
How data can impact pharma educational focus from Japan to the world

The journey that any single treatment option takes from germinal idea to positive clinical outcome is both lengthy and expensive. Clinicians, health practitioners of all types, and pharmaceutical companies know that designing effective treatments for individual patients has to be done based on empirical evidence of effectiveness.
In an age where every medical intervention and clinical touchpoint is documented digitally, it’s easy to imagine that one day, every treatment could be based on personal medical history considered against every possible treatment option. Thankfully, that day is getting closer, but it’s not here yet. Its arrival has been hastened by a data partnership between Japanese company MDV and Prospection, an Australian real-world data analytics outfit featured several times on these pages to date.
Prospection’s unique offering in the sector is that it can take longitudinal patient data (information that spans a significant portion of an individual’s life) and, for instance, cross-reference treatments based on similar pathologies to discover the best drugs, dosages, and other options to achieve the best patient outcomes.
Its data agreement with MDV was the foremost subject when we spoke to Greg Hughes, Head of Japan and APAC for Prospection. Greg told us that the MDV role is particularly significant, given that the company holds detailed medical histories for some 40 million Japanese patients drawn from the length of the country. With Prospection’s specifically-engineered data processing capabilities and MDV’s comprehensive data sets, clinicians and pharmaceutical companies are beginning to get significant insights into treatments for multiple conditions on a local, national and per-condition basis.
When the Japanese data set from MDV is combined with Prospection’s extensive data resources drawn from right across the world, the algorithms’ effectiveness is enhanced. That refinement and ever-expanding data amalgam improves quality of outputs with every iteration and addition: “The key is longitudinal data. We’re tracking patients over a 5-10 year period. The best longitudinal data is where you’re guaranteed to see each interaction with the healthcare system, [but] those data sets are rare. Australia has one so is a very good country to train advanced predictive algorithms […] There’s a very good case for using algorithms developed in Australia in Japan [because it] has pathology data. So we have taken algorithms trained on complete data and for further trained those algorithms on MDV data, which is inclusive of pathology data to improve those algorithms, and then we can take them to the US where the data is quite fragmented, [where] harder to get guaranteed longitudinal data.”
However, if the data sets already exist, surely any pharmaceutical company can pick up the information and run with their own data science teams? It’s possible, Greg stated, but it’s a case of resourcing and costs. “Historically, a pharmaceutical company may have been buying a data extract from MDV. It takes them months to figure out how to transform that data into a meaningful data set to find the stages of a disease, to understand lines of therapy, etc, and it can take them months before they get to any meaningful insight. With [Prospection] they’re there within weeks,”
That turn of speed comes from a focus on discovering correlations between patients’ histories, treatments and outcomes developed over time and ‘written through’ Prospection’s algorithms. The company is “specific in that we serve the pharmaceutical industry,” Greg said. “We look at the patients on a particular medicine or a series of medicines for a disease. What characteristics drove a better outcome? Why do some patients with the same diagnosis, the same stage of disease, do better on drug one, when some patients do better on drug two?”
In the specific case of Japanese patient data from MDV, for example, companies will be able to focus down to the level of the prefecture to get the most effective results from their outreach programs. Patient outcomes will be improved where needed most, and the data set’s invaluable insights will drive an overall clinical improvement across various conditions in people’s overall well-being.
At the time of writing, it was, Greg said, early days. “We’ve been looking at this regional data from MDV now, only for six weeks, because the deal [between MDV and Prospection] is relatively new. But in every therapy area looked at so far, we are seeing significant regional differences, and that can determine, in what region does the pharmaceutical company focus its educational activities. In some areas, you might have an adherence issue (in that patients aren’t adhering to the drug regime). Or you’ve got a situation where the patients that have the certain characteristics that indicate they’ll do better on a drug aren’t being prescribed [it]. So that again creates a very different form of activity in that region.”
The MDV data holds lengthy patient records for 40 million individuals from a total population of 120 million – an amount of information that is statistically highly reliable (the smaller a data set, the more likely anomalies are to be presented as trends). In combination with the dedicated data science of Prospection, healthcare’s prognosis in Australia, Japan, and further afield looks much more promising.
To learn more about the MDV and Prospection’s new partnership, there’s a detailed article here, and to find out more about the effectiveness of Prospection’s data transformations, click through to read more.