A blueprint for a scalable health data strategy

I like the idea of a blueprint for developing a scalable data strategy. It isn't smart to recommend a step by step recipe for something as abstract and heterogenous as an analytics strategy. I imagine many organizations or even individuals distribute around the mean of a normal curve and are smug and self congratulatory regarding data frameworks or even a process for evaluating data on hand.

A blueprint is created with the entire framework integrated and included in the evolving architectural plan. We need to see what is possible before selecting the "finishes" or design elements.

You will need to trust me up front or at least stay while I walk through a case study--when I tell you a large part of ineffective data strategy stems from "not knowing what we don't know". The types of questions and even answer options integrated into a survey will definitely distort your findings and diminish insights pulled from analyses. Remember the old adage, "garbage in, garbage out"?

Data out in the real world is complicated. Understandably the patient populations in clinical trials need to  be homogenous so that we can make assumptions about any signals observed during the trial period.

​Unfortunately we are limited in what the data will mean outside in the real world. Your clinic waiting room is not curated. Patients have comorbidities, differences in sex, age, ethnicity, social correlates, and will likely respond differently out in the real world beyond the limits of small study populations and limited evaluation periods.

The Medical Expenditure Panel Survey

You should build your strategy with publically available data first. The Medical Expenditure Panel Survey or MEPS collects core interview content:

• demographics
• charges and payments
• medical conditions
• employment
• health status
• health insurance
• utilization

There are also supplemental sections where depending on interview rounds participants are also asked about:

• access to care
• child preventative health
• satisfaction with health plans & providers
• assets
• income
• preventitive care

Additional supplemental questionnaires are also completed when patients are identified:

• Diabetes Care Survey (DCS)
• Adult Self-Administered Questionnaire (SAQ)
• Cancer SAQ

---

The 2-day workshop provided guidance on what to do when you need additional granularity?

Understanding what is happening under the hood and how the questions are coded was helpful information. I have been using MEPS data for over a year now but a deeper dive into understanding the framework was a tremendous asset to current and future projects.

​Stay-tuned...

The goal is to introduce a few sources of freely available data and then show how they can be combined to answer  a variety of relevant questions about healthcare utilization, real world evidence, and patient populations or providers.

The blog will be a general guide for those of us with a bit of experience accessing data in a variety of formats. If you would like more granularity or detail you can pre-order (2.99!)

​Blueprint: A Scalable Healthcare Data Strategy

---

Grab your free e-copy of Improving Numeracy in Medicine--a brief give-away so get it while its hot!

I will keep the statistical information "light" here on the website but the short guide to common statistical approaches in the literature can be useful.

Thoughtful discussions about content development and outcomes analytics that apply the principles and frameworks of health policy and economics to persistent and perplexing health and health care problems...