The company, FARENTA, specializes in coordinating pharmaceutical companies with pharmacies in their provision of services related to drug development, marketing and distribution.

In 2017, Finland changed the rules for refunds on anti-diabetic medicines. It was expected that this would influence the market for drugs. To track changes, FARENTA, in cooperation with the University of Eastern Finland and the Endocrinology department in a hospital in Helsinki, launched a longitudinal study of patients with type II diabetes. This tracked changes in the patterns of use of these drugs, underlying explanations and the effects. The study was run in three waves: winter at the turn of 2016/2017, mid 2017 and the end of 2017.

Data were collected in extensive surveys. Analysis of complex structured surveys, especially longitudinal studies is well known for its difficulty. FARENTA collaborated with Transition Technologies:

  • to automate the creation of result presentations from successive waves of the study
  • to create an interactive application for analysis purposes.


Initially, results were analysed using the Office suite and then transferred into the presentation. The problem with this approach was changing data: some patients applied their rights to personal data protection and withdrew from the study, requesting deletion of their data. Using manual analysis and presentation, all tables and graphs had to be repeated when patients opted out. Imperfect data also needed to be tracked down and corrected, involving the recalculation of results.

Transition Technologies converted data to formats recognized by standardly available statistical packages. The master presentation was transformed into an R script that would generate the data presentation when run. The conversion of the entire presentation from entirely new data to a complete presentation then took no longer than a few seconds! An additional benefit from the change of preparation method was the implementation of interactive tables and charts.

The application allows:

  • sequential analysis of respondent responses to any question in the cross-section of the three waves of survey
  • analysis of therapeutic decisions made by users of arbitrary types of medicine: medicaments from a particular manufacturer or specific drug (What decisions did insulin users make? Which choice are they using now?)
  • analysis of the consumer behaviour of patients that qualify for 100% reimbursement of the drug costs – „the Kela threshold“
  • analysis of the multiple occurrence of medicines in patients’ homes – Which medicines are purchased together, what changes were observed between waves of the survey?
  • analysis of results of treatment – Were patients impressed? How was their weight and blood glucose?
  • export of reports to PDF, Excel, CSV and forprinters

Holding the mouse over a chart element, as shown, displays additional information, which could not otherwise be seen on an ordinary graph. When appropriate, it is also possible to enable/disable visibility of chart elements, to focus on the most relevant data. Each chart, without any loss of quality, can be zoomed, cropped and saved to the computer drive for a record that can be used later.

R algorithms also allowed several waves of the study to be explored and dozens of drugs checked according to uptake in prior waves. Using the Office package, this procedure was repeated for the list of drugs, making it a long and tedious procedure. The automated procedure was very quick and much less prone to error.

An automated procedure was also applied to statistical tests, which generated summaries of results and interpretations. These were based on observed differences with respect to thresholds for statistical significance.