Michael JohnsonComputational models of individual patients will be the focus of a joint study by Microsoft and Boehringer Ingelheim. The companies announced the collaboration on October 22, 2026: the first area of focus will be lung health, and the work will center on assessing scientific and technical feasibility.
The partners define a digital twin as a computational representation of a patient that uses clinical and other data to model disease progression and possible outcomes under different scenarios. The research program will cover disease-course modeling, outcome prediction, and clinical trial simulation. This describes planned work: the company announcement set out research goals, not a finished system or results from trials involving patients.
Why model a control group
One idea being assessed is whether these models could support synthetic or hybrid control groups. In a conventional trial, a control group helps compare a treatment’s results with an alternative scenario. A synthetic control group uses existing data; a hybrid group combines such data with data from a traditional control group. If the approach produces reliable results, it could potentially reduce the number of participants needed for the control group and help with trial planning. The partners say they plan to study this possibility; they are not claiming it has already been achieved.
For clinical research developers, a practical point of reference is the quality of the evidence underpinning the model: how comparable the data are to the target population, how differences between predictions and observed outcomes are assessed, and what checks are needed for a specific disease. Microsoft and Boehringer said they plan to use modeling, prototyping, and evidence generation across several disease and therapy areas, starting with lung health.
Feasibility first, then application
In the partnership announcement, Boehringer Ingelheim’s Chief Medical Officer, Lykke Hinsch Gylvin, specifically stressed the need to maintain scientific quality, patient safety, and regulatory compliance. These conditions will frame future evaluation: a digital twin must demonstrate its value for a specific clinical scenario and undergo appropriate validation before its findings can influence trial design.
The collaboration also includes the modernization of data centers and security systems. The companies are considering potential joint projects in animal health, business processes, and technology infrastructure; details of these areas have not yet been disclosed.
At this stage, the outcome is a research program, not a replacement for standard clinical trials. The project’s value to the industry will depend on whether modeling based on clinical data can produce sufficiently verifiable estimates to improve trial design while maintaining safety and evidentiary standards.