Fractional outcomes research for healthcare

Know what your program
actually changed.

I help health plans and care programs answer the question every budget review asks: did it work, and was it worth it? Decision-grade evidence: rigorous causal evaluation and program ROI, explained in plain language.

Who I help

Teams that need decision-grade evidence

Usually a program is live, a renewal or budget decision is coming, and the current numbers would not survive a skeptical CFO.

P

Health plans

Evaluate care management, disease management and vendor programs. Know which ones reduce cost and which just look good on a slide.

V

Care programs and digital health

Build the evidence your buyers will ask for, with a methodology their actuaries and analysts can check.

R

Value-based and provider groups

Measure the impact of interventions on utilization and total cost of care across your attributed population.

Services

Three ways to work together

Program ROI evaluation

A retrospective look at a program that is already running, using your claims data.

  • Matched or weighted comparison group
  • Difference-in-differences savings estimate
  • PMPM savings and ROI with confidence intervals
  • Executive summary plus full methods appendix

Evaluation by design

Before launch or contract signing, agree on how success will be measured.

  • Outcome definitions and attribution rules
  • Comparison strategy and sample size check
  • A written study plan both sides sign off on

Fractional outcomes lead

Ongoing part-time support for a team without a dedicated outcomes researcher.

  • Methods review of internal or vendor analyses
  • Mentoring analysts on causal methods
  • A repeatable evaluation playbook
Approach

From question to decision

Every evaluation follows the same simple path, and you see the reasoning at each step.

01

Question

Start with the decision the result will drive: renew, expand, redesign or stop.

02

Design

Define who was treated, when, and who they should be compared to.

03

Estimate

Build a fair comparison group, check the assumptions, and measure the difference.

04

Decide

Translate the effect into dollars, with the range of plausible outcomes, not one number.

An honest counterfactual

Pre/post comparisons overstate savings because high-cost members tend to get cheaper on their own. Every estimate is measured against a credible "what would have happened anyway."

Assumptions in the open

Each result comes with the assumptions it rests on, how they were tested, and what would change the answer.

Your data stays with you

Analysis runs in your environment. Only aggregate, de-identified results leave it, with small cells suppressed.

Examples

The kinds of questions I answer

Illustrative scenarios, not client case studies.

Care management

"Our vendor reports 3x ROI. Is that real?"

Rebuild the analysis with a matched comparison group and difference-in-differences. Separate the effect of the program from regression to the mean, then report savings for everyone offered the program and for those who actually engaged.

Program design

"We launch in Q2. How will we know it worked?"

Write the evaluation plan before day one: eligibility, index dates, outcome measures, comparison strategy and the minimum effect the program needs to show to pay for itself.

[Headshot placeholder]
About

Hi, I'm Olivia.

I'm a healthcare outcomes researcher focused on causal inference and program evaluation. [Add 1 to 2 sentences on your background: years of experience, types of organizations, notable work.]

[Add credentials or education, if you want them listed.]

I believe analytics should sharpen judgment, not replace it. The people who run a program own the question and the decision. My job is to give them an answer that is rigorous, transparent and honest about its uncertainty.

[LinkedIn URL]

Have a program that needs an answer?

Book a free 30-minute call. We'll talk through the decision you're facing and whether an evaluation would help.

Book a call  Email me

Placeholders: booking link (e.g., Calendly or Cal.com); email assumes hello@decisiongrade.co.