Data Analytics for Pharma Development 2018

John Mulcahy spoke at Data Analytics for Pharma Development 2018 in Amsterdam on the 12th September 2018 about Leveraging Digital Health to transform patient care

The conference had a number of themes that played out across the speakers:

  1. The economic and clinical drivers of pharmaceutical development.
    Many of the presentations looked at the challenge of developing new drug products that are more effective and are safe/tolerable, because many of the best drug candidates have already been identified, developed and exploited leaving less scope for clinical improvement from new drugs.
    StandardOfCare-PharmaROIfromNewNME-withoutRoI-noredline

    At the same time, the cost of developing those new drugs has been rising significantly driven mostly by costs of drugs that fail to meet their safety/efficacy targets (since the scope for improvement from new drugs products is lower) and increases in cost of performing clinical trials. and then proposed:
  2. Data analytics tools and techniques to support development of new pharma product to optimize the identification of candidate drug products:
    Such as https://Siren.io : platform for performing investigation of potential targets/molecules across big, complex and partially unknown data sets & graphing the linkages.
  3. Approaches that organizations and consortia are adopting to enable data curation and sharing so that we have data of sufficient quality and availability so that data analytics can enable development of new pharma product and improve care. Real World Data is uncontrolled, and of varying quality.
    ClinicalTrialData-vs-RWD

    There are multi-national and industry initiatives looking to address the challenge of incompatible data sources by mapping the data into an OMOP common data common data model and enabling the data to be used. These initiatives build on each other:
    1. EHR4CR Electronic Health Records for Clinical Research which ran from 2011-2016
    2. European Medical Information Framework, EMIF, which has a mission to provide one platform for data discovery, assessment and (re)use.
      Identifying predictors of metabolic complications in obesity, and Identifying predictors of Alzheimer’s Disease in the pre-clinical and prodromal phase are EMIF’s first two focus application areas.
    3. The European Health Data & Evidence Network, EHDEN, will start in late 2018 with an aim to map 100 million health records across the EU via a common data model (OMOP), supporting research, the BD4BO IMI2 programme, and outcomes-based healthcare
  4. The need for new and non-traditional data sources in order to be able to answer the key questions. Many speakers talked about the need for patient level data (including that coming from consumer wearables or medical devices).
    The conference opened with the challenges of using DNA or RNA for drug development, or as a means to identify which people will respond to a drug and whether or not the patient will experience adverse events both of which would be important for precision medicine and targeting treatments (and potentially
    reimbursement and access) to patients who will benefit from a treatment.
StandardOfCare-PharmaROIfromNewNME-withoutRoI-redline

In HealthGenuity’s presentation I looked at the question of pharma development from the patient perspective, showing the actual current real-world standard of care that patients experience, and why this is often significantly lower than the standard of care possible with existing drug products. Bringing real-world care up to the standard of care possible will also be important as Pharma companies become more responsible for clinical outcomes as part of the move to value based care.

I then provided examples of how digital health and data analytics can help address this gap. Addressing the motivation and behaviour challenges that people experience for long term treatments, especially conditions where there are deferred consequences when a patient falls off therapy; and those where the best care pathway involves significant behaviour change.

Since each patient is different and that their engagement with treatment changes over time, this calls for solutions that enable us to provide personalized, tailored care for patients, to detect and understand when they are experiencing challenges or their circumstances have changed, and provide support before the patient relapses or experiences sub-standard therapeutic outcomes. Behaviour is established and maintained day-by-day. Trying to find out what happened 3 months or 6 months later and trying to get the person back on therapy at the next clinic visit may be too late if new habits have already formed. Therefore, it is crucial to detect issues and engage with patients in a timely manner, at the moment where they experience challenges before they start to fall out of the habit of following their care plan. This has been achieved by using digital health solutions that are tailored to the person’s needs with medical devices to monitor and track how each patient is performing over time; and then intervening to support patients at the time when they need it.

I see three key mid-term applications where data analytics can be coupled with digital health solutions to advance the state of care:

  1. Insights to enable Physicians keep up to date with latest research relevant to each patient:
    Automatic generation of clinical insights to present latest research relevant to each specific patient. Enabling physicians to practice at the top of their licence by keeping abreast of latest scientific developments while building a better therapeutic alliance with patients who will increasingly also be going online to understand their condition and options for care.
  2. Predictive Analytics
    As we build up (a lot) more data with more patients over time, Predictive analytics to be able to predict clinical outcomes for patients based on current patterns of behaviour compared to matched cohorts of similar patients in the past and could help to show the consequences if current behaviour is maintained and thus encourage existing healthy behaviour and/or adoption of desirable new changes in behaviour
  3. Analytic monitoring of performance against outcome-based contract terms
    Outcome based contracts have had limited adoption to date, often because of the hassle of setting them up and monitoring performance. Automatic self-monitoring of performance against outcome-based contracts using data analytics could address some of the issues, enabling a lower monitoring burden for payers/providers, and improved performance by company.

Leave a Reply

%d bloggers like this: