AI Models in Clinical Research

Target group

• Healthcare, incl. research, SMEs, start-ups, health insurance companies, regulatory authorities

Your requirements

You need support on:

• Developing or validating AI models in clinical research
• Planning a clinical trial to evaluate AI models
• Integration of AI models in medical care

Our offer

We offer support in the development of suitable AI models for research with medical data. In addition, this service aims to successfully integrate existing AI prediction models into routine care. For an adequate transfer of AI prediction models into medical care, models must first be externally validated in the respective setting, calibrated if necessary and evaluated with regard to their benefit and harm (impact analyses). The means of choice for this can be suitable real-world data or prospective clinical trials.

For clinical trials, methodological planning or consultation is a decisive factor for success and quality. The planning includes the choice of parameters, endpoints, suitable (adaptive) study designs, evaluation methods, sample size and power calculation, as well as the choice of suitable database solutions for the collection of study data.

Our services include:

• Support in the development of AI models in clinical research
• Support in evidence synthesis to identify suitable models for a defined prediction problem
• Support in external validation and calibration of existing AI models (e.g. external validation based on real-world data) for the respective application context
• Support in methodological planning of clinical trials (e.g. sample size calculation, power calculation, selection of appropriate endpoints)
• Advice on DSGVO and GCP-compliant database solutions in clinical trials
• Support in the integration of AI models in medical care

Requirements

• Occupation in the health sector
• Initial project idea

Success stories

• https://doi.org/10.1093/ehjdh/ztad021
  

• https://doi.org/10.3390/cancers12103010
  

• https://doi.org/10.1093/pm/pnx242
  

• https://doi.org/10.1093/ehjdh/ztab054