988 Suicide & Crisis Lifeline: Independent Evaluation & Predictive Modeling

In 2024, we were retained to act as the independent evaluator for the 988 Suicide & Crisis Lifeline program implementation in the State of Oregon. Operating under a federal grant subcontracted through the Cornerstone Whole Healthcare Organization (C-WHO), our engineering and analysis team plays a pivotal role in reviewing the state's operational data processing, improving data quality, and modeling call center resources.

Independent Evaluation & CMHP Technical Analysis

Our evaluation scope includes providing suggestions and technical analysis of Community Mental Health Programs (CMHPs) across Oregon. We collaborate directly with two primary crisis call centers—Northwest Human Services and Lines for Life—alongside the state itself via the Oregon Health Authority (OHA). By conducting technical analysis of how these agencies collect, manage, and process call data, we assist them in structuring telemetry to meet federal reporting obligations.

"By bridging the gap between local clinical operators and state analytics infrastructure, we help build an integrated public health reporting ecosystem."

Developing Custom Automated Data Pipelines

A major bottleneck in state-level compliance is the manual effort required to aggregate and submit reports. We are developing secure, automated data pipelines designed to ingest crisis transmission events from call centers and transmit them to the state's central registry. Due to strict NDAs surrounding HIPAA regulations, public health data handling, and proprietary network topology, specific architectural blueprints are confidential. However, the system leverages secure modern protocols, automated ETL pipelines, and robust validation checks to verify data integrity before submission.

Predictive Call Volume Forecasting Models

To assist call centers in operational scheduling, our team engineered predictive call volume models. By analyzing historical trends, calendar markers, and public health indicators, these machine learning models generate forecasts of call, text, and chat volume. This allows center directors to align clinician staffing with expected call volumes, reducing hold times during critical hours. Our current efforts focus on wrapping these models into production pipelines, with the goal of deploying live, real-time sync pipelines before the close of the grant in late 2026.