AI-based readmission prediction

Customer Stories

Integrated Care Solutions (ICS) – Dubai-based digital health company focused on enhancing patient outcomes and care coordination.

Healthcare

Industry

Dubai

Location

AI Software Development, Predictive Analytics, MLOps & Model Deployment, Frontend & Backend Engineering

Services

$180,000

Budget

Challenge

ICS needed an automated, explainable, and integrated solution to identify at-risk patients in real time and support personalized interventions.

See what we can do for you

Outcomes We Deliver

Solution

We developed and deployed an AI-based readmission prediction service integrated into ICS’s clinical platform. The system uses 6 years of patient history and real-time episode data to assign readmission risk scores at 3 stages of the care journey.

Let’s talk about what’s possible

Tech Stack

To developed and deployed an AI-based readmission prediction service, Blackthorn AI applied a production-grade tech stack including:

Python

TypeScript

XGBoost

Docker

Azure

SQL

Pandas

Python

TypeScript

XGBoost

Docker

Azure

SQL

Pandas

Roadmap

Project duration

01-02 Weeks

Goal Alignment & Data Audit

Validated business objectives with clinical stakeholders, conducted high-level audit of available patient records, defined initial scope and KPIs for prediction accuracy and response time.

03-04 Weeks

Data Infrastructure & Compliance Prep

Mapped data schemas across historical and real-time sources (EHR, episode data, notes), established secure access protocols, and ensured compliance with privacy and clinical data regulations.

05-06 Weeks

Exploratory Analysis & Feature Engineering

Performed in-depth data profiling, identified key predictors of readmission, handled imbalanced classes, and engineered structured + time-series features from up to 6 years of data.

07-08 Weeks

ML Modeling & Risk Stratification Logic

Developed and tuned predictive models using XGBoost and logistic regression, implemented logic to assign patients into three actionable risk categories based on outcome likelihood.

09-10 Weeks

Model Evaluation & Clinical Validation

Benchmarked model performance (accuracy, precision, recall), tested outputs with ICS care teams, and built early explainability features for nurse-facing decisions.

11–13 Weeks

MLOps & API Architecture

Deployed models in a containerized pipeline (CI/CD), designed a standalone prediction API with retraining triggers, enabled secure integration into ICS backend and web dashboard.

14–16 Weeks

Frontend Integration & Release

Integrated model output into production frontend, finalized UX for nurses and clinicians, automated model training with new data, and released live MVP for evaluation across active patient flows.

01-02 Weeks

Goal Alignment & Data Audit

Validated business objectives with clinical stakeholders, conducted high-level audit of available patient records, defined initial scope and KPIs for prediction accuracy and response time.

03-04 Weeks

Data Infrastructure & Compliance Prep

Mapped data schemas across historical and real-time sources (EHR, episode data, notes), established secure access protocols, and ensured compliance with privacy and clinical data regulations.

05-06 Weeks

Exploratory Analysis & Feature Engineering

Performed in-depth data profiling, identified key predictors of readmission, handled imbalanced classes, and engineered structured + time-series features from up to 6 years of data.

07-08 Weeks

ML Modeling & Risk Stratification Logic

Developed and tuned predictive models using XGBoost and logistic regression, implemented logic to assign patients into three actionable risk categories based on outcome likelihood.

09-10 Weeks

Model Evaluation & Clinical Validation

Benchmarked model performance (accuracy, precision, recall), tested outputs with ICS care teams, and built early explainability features for nurse-facing decisions.

11–13 Weeks

MLOps & API Architecture

Deployed models in a containerized pipeline (CI/CD), designed a standalone prediction API with retraining triggers, enabled secure integration into ICS backend and web dashboard.

14–16 Weeks

Frontend Integration & Release

Integrated model output into production frontend, finalized UX for nurses and clinicians, automated model training with new data, and released live MVP for evaluation across active patient flows.

Team Size

1 x AI Solutions Architect

1 x Data Scientist

1 x Data Engineer

1 x MLOps Engineer

1 x Frontend & Backend Developers

5 team members from Blackthorn.ai

1 x AI Solutions Architect

1 x Data Scientist

1 x Data Engineer

1 x MLOps Engineer

1 x Frontend & Backend Developers

Delivering Impact

Beyond the values already highlighted, there’s even more to discover. Our commitment to innovation, client success, and impactful results sets us apart

Book a Meeting

32%

Drop

In readmissions within 90 days of pilot rollout

$8M

Savings

During 3-month observation period

84%

Of flagged high-risk patients

Actually readmitted – enabling targeted care

32%

Drop

In readmissions within 90 days of pilot rollout

$8M

Savings

During 3-month observation period

84%

Of flagged high-risk patients

Actually readmitted – enabling targeted care

AI-Based Readmission Prediction Service

Integrated Care Solutions (ICS) – Dubai-based digital health company focused on enhancing patient outcomes and care coordination.

Healthcare

Dubai

AI Software Development, Predictive Analytics, MLOps & Model Deployment, Frontend & Backend Engineering

$180,000

ICS needed an automated, explainable, and integrated solution to identify at-risk patients in real time and support personalized interventions.

Solution

To developed and deployed an AI-based readmission prediction service, Blackthorn AI applied a production-grade tech stack including:

Project duration

01-02 Weeks

03-04 Weeks

05-06 Weeks

07-08 Weeks

09-10 Weeks

11–13 Weeks

14–16 Weeks

01-02 Weeks

03-04 Weeks

05-06 Weeks

07-08 Weeks

09-10 Weeks

11–13 Weeks

14–16 Weeks

Team Size

Delivering Impact

32%

$8M

84%

32%

$8M

84%

Discover More Related Customer Stories