This is Part #2 of a blog series describing how disruptive innovation technologies such as Saama’s Life Science Analytics Cloud is enabling the Pharma industry to bend its clinical development cost curve sharply downwards while bringing cures to patients sooner. This blog, co-authored by Nekzad Shroff and Amit Gulwadi , explains the ROI benefits of improved trial planning decisions through the smart use of data driven insights.
Clinical trial planning is an iterative and strategic undertaking which can have an outsized influence on the overall outcome of a clinical trial and the eventual fate of the candidate therapy. There are several decisions which are made in the planning phase which have traditionally required painstakingly manual data collection, research and analysis in order to give decision makers insights to enable the trial the best chance of success. And, in spite of best efforts, trial planning decisions still leave a lot of room for improvement, given some of the unflattering statistics on trial performance available in the industry. For example:
- Sponsors only manage to identify and recruit about 5% of potential patients who could benefit from a clinical trial
- 25% of sites contracted for trials fail to recruit a single patient
- 80% of trials are delayed due to inability to recruit the right patients
There are several processes during the planning phase of a trial which stand to benefit from disruptions related to improved analytics capabilities available today. Through our extensive understanding of the clinical development continuum, we have identified the following potential opportunities which are addressed as use cases in the Study Planning module of Saama’s Life Science Analytics Cloud (LSAC) solution. We also modeled the likely benefit for a typical drug manufacturer in terms of quicker time to approval and reduced development costs in bringing a drug to market. We conservatively estimate that a sponsor implementing these capabilities can shorten the planning phase of a typical trial by 8 weeks, along with saving about $285,000 in planning costs of a single trial, which translates to a program level savings of $13 million in getting a single drug approved!
- Trial Plan Creation
Today, building a trial plan is an exercise in spreadsheets and frustration. This strategic activity requires a small but very senior team of trial manager, trial physician and therapeutic area leadership to make critical decisions on the key trial parameters which will give the trial the best chance for successful outcome. This is a time sensitive, gating activity which requires optimization across a diverse set of input data such as protocol design, primary and secondary endpoints, country and site capabilities, historical site and investigator performance benchmarks, current competing trials, incidence prevalence and overall feasibility. Much of this data is spread across a variety of systems and spreadsheets and requires feasibility experts to spend time very inefficiently in data wrangling and organizing tasks. Basic feasibility decisions require days of effort and “what-if” scenario analyses with these data inputs which have to be manually built out as spreadsheet models from scratch every time. The whole process can easily take this team working together about 3-4 months
Saama’s LSAC for Study Planning directly addresses these inefficiencies by bringing together a variety of critical data points into an analytics interface optimized for making trial planning decisions. For example, building custom cohorts of patients to analyze protocol feasibility, or visualizing patient and PI clustering to model rate of recruitment. The trial planners can quickly dive into insights, predictions and guided analytics specifically designed for creating the Trial Plan, including scenario planning and tradeoffs of timelines, protocol, Rate of Recruitment (RoR), screen failure rate (SFR), dropouts, data quality and financials. We conservatively estimate that this can reduce the planning phase timeline by at least 25% and the manual data collection and analysis effort by 50%!
- Country Selection and Approvals
The Country Selection process now involves a much larger team, distributed across geographies. Each country in consideration generally would have a country manager, medical director and clinical research associate (CRA) all providing inputs on country specific feasibility parameters such as commercial market size, IP protections, incidence prevalence, commercial regulatory requirements, ethical and regulatory requirements, regulatory timelines, import feasibility, standard of care, reimbursement and country historical performance. The country team would have to put together all this data from their own research and potentially working with sites. Traditionally, it is left to the trial manager or feasibility expert in the home office to manually collect and collate all these questionnaires from the countries into a cohesive (spreadsheet) model for country feasibility assessment.
Saama’s LSAC acts as an intelligent store of country and site level data whose capabilities grow exponentially over time as it learns from more and more trials being conducted. Contrary to the manual approach, the system does not forget historical inputs. Powerful machine learning algorithms embedded into the workflow are able to process these multi factorial inputs to arrive upon predefined and configurable recommendations of country feasibility which can be applied to various scenarios for decision making. Country selection decisions can be more accurate and data driven and the timeline for making these decisions can be (conservatively) reduced by at least 25%!
- Site Selection
Site selection can be one of the most time consuming processes, stretching for months or quarters with a large global team collaborating on making site and investigator selection decisions. When the trial is outsourced to a CRO, there is often extensive reliance on CROs to provide historical site performance inputs based on the trial protocol and feasibility inputs from the sponsor. The list of sites with historical studies provided by the CRO then have to be analyzed to shortlist the sites to be selected based on the CRO provided metrics. These sites are then contacted with questionnaires for site inputs which are faxed or emailed back and forth and have to be manually analyzed and synthesized in order to make the overall site selection decision.
This is an area which is ripe for innovation. Saama’s LSAC site selection capability has powerful machine learning models which are able to make recommendations on the best sites and investigators for the trial based on optimizing across a variety of predictive inputs such as Infrastructure (IRB, contracting, facilities, personnel), Patients (protocol criteria, access to patients), Investigators (expertise, capability, KOL strategic value), Feasibility questionnaire responses and Operational performance (speed, productivity, quality). Site selection decisions are much more accurate and can leverage hidden patterns which manual human analysis would not be able to process. We modeled that it is typically possible to reduce the painstaking site selection process by at least a full month. However, the true benefit of better site selection would show up on the study conduct numbers, with dramatically improved rate of recruitment, higher performing sites, fewer costly rescue sites and correspondingly shorter timeline for study conduct.
Innovative pharma and biotech sponsors have just started to understand the huge potential of disruptive intelligence in the trial planning process. With healthcare costs being high on the radar and advanced data capabilities becoming mainstream, the time for transformation is not on the horizon. It is now!
 Compel Data to Predict High-Performance Clinical Trial Sites, Amit Gulwadi
 Baseline trial considered for estimation – Oncology, Phase 2 or 3, 10 countries, 100 sites, 2 year execution.