Medial Research Network
UK TEL: +44 (0)1908 261 153 | US TEL: +1 (847) 779 7857

Blog Posts

My thoughts from Pharma Integrates

The Pharma Integrates conference a couple of weeks ago gave a fascinating insight into the broader industry view of the Patient Centric movement and its intertwined relationship with ‘big data’.

For me, the main pharma centric insights came around the concept of using the large amounts of data we can gather on ourselves as patients to:

  • Help create more sophisticated patient and illness classifications for better prognostic predictions
  • Help personalise the medical approach to our illnesses
  • Help understand how to motive patients to engage with their therapies
  • Help us understand the meaning and value of real world data.

Patient And Illness Classifications
Medicine spends a lot of time and effort examining patient behaviour empirically – simply through observation. Clinical outcomes are descriptions of the effect a condition has on patients’ ability to conduct their lives in a normal fashion. Our treatments are designed on a risk vs benefit assessment; allowing us to refine the interventions to be best suited per patient. Additionally, we classify the sub-types of conditions based on observed criteria to correlate with different outcomes. This helps us predict benefits for patients. The advent of huge volumes of data about people through social media and computerised equipment, now owned by billions of people in the form of smart phones and wearable devices such as smart watches, coupled with descriptions of their illness and the outcomes they record, should allow us to generate much better observations of patient and illness classifications, to hugely enhance our prognostic information for patients to determine their best therapies.

Personalised Medicine
Personalised medicine is of course an old concept – it has been around for decades. The message has always been two fold – first we want to know how to better choose therapies for patients to reduce risk and increase benefits, and second we need to ensure we understand what patients actually want out of their therapy, and target those particular outcomes. It became clear in several of the debates that our ability to harness the huge volumes of data will allow us to move much closer to both of these objectives. In the first instance it will help us better predict which patients should be offered the more aggressive therapies and which patients should not be offered therapies that have little chance of working, based on the approach outlined above, and second it will allow us to collect much more data that the people with the condition themselves feel is important – a kind of ‘billion patient’ focus group. There was a reasonably clear patient voice in the room driving the industry to these outcomes – but I missed the presence of any medical educationalists who need to be addressing the education of our physicians to understand and use this new data for the holistic treatment of patients.

Patient Engagement With Their Condition
A well represented element on the discussion panels were the professionals who spend their time determining how to motivate patients – both academics and commercial researchers and solution providers, who were describing how we can use the proliferation of this data to predict not just what the prognosis might be, but also how that will be affected by the mental state of the patient, and their engagement with their condition and how to treat it and – perhaps the most innovative thing I heard – how healthcare professionals can change the engagement status of the patients to give the therapies the best chance of success – again making me feel some medical educationalists were missing. In truth, there were almost no physicians there who are in medical practice, which was a shame as they could have added a significant dimension to the debate as they are also committed to patient specific therapies, holistic approaches to treating a person’s condition, and listening to what the patient wants to achieve from their therapies.

Real World Data And Clinical Trial Implications
This is where the empirical nature of medicine and the evidence based approach to medical interventions starts to separate in this story. To determine if interventions are worth doing we need to have evidence they have an impact on the target condition. Just observing correlations does not cut it. If we have huge volumes of data over which we run machine learning algorithms to search for correlations we know for certain that due to the randomness of data we WILL see correlations that are false positives. It was refreshing to hear clinical scientists saying that it is therefore not enough to create these correlations – we then have to move from this hypothesis generation process to a hypothesis testing process – from real world data to blinded, controlled clinical trials. Not only that, but we should do so only when there is a clear reason to believe there is a possible causal link between the intervention and the outcome.

The implications in my world – the clinical trial world – is that clinical trial designs might change to incorporate smaller control groups, that patient classification might become much more critical, that the types of data we collect might expand to be more holistic and more focused on the outcomes patients are interested in, that wearable and smart phone tech might be more useful in collecting data, but some things will remain the same as the bedrock of the science we are running – devices still need to be accurate and validated (and many are not), data still needs to have low variance, the data we collect has to be relevant to the condition we are treating and should ideally be collected prospectively in a blinded fashion where patients are randomised to therapies to reduce bias. And we still need to use carefully chosen patient groups with inclusion and exclusion criteria to first see if (in a clean population) the drug can work so that we can further refine that message through real world data to see if we can find as many situations as possible to see that its effect is valuable overall.


Graham Wylie


Posted by:
Zara Broadfield
Share this post!