The good thing about SCDM conferences is that it is a conference focused in the data management and sciences space which gives greater content depth and is reflective of the speakers and panelists who have a great deal of hands-on experience and eagerness to share. The program had topics that covered current practices as well as emerging models like analytics and risk based monitoring. It was also pleasant to catch-up with old friends in the industry as well as forming new. I had been to most of the sessions which I think were good. I had taken mental notes of points that stood out to me during these sessions as well as during interactions with industry leaders. This helped me sketch my own perspective, some of which are summarized below:
Don’t forget the “Bio” in Bioinformatics: This statement resonated well with me. Given, how data management space has specialized over recent years in pursuit of efficiency, focus and adherence to standardized business processes, it would serve us well to remember that a holistic understanding of the science behind the data, its objective and impact would really make a difference to the way we support the science of disease management and drug discovery. A moving testimonial read out by Demetris Zambas in the beginning of the conference of a parent whose child was cured from Acute Myeloid Lukemia (AML) by a drug from Novartis is a reminder of the impact that we generate in human lives and the importance of keeping science behind the data central to the work that we do.
Influence of technology on patient behaviors: Giant strides in technology innovation is making new technologies forays into the clinical research space, especially those that directly impact subjects. An important aspect to be cognizant of is that many of technology tools that are being provided to patients can also drive positive behaviors in illness management by patients during the trial. This in turn could significantly increase placebo effect. A good example being utilization of wearable gadgets to track exercise levels in diabetic patients. Understanding how that impacts the study design and outcome is important before jumping into a decision.
Predictive analytics may not be accurate always: With exploding innovation in analytics, especially in the space of machine learning and predictive modeling, the clinical research industry is also in hot pursuit to achieve accurate predictions in many spheres of disease management and clinical research conduct. Parallels of how machine learning and prediction is successful in the aviation industry and entertainment industry was insightful. However, the failings of analytics are a lesson that teaches us that technology needs to be contextualized to an industry and utilized. The example of Google not being able to predict flu outbreak in 2013 based on past data and algorithms exposes a latent aspect in the healthcare industry. The variables that we choose for modeling and algorithms as well as the completeness of understanding the biological mechanisms and pathways is important for getting the predictive modeling accurate. Having said that it will be a matter of time before Google gets it right. The difference for our industry though is that we need to get it right the first time.
Reducing cost of clinical development: In the Leadership forum at the end of Day 1, the panel discussion deliberated on the need for clinical development costs to be reduced and what needs to go into that from a data management and operations viewpoints. Interesting perspectives emanated from the panel as well as the audience. I believe that there is indeed an opportunity to improve efficiency in the processes within data operations. However, I also believe that the probability of success of a drug discovery remains elusive. This renders opportunity costs to be loaded into the cost of the medicine. Not only a disruption in the model of pursuing drug discovery is important but also new business partnership models need to emerge that can absorb the shock of failure of drug discovery.
Who is better suited to monitor risks in the risk based monitoring model?
Interesting viewpoints were debated in the panel session that I participated. While it was unanimously agreed that implementing RBM required all functional stakeholder to collaborate, the question that still lingers is who is better equipped to review and monitor risks. Though it can be eloquently debated both ways – CRA or a data manager, I articulated that basis of decision will depend on multiple factors such as size of the organization, business appetite to implement it in a full-scale model or a hybrid model, availability of analytics-skilled talent or ability to upskill existing talent in operations. Hence, a cookie-cutter approach needs to be avoided and the solution needs to be tailored to the organizational situation and assessment of teams. Of course, in the interim, the time is right for the industry to commence the transition by initiating RBM training, enablement sessions and familiarizing the use of best of class analytics tool within teams so that it will help the individual and the industry at large.