Technology and regulation updates.

A practical weekly update for pharmaceutical and life-sciences companies.

Latest updates

4 reviewed updates

Regulation

EMA opens early feedback route for non-animal test data

The EU medicines agency, EMA, has opened a pilot for new ways to test drugs without animals. Some methods use tissue grown in a lab. Others use computer models. These models may include AI to predict how a drug could act before tests in people. The pilot covers other methods too, but it gives AI-based tests a route into early talks with regulators.

Drug firms, labs, and method makers can apply. EMA will choose some applicants to send a full set of data. They can then meet experts from across Europe. The experts will point out strong parts and gaps in the data.

The advice can help a firm plan its next study and see what proof it still needs. EMA will make later decisions through its normal review. Entries stay open until the third quarter of 2027. EMA pilot announcement

Our thoughts.

For a company building an AI-based test, our advice is to use this route early. Ask about one live method before funding the next study or planning its use in drug development.

Technology

IQVIA connects AI to trial planning, recruitment, and data cleaning

IQVIA has launched an AI service for several parts of a drug trial. It can help plan a study, open trial sites, find patients, and clean data.

IQVIA says it helped start studies 33% faster and clean data 50% faster. It also says more patients joined and gaps between trial phases fell. The release does not explain the study plan or sample size behind these figures. Sponsors should use them as a guide for what to test.

The main value is one service across tasks that often sit in separate systems. A sponsor could use it to choose sites and plan recruitment. It could also help find data issues sooner and lock the final trial database with fewer delays. IQVIA press release

Our thoughts.

We see the strongest use where delays build up across several trial handoffs. Start with one study and set a baseline for startup, recruitment, and data cleaning. Then test the full path before wider use.

Technology

AWS adds a score for each record match

Amazon Web Services (AWS) runs cloud tools for companies. It added a score to each match made by its Entity Resolution service. The service checks if two records may describe the same person or group. A record is one entry in a data set.

For example, a drug company may have the same doctor in an event list and its contact system. A high score suggests both entries point to that doctor. A low score can send the match to a person for review.

Before this update, each record in a matched group had the same score. Data managers can now set a review level for each match. AWS says the feature is ready now. Its notice gives no test results for drug company data. AWS product announcement

Our thoughts.

We would start with a data set the company already knows well. Count wrong and missed matches before choosing when a person must review them.

Technology

Servier adds an AI tool and patient data to cancer research

Servier will use Owkin's K Pro AI tool and patient data to study cancer. The deal gives its researchers several types of health records from the MOSAIC network.

Servier plans to use the tool to study patient groups and possible drug targets. It may help researchers ask better questions, test ideas sooner, and choose which leads deserve more study. It may also help them link signs that sit in separate records.

Owkin's release confirms the deal and its planned use. It reports no research results yet. The main opportunity is to find useful patterns across several types of patient data. Those patterns could point to a smaller patient group or a new line of research. Owkin announcement

Our thoughts.

In the future, this data could help researchers focus on the strongest leads. Our advice is to agree how each finding will be tested and moved into the next study.