Analysis Listings × Treatment Pathways
Shows how patients move between lines of therapy as a Sankey diagram. Click a node to narrow the cohort, and check calendar trends and data quality on the same screen.
- Python
- FastAPI
- React
LIFE SCIENCE × HEALTHCARE × GENERATIVE AI
Beyond the proof of concept.
Drawing on clinical and data analysis experience, we build generative AI into healthcare and pharma work — from scoping requirements through development and operation.
01
WORKS
Work HerzLeben has delivered. For some of it, we have published how it was built.
Systematic Review × AI Subagents
A workflow that rebuilds the TrialMind paper's procedure with Claude Code: subagents handle PubMed search, abstract screening and full-text data extraction, and the results are checked with the paper's own metrics. The output can be explored in a sample report.
Claude CodeSubagentsPubMedPython
Medicare Part D × Text-to-SQL
An app for exploring 80.7 million rows of US prescription data in plain language, without writing SQL. Ask a question and Claude writes and runs the SQL on BigQuery, then returns maps, charts and an interpretation.
ClaudeBigQueryCloud RunPython
More delivered work
Clients and data are not named; screens are recreated with synthetic data
Claims Data × Prescription Gaps
Calculates the days between consecutive prescriptions for each patient and shows the distribution by drug, disease and period, including the share of gaps over a threshold.
- R
- React
- BigQuery
Claims Data × Co-prescription
Pick a disease and a target drug to see which drugs were prescribed on the same day, scored by Support, Confidence and Lift. Teams can explore it themselves without writing SQL.
- Python
- React
- BigQuery
Economic Evaluations × Auditable Models
A reviewer approves the assumptions a paper leaves unstated, and an R model is generated from the locked specification alone. It then shows which assumption explains any gap from the published result.
- R
- Shiny
- Generative AI
Cost-Effectiveness Model × Sensitivity Analysis
Runs a two-arm cost-effectiveness model in the browser. Switch survival distributions and review one-way, probabilistic and scenario analyses as tornado charts and acceptability curves.
- R
- Plumber
- React
Past Projects × Knowledge Search
Search past analysis plans and programs by chat to find definitions and analysis logic. Each answer shows the source passage and original file, easing the load of internal questions.
- Python
- Streamlit
- OpenAI
TLF × Automated layout checks
Reads TLF PDFs and flags pages with unusual blank space at the bottom or right, based on pixel density. The same screen also checks whether figure colours are hard to tell apart under different types of colour vision.
- R
- Plumber
- React
02
PROJECTS
We take on projects that put data and generative AI to work, mainly in life sciences and healthcare.
PHARMAPharmaceutical
Data & analytics
Data & analytics
Embedded in the team from analysis to generative AI, continuously building several apps. Replaced a purchased external service with in-house tools, and moved repetitive aggregation and first-draft documents onto generative AI.
DevelopmentEmbedded support
GOVERNANCEPharmaceutical
RWD governance
RWD governance
Joined the build-out of a foundation for sharing and using real-world data across the organization. Beyond selecting and configuring the tooling, designed the operating rules and ownership model — who maintains what — and stayed through to the point where teams actually use it.
FoundationsEmbedded support
ACADEMIAAcademic research
Large-scale data
Large-scale data
Ran the analysis for a study combining a large commercial health dataset with DPC claims data, turning the researchers’ “this kind of patient” into variable definitions across two datasets with different granularity and coding. The work carried through to an accepted paper.
Foundations
DATA VENDORHealthcare data vendor
Offer design
Offer design
Clarified what pharma buyers want, from time on the buying side and fresh interviews. Rebuilt the offering, the proposal structure and the PoC format — and it led to a signed deal.
Embedded support
03
WHAT WE DO
01
Analysis tools and business apps
We build the screens people use to work with medical data, from sorting out requirements to implementation. Not a one-off analysis report, but something the team touches every day.
We start from how it will be used — for example, clicking a chart to narrow the patient group and checking the distribution or trend on the spot.
Treatment pattern visualizationCo-occurrence analysisExploratory dashboardsAutomated reports
02
Automated visualization and reporting
We replace the aggregation, charting and slide-making that medical affairs and commercial teams repeat every month with a system that has generative AI built in. People only check whether the numbers hold up and how to read them.
From PoC design through to running it in production. Built around guardrails and human-in-the-loop, so output never flows through unchecked.
Automated recurring reportsAutomated dashboard refreshCollecting and triaging papers and trialsGuardrail designHuman-in-the-loop
03
Training
So that your team can keep growing the tools we deliver, we also hand over how they are built, through exercises on real data.
See TRAINING →04
CLIENTS
The people who most often come to us are in roles like these.
PHARMA / MEDTECH
Pharma &
medtech companies
01medtech companies
Data & analytics, medical affairs, commercial and digital transformation teams
TYPICAL ASKS
- Monthly reports take days to put together
- We want to use RWD but lack hands in-house
- Our generative AI PoCs never reach production
DATA VENDOR
Healthcare data
companies
02companies
Data vendors and health-IT companies looking to take their products to pharma
TYPICAL ASKS
- We need to reframe our pitch in pharma’s language
- We need a demo or test environment, fast
- We need someone who understands medical data
HOSPITAL
Hospitals
03Planning, medical billing and nursing management teams building in-house analytics or adopting generative AI
TYPICAL ASKS
- We want to use DPC claims data for management decisions
- We want to grow people who can analyze data in-house
- We want to stop tallying by hand
05 WHY
The field. The data.
The build.
HerzLeben draws on four settings: data analysis inside a pharmaceutical company, digital transformation for life-science clients at a consulting firm, product rollout at a medical device manufacturer, and clinical practice. From there, we build generative AI into the day-to-day work of life sciences and healthcare.
Depending on scale and domain, we team up with outside specialists — so one point of contact can carry a project from requirements through implementation to adoption.
06
TIPS
Notes on how we build, and what we learn along the way.
Claude Code × Medical Application
[Claude Code] Systematic Review × AI Subagents Part 4: Discussion and How to Reproduce
Oct 11, 2026
Claude Code × Medical Application
[Claude Code] Systematic Review × AI Subagents Part 3: Data Extraction and Full-Text Handoff
Oct 6, 2026
Claude Code × Medical Application
[Claude Code] Systematic Review × AI Subagents Part 2: Screening and the Subagent/Hook Split
Oct 5, 2026
07
CONTACT
We help life-science and healthcare teams put data and generative AI to work in practice.