ATG Health
DELIVERING ANALYTICAL INSIGHTS
1. Predictive Modeling
Statistical Algorithms have been applied to claims, member demographic and utilization to predict likelihood of a member going into care.a. Applied to HEDIS measures and helps to target members for gap closure
b. Target List can be produced on a regular/monthly basis
c. Methodology is scalable to other plans/industry
d. Outcomes: Helped major plans to move from bottom of the list to be consistently in the top 2 among competing MCOs. This has helped out health promotions department to be more efficient in making outreach efforts. Their operational expertise has helped plan to reach the plan's Value Based Purchasing targets.
e. Key part of the presentation to the department of health who expressed high interest in it as a tool for population health management.
2. Disparity Analytics:
Conducted to find difference in HEDIS rates/utilization across different race/ethnicity groups.Also looking into disparity in healthcare between urban and rural areas.
Analysis developed using SQL and BI toola like Tableau that can be consumed by different users in a self service mode.
3. Population Health
a. Developed key metrics on utilization, maternal child and chronic conditions used at the national level and state levelb. PIP: Pediatric Asthma, Asthma Medication Ratio and Medication Management, Diabetes and Behavioral Health developed for different Plans/Statea. HEDIS methodology applied and logic developed according to regulatory requirements.
c. P4P: LANE, PPA and Readmits internal reports developed and processed monthly. Data applied to internal process improvements and outreach programs. Rules maintained, modified and updated according to District of Columbia regulatory requirements.
d. Population Analysis conducted on a variety of demographic segments.
4. Analytics (chronic conditions) using statistical algorithms to identify high risk/potential high cost members with chronic conditions who will benefit from an intervention
a. Applied to members with Diabetes, COPD and Heart Failure
b. Data set produced on a monthly basis
c. Methodology is scalable to other plans
d. Analytics employs lab data, medication adherence data
5. Medication Adherence - advanced logic created to calculate medication adherence at the member level.
At the member level we can know what the pdc (proportion of days covered) rates are by different chronic conditions.
6. COC Ideation: new ideas for cost of care are being continuously generated.
a. Asthma: $120 K savings
b. Diabetes : $300K estimated
c. Heart Failure: $1 million
d. Sickle Cell: $600 K realized in 2018. This will be scaled to get more savings
e. Ideas are scalable across plans
7. Disease DataMart with clinical outcomes and lab data.
Key clinical metrics were identified and pulled from claims.
Based on the data set that was created, we have created analysis and insights that has helped to create one of its kind diabetes intervention program that potentially could be replicated across the different plans and states that the MCO is in.
8. Geospatial analysis:
we have developed custom code to calculate distance between members and providers.This utilizes geo coordinates. Based on user requirements, we also display data on Tableau.
This has been used for value based program to identify closest radiology center/lab to the member. For different radiology center, clusters of members based on a radial distance from the radiology center have been auto generated.
9. Operational Tools Created
Population registry user defined reportPCP/HOSPITAL Report Card
PCP Trend Report
Case Management Report Card
Health Risk Assessment Database
HCMS/Inpatient Dash-Boards
User defined reports by code/provider/group/TIN
Case Management Watch Listing