Case Studies

Proven Results

Real outcomes at leading health systems across Dallas-Fort Worth and beyond. Every engagement is measured against benchmarks you helped define.

$2.6M+

Annual savings generated

94%

Scheduling error reduction

62K+

Patients screened

200K+

ML training cases

Parkland Health

Early Diabetes Detection

Healthcare Analytics · ML Screening

Personalized ML screening across 62,000+ patients cut costs by 33% and doubled health benefit.

The Challenge

Parkland Health needed a more targeted approach to diabetes screening — one that could improve outcomes across a large, diverse patient population without proportionally increasing costs.

Our Approach

We developed a personalized ML screening model that identified patients most likely to benefit from early intervention, rather than applying uniform screening criteria. The model was trained and validated on Parkland's own patient data, ensuring it reflected the actual population being served.

The Outcome

The ML-based screening approach cut screening costs by 33%, doubled the measurable health benefit, and generated an estimated $2.6 million in annual savings — while improving care for 62,000+ patients.

Key Metrics

33%

Reduction in screening costs

2×

Health benefit achieved

$2.6M

Estimated annual savings

62,000+

Patients screened

Children's Health Dallas

Surgical Scheduling Optimization

Operational Decision Support · ML Prediction

ML-based prediction reduced median scheduling error by 94%, saving $407–$701 per case in OR time.

The Challenge

Inaccurate surgical scheduling was creating significant waste at Children's Health Dallas — both in direct OR costs and in downstream scheduling inefficiency. Small errors in predicted case duration compound across an operating room schedule to produce large losses.

Our Approach

We built an ML-based case duration prediction model trained on more than 200,000 historical surgical cases. The model captured procedure-specific, surgeon-specific, and patient-specific variation that rule-based scheduling systems can't address.

The Outcome

The prediction model reduced median scheduling error by 94%, translating to an estimated $407 to $701 in savings per OR case — across one of the largest pediatric surgical programs in the country.

Key Metrics

94%

Reduction in scheduling error

$701

Max savings per OR case

$407

Min savings per OR case

200K+

Cases in training dataset

Parkland Hospital

Dialysis & ED Congestion Reduction

Workflow Re-Engineering · Analytics

Analytics identified a small dialysis cohort driving 20–30% of ED visits. Workflow redesign reduced strain without adding capacity.

The Challenge

Parkland Hospital's ED was experiencing persistent congestion that appeared to require significant capacity expansion. Standard approaches to ED throughput improvement weren't identifying the root cause.

Our Approach

Analytics revealed a non-obvious driver: a relatively small dialysis patient cohort was responsible for 20–30% of ED visits, many of which were avoidable with better outpatient coordination. We mapped the specific workflow gaps causing these visits and designed targeted interventions.

The Outcome

A workflow redesign targeting this cohort reduced ED strain measurably — without adding physical capacity or significant new resources. The engagement demonstrated the value of analytics-first problem framing before reaching for capacity-based solutions.

Key Metrics

20–30%

Of ED visits tied to cohort

Additional capacity required

↓

ED congestion reduced

Targeted patient cohort

What We Measure

Health systems that work with us come away with more than a set of recommendations. They come away with a different way of running their operation.

Avoidable costs identified, quantified, and eliminated
Clinical and operational workflows that are reliable, not just improved on paper
Decision-making capability embedded in your team, not dependent on outside expertise
Measurable outcomes tied to benchmarks you helped define

Book a 30-Minute Discovery Call

Send us your hardest operational problem. If we can't tell you something useful in 30 minutes, we're not the right firm.

Schedule Your Call