Longitudinal Data & Individualized Healthcare

When summary statistics aren’t enough.

by Anurati Mathur, Co-founder & CEO at Sempre Health

Population-level healthcare statistics have long been used in the realm of public health as barometers of health system effectiveness. In the last decade, these metrics have grown to become management tools within value-based healthcare delivery as well. From HEDIS (Healthcare Effectiveness Data and Information Set) to MIPS (Merit-Based Incentive Payment System), measures like ‘Percent of diabetics with A1C > 9’ are the widespread basis of clinical and payment decisions alike.

But, they remain imperfect.

Not only do statistical measures typically represent a single point in time, they ignore the heterogeneity of the target population. And, while only an estimated 26% of diabetes patients have an A1C > 9, for a patient who falls in that 1 of 4, living with uncontrolled diabetes is the only experience. In a world that is increasingly personalized and consumer-centric, it isn’t enough to know you fall in the bottom quartile relative to others. The individualization of healthcare is long overdue.

As US healthcare data volumes approach 2,000 Exabytes in 2017, it is increasingly easier to construct longitudinal profiles for individuals in a way that hasn’t been feasible. Traditionally captured data (prescriptions, medical claims, EMR notes, etc.), when further explored on the axes (like time), reveal trends and anomalies at the individual level, which can then inform interventions in a newly nuanced way.

There are two immediate benefits of the longitudinal approach:

1. Population-level measures can be deconstructed to reveal trends amongst cohorts. Disease progression is better detected, and patient interventions become more targeted.

In the example below, disaggregating the summary statistic on the left reveals subgroups with varying clinical trends. Further examining the demographic and clinical composition of these subgroups can inform personalized and refined patient programs.

2. Patients can be treated proactively and not simply when their outcomes exceed measure thresholds.

In the example below, the patient’s A1C is greater than 9 in months 6, 7, 10 and 11. Intervening only when the patient exceeds the measure’s threshold often means a missed opportunity for preventative care. Conversely, managing to longitudinal trends allows for proactive and focused patient-level management.

Approaching populations as a homogenous group risks wasting time and resources on ineffective interventions. Using multi-dimensional data to understand a patient – clinically, financially and socially – has the potential to redefine care management. Reliable patient understanding proves especially important in addressing complex healthcare issues, like medication nonadherence.

Five years ago I was prescribed allergy eye drops. I learned that I would have to pay $150 out-of-pocket and at so high a price point, I abandoned the prescription. In the week that followed, I received several phone calls and text reminders to pick up the medication, but none of that could address my biggest problem – price.

This example may seem simple, but the point is important. Identifying the right intervention for the right patient at the right time can be the difference between sickness and health. And, in a world moving rapidly towards value-based care, the wrong approach can be costly!

At Sempre Health, we’re financially aligning patients with their care. We parse millions of prescription and clinical records to better understand you. Then, we make your healthcare benefits individualized and responsive. For example, you might earn discounts on your copays when filling your prescriptions on time. This means you can finally share in the savings you generate when you make healthy choices. The future of individualized healthcare is here.

To learn more about Sempre Health, visit: www.semprehealth.com.

Leave a Reply

Your email address will not be published. Required fields are marked *

Security Captcha * Time limit is exhausted. Please reload the CAPTCHA.