top of page
Post: Blog2_Post

Medicine shortages: what the data can reveal before official reporting

12 hours ago
4 min read
MySerial AI, a solution dedicated to MVOs, makes it possible to identify 30 out of 33 drug shortages detected.

A medicine shortage has an official reporting date. The data may tell a story that starts much earlier.


Falling dispensing volumes, an unusual change in quantities placed on the market, or supply pressures affecting the same medicine in several countries: when brought together and assessed in context, these signals can reveal a risk worth investigating.


For Medicines Verification Organisations (MVOs), the aim is to help authorities identify situations that need closer monitoring sooner. The central question is which data can genuinely provide earlier insight into emerging supply pressures.



Official reporting is a reference point, not the beginning of the story


To assess a detection method, its findings need to be compared against an identifiable event. The official reporting of a shortage provides that reference point.


However, three dates must be distinguished: when supply pressures first emerge, when they are detected in the data, and when the authorities officially report the shortage. These events do not necessarily occur at the same time.


Detecting a signal before official reporting does not necessarily mean detecting it before the shortage actually begins. This distinction is essential when interpreting the results of an early detection tool.


Even so, it offers a practical opportunity: to investigate a situation that may warrant closer monitoring before the shortage is officially reported.



A fall in volumes needs an explanation


A decline in dispensing volumes may indicate a supply problem. It may also reflect seasonal variation, a change in demand or a shift towards another medicine.


Consider an illustrative example: dispensing volumes for a medicine fall over several weeks. On its own, this observation is not enough to establish that a shortage is occurring.


The analysis becomes more informative when the wider context is considered.

  • Has the same decline occurred at this time of year before?

  • Have volumes placed on the market also changed?

  • Are dispensing volumes rising for competing medicines?


These comparisons help teams develop hypotheses and identify situations that warrant investigation.


Historical data plays a crucial role. A few weeks of data reveal a recent change. Several years allow that change to be assessed against longer-term patterns, including seasonality.



Why examine the same medicine across European countries?


MySerial AI, a solution dedicated to MVOs, makes it possible to identify 30 out of 33 drug shortages detected.

An analysis confined to one country leaves an important question unanswered: is the pressure local, or are similar signs emerging elsewhere?


Comparing the situation for the same medicine across European countries provides additional context. An isolated change and simultaneous supply pressures across several markets call for different lines of investigation.


This international perspective can help teams decide where to look more closely. It does, however, require comparable data. Reporting periods, pack sizes, units of measurement and source coverage all need to be considered.


A difference between countries becomes useful when it is clear what is being compared. Otherwise, differences in data collection or presentation may be mistaken for differences in medicine availability.



MySerial AI: an initial retrospective analysis in France


Developed by Design Data, MySerial AI is a solution exclusively for MVOs, designed to identify early signs of pressure on medicine supplies.


It combines serialisation data with European and international sources. The analysis considers up to five years of dispensing history, volumes placed on the market and trends in competing medicines. It also allows teams to examine the situation for the same medicine in other European countries.


France was the initial setting for the analysis. The work draws on serialisation data from France MVO, reimbursement data from Ameli, the French national health insurance service, and shortage information from ANSM, the French medicines agency.


In this retrospective analysis:

  • 30 out of 33 breaks were identified before their official reporting, with at least two signals detected;

  • the findings showed a lead time of around 100 days before official reporting;

  • this lead time reached up to 144 days when three signals were present.


These findings relate to the French dataset analysed. They measure the time before official reporting by ANSM, rather than before the actual onset of a shortage. They do not guarantee the same performance for every medicine or country.


The first operational trials in France are being prepared.



Turning detection into effective monitoring


MySerial AI, a solution dedicated to MVOs, makes it possible to identify 30 out of 33 drug shortages detected.

A retrospective analysis can establish whether signals were present before known shortages. Operational use raises a further question: how should teams respond to signals emerging today, when the outcome is still unknown?


The number of shortages identified must therefore be considered alongside the number of situations flagged, the time needed to investigate them and the quality of the available information.


Missing data, for example, must not be treated as an absence of dispensing activity. Limitations in a source’s coverage must also remain visible when interpreting the results.


For teams to act on an alert, they need to understand what triggered it, examine the underlying data and decide what to do next.


The tool’s role is to guide that analysis. Assessing the signal and deciding on the appropriate response remain essential.



Extending the approach to other European MVOs


The initial French findings provide a basis for exploring the approach with other MVOs. Applying it elsewhere requires an assessment of the data accessible in each country, the complementary sources available and the conditions governing their use.


Confidentiality is integral to this work. MySerial AI relies on restricted-access data and is not offered to pharmaceutical companies. Public findings are presented in aggregate form, without disclosing the underlying data.


For an MVO, the starting point is to identify the information already available and assess how combining it could support national monitoring.


The aim is to give authorities more time to understand emerging supply pressures and prepare their response.


Do you represent a European MVO?

Explore the findings from France and speak to Design Data about the potential for using this approach in your country.





Design Data Consulting, Consulting & Intelligent Management Dashboards to Simplify Decision-Making.

 
 
bottom of page