healthcare for $3,000, day 7
  • A Surgeon for a consultation a month later at C Healthcare for $3,500, day 37
  • An annual checkup 3 days later at A Healthcare for $200, day 40
  • A Surgeon, a week after the consultation, for wrist surgery at C Healthcare for $20,000, day 44

The surgery is the last appointment in the dataset that falls within the two-month timeframe following that patient’s movements around the state of Oregon. It is possible this patient had a few follow-up consultations for their wrist or other associated appointments; however, if those appointments fell outside of the two-month window, they were not included in this dataset. It is also unlikely that the appointment-level dataset includes price; however, this information is easy to join in by the appointment field/procedure code.

Using the bulleted list above and advanced appointment groupings, A Healthcare leaked $26,500 and B Healthcare leaked $23,500. That is a lot of money — and for just an single patient. Understanding the impact of multiplying like numbers by total patients, A and B Healthcare Networks will be incentivized to utilize leakage analytics to better understand which providers and practices are the biggest “leakers” of potential hospital dollars, and work on a comprehensive anti-leakage campaign. Let’s focus in on the potential mitigation strategies of the B Healthcare network.

On the provider side, B Healthcare conducts leakage-awareness workshops to ensure providers are aware of the cost of care being lost to other healthcare networks. Additionally, B Healthcare may incentivize its providers to send patients to other specialists and generalists within the network — a $10,000 bonus for 100 providers leaking 3% or less of their patients is less expensive than the potential $15 Million, 12% of a hypothetical provider’s cost of care, leaked by just 1 of those 100 providers the previous year. Alternatively, B Healthcare may create a new role in the network — a referral specialist. The referral specialist is the source of truth for the best providers, in-network of course, to see for a follow-up or independent appointment, thus taking the leakage onus completely off of providers and avoiding monetary rewards altogether.

On the campaign side, B Healthcare will try to further their understanding of leakage in an attempt to eliminate it. Using the same appointment-level dataset but only analyzing the appointment and time detail fields, B Healthcare may see which appointments are on the rise, “trending”, or in decline. As an example, lets say in a one-year window of the appointment-level dataset, B Healthcare sees heart disease has been trending in Oregon. B Healthcare will especially want to prevent leakage in this trending category. So much so that they join patient-level information in to their analyses. Which patients that have visited, and therefore have given their information to, B Healthcare are at risk of heart disease? Using the combination of factors that push a person into an at-risk category: age, medical history, family history, and zip code as a proxy for diet, wealth, and overall health, B Healthcare may then use the patient’s contact information to text them around the age of 65 years of age — when they are most at risk of heart disease — suggesting they come in for a check up. Or, better yet, B Healthcare decides to text them 5 years earlier, at age 60, getting ahead of other networks and picking up the patient. Using this proactive strategy, B Healthcare avoids loosing this patient to another healthcare network to the best of it’s ability — attempting to retain the patient for a disease they very likely do not yet know they have. As our understanding of the exact timing and circumstances of certain ailments becomes more advanced, such leakage mitigation methods may become more and more targeted and exact, perhaps alerting patients of their potential ailments with a probability attached.

Beyond provider-level action and more technically-detailed campaigns to reduce leakage, healthcare networks may use appointment-level datasets for other insights and decision-making power. Other areas include:

  • Real estate: providers, patients, and administrators are those people that make up the population most frequently using a Healthcare network’s buildings. If a certain segment of the population shows to be steeply declining in a particular appointment area that the network has a specialized space for it may be advantageous for that network to sell off some of its space in anticipation of declining appointments.
  • Targeted drug prescriptions: a network may join prescriptions by appointment data into the appointment-level dataset to understand which drug providers are most frequently prescribed for a particular appointment category. Healthcare networks ensure that those contracts are strong or encourage a switch to other, similar medications, with stronger contracts.

All of these practices are not as far away as you may think — as there are many third party data providers in the market of providing this exact dataset service. The only thing healthcare networks need to do is accurately collect and enter their data. Of course this is easier said than done, but with other technologies and algorithms on the horizon, historical data cleansing and the simplified maintenance of valuable data may already be in the works…

--

--

--

Consultant / Data & Analytics

Love podcasts or audiobooks? Learn on the go with our new app.

Recommended from Medium

How To Use PyCall.jl: Python Libraries In Julia

How to Become a Machine Learning Engineer? | 365 Data Science

Conner Huertas Del Pino v Jorge Brian Panta Live[Stream•]?!

Online live stream search engine

10 Python Libraries to Help You Work With Data

Survival of the Fittest…Model

PULSEPAD WHITELIST — The first IDO platform for Pulse Chain

Simulating the Coronavirus Epidemic in Switzerland with a realistic model

Getting data from the previous row in your table visual in Power BI

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Abby McCulloch

Abby McCulloch

Consultant / Data & Analytics

More from Medium

How AI will Revolutionize the Way Businesses Work

#5 Founder’s Showcase: Celadon (Sophie and Christian Silver)

How Do Patients Feel About AI?

Go Automatic In A World Of Stick Shifts