Transforming life sciences with AI and automations


Dr Sameer Thapar, Global Pharmacovigilance Director, Oracle Health Sciences Consulting, talks about the need to deploy technology to improve efficiency in the lifesciences sector

Many previously labour intensive functions in healthcare and life sciences are now becoming commoditised with the aid of technology. Digitisation of data elements is allowing health professionals to utilise technology to assist in the management of activities. Deploying a database is the first step to raise the level of efficiency that a healthcare organisation experiences from baseline. Most solutions today have the means to raise the standards of efficiency within the organisation. This is done when workflow best practices and automation configurations are implemented in line with the deployment. The result of these implementations is to transform the organisation beyond the realm of industry best practices. There are new tools available to healthcare professionals to overcome the increasing demands that they face on resources and budgets in the management of routine activities, such as implementations of automation, use of intelligent bots and data mining tools. These are just few of the more commonly deployed from the list of known technology. All can immediately be leveraged to elevate the organisation from operating at an industry best level to a transformational one.

Automation is the buzzword of this year but in-depth understanding of the components that make up automated processes is not known well. Automation is simply applying rule-based algorithms to tasks. If a task is repetitive and non-changing, then it is a good candidate for automation. Automation is an excellent modality for data mining activities. Big Data has the challenge of discovery of pertinent elements. Up till recent advanced in big data mining using automation, it has been a laborious task to sieve through the records using various algorithms at each pass through. Automation is sometimes used interchangeably with Artificial Intelligence (AI) and this is not accurate.

As mentioned, automation is rule based whereas artificial intelligence is machine learning. AI is allowing the computer, database, or machine to process and deploy algorithms as it sees fit. The machine is ‘fed’ data and ‘learns’ from it. It then applies this learning to other similar tasks. If it experiences an outlier, a piece of anomalous data, it deploys a best fit for that, therefore an ‘intelligence’. According to Forrester Researchi, investment in Artificial Intelligence (AI) is expected to triple this year. Tech giants such as Google, Microsoft and Amazon are already tapping the potential of machine learning with inventions like Home, Cortana and Alexa.

In order for AI to operate well, it needs to mine data at an exponential rate. Large data sets in the trillions of bytes have to pass through the AI machine for it to grasp and formulate its rules and algorithms. For healthcare and life sciences domain, the need for real-time data mining is crucial. It is not advantageous in these domains to shift through data elements as a backlog activity. Actionable insight is always on elements being generated in the present. However, this has been constrained by technology and labour requirements. But if technology could be upgraded along with the near elimination of labour resources, we would have an automated data mining system. Such a system is in its early stages at the FDA.

The FDA Data Mining Council (DMC) was formed in 2007. The DMC is collaborative and explores methods and best practices recommended by experts from other federal agencies, industry, and academia—all of whom have analogous experience in knowledge discovery through various data mining approaches.ii
The council serves as a forum for FDA scientists to share their experiences and challenges in analysing data contained in the vast databases the FDA maintains, as well as to discuss new methods for such analyses. The FDA currently receives approximately two million adverse event, use error, and product complaint reports each year from consumers, healthcare professionals, manufacturers, and others. Since the early 1990’s, FDA has advocated data mining to the industry in an effort to better understand the signals within the safety data. Now, FDA data mining experts have expanded their attention to adding more sophisticated data mining methods and applying data mining to other types of product safety-related FDA and non-FDA databases.

The Proportional Reporting Ratio (PRR) is the foundational concept for many disproportionality methods. However, because this method does not adjust for small observed or expected numbers of reports of the product-event pair of interest, other more advanced statistical methods are employed, such as the Multi-Item Gamma Poisson Shrinker (MGPS), which produces Empirical Bayesian Geometric Mean (EGBM) scores iii. Several FDA Centres including CDER, CBER, and CFSAN, use the MGPS algorithm for their routine surveillance activities. Various commercially available software programs generate PRR and/or EBGM scores.

CDER has applied Empirica Study to analyse drug clinical trial data in either new drug applications or supplemental applications. Empirica study interfaces with data that conforms to the standardised Study Data Tabulation Model (SDTM) of the Clinical Data Interchange Standards Consortium (CDISC) data standards to create a wide set of automatically generated analytical outputs and tailor-made, reusable tables and graphs. These outputs have helped reviewers to more efficiently analyse potential safety issues in clinical trial data of drugs approved by the FDA. The FDA notes the benefits of data mining with these tools in the areas of standard processes (because data mining is automated, the outputs are statistically objective and devoid of manual analyses), simultaneous analysis (across an entire database at once), efficiency (analyses computed in minutes), and the benefits of automated signal investigations (transparency with audit trails, drill-down capability, observation of signals over time, and study of a product in populations). These tools are in the early stages of the automation efforts at the FDA.

However, the FDA operates in a regulated environment. Several pharmaceutical companies are now experimenting with the tools of automation and artificial intelligence, such as bots, which are outside the purview of current regulations. Chatbots, or bots as they are simply known, are assisting with routine tasks such as getting weather updates, flight messages, booking hotels, and now, answering simple health inquiries from physicians.

Merck-Sharpe-Dome (MSD) Italy has launched the MSD Salute Bot in 2016. The first focus has been the immuno-oncology and soon to come there will be many more areas. The Bot is running on Facebook Messenger. From the MSD prospective physicians are digital consumers looking for relevant information for their professional activity iv. Some key factors like the increase of media availability, mobile devices penetration and the decrease of time available, are resulting in a reduction of time spent navigating and searching on the web. Only available to registered physicians at this time, the Bot has interacted with hundreds during the 2016 winter holidays.

Bots are the first layer of accessible and deployable AI that encompasses natural language processing (NLP) on the general lexicon. Anyone can deploy a Bot via online free tools and building kits. As of now, Bots are the for the outreach activities and do not decipher well the complicated medical lexicons. Currently, the Bot, no matter how strong the AI or NLP element, could not replace the human interaction for in-depth conversations on health, disease diagnosis, or exploratory conversations with the health professional. Its aim is different, and that is to relieve the health professional from answering the routine questions. Bots are best when utilised for social media outreach activities and fare extremely well in instances requiring a customer service orientation.

Yet, technology is advancing and these tools will as well. New processing power in chip technology, coupled with advances in algorithmic analysis will yield better tools for the healthcare and life sciences industries. AI, NLP, Bots, automation, machine learning and data mining are not new. They have been used in other industries such as manufacturing, customer service, and finance. Now progress in these other industries and the elimination of the requirement for extensive resources has caught on with the health sciences. As new adopters of these technological advances share their experiences, the entire health and life sciences industry will benefit, having access to a new set of tools to transform itself.


i. Accessed 29Oct2017

ii. Data Mining at FDA, Hesha J. Duggirala. 2016 Accessed 29Oct2017

iii. Oracle Empirica Manual 2015

iv. 30Oct2017