Utilizing Tech Tools to Accelerate Drug Development

In today’s multi-provider clinical trial environment, sponsors need to rapidly and accurately assess vast volumes of operational, clinical and safety data. This is leading to a switch—from resource intensive manual methods to utilizing technological tools which provide near real-time data access and automated processes while still ensuring regulatory compliance.

However, this new era of technological advancement is not about replacing human expertise. On the contrary, it is about providing the right tools to deepen the understanding of all users and break down silos while keeping costs down and offering the customization needed to fully optimize processes on every clinical trial.

So why do we need to move away from traditional methods and what new opportunities can technology offer us?

Traditional approaches—time consuming and inefficient

Traditional onsite monitoring approaches which rely on manually verifying source data verification are labor intensive, inefficient, slow and static. Without real-time, or near-real-time, data, sponsors are missing out on the insights which enable them to make informed decisions, impacting clinical trial timelines and costs.

Conventional data processing takes place in silos with data from sources such as case report forms, labs, and real-world data managed individually and only merged when ready for study result reporting. At the same time, a reliance on paper-based systems results in significant data handling challenges—including an increased risk of errors and difficulties ensuring data integrity and consistency.

These problems have only been exacerbated as clinical trials have become more complex.

In contrast, by utilizing technology-driven tools more effectively, we can drive efficiencies, improve communication, and provide better insights. For example, a portal to intake and integrate previously siloed data can enable near real-time cross-sector data review and comparisons, leading to earlier identification of risk.

Customizable tools can enable users without statistics or programming backgrounds to analyze data themselves and get the insights in real time, reducing the waiting time for analysis. These tools can also help with budget control, particularly in small biotech organizations which do not have the budget or demand to hire a programmer or data manager.

Combined, these tools lead to confident, informed decision making, and ultimately, create savings.

Breaking down silos and unlocking new insights

One technological advancement which is helping to transform clinical trials is the use of application programming Interface (API) integrations to connect different, previously siloed systems. This means platforms can become tool agnostic, taking data from various sources and structuring that data in a way which allows us to deepen our understanding.

This agnostic approach allows the integration not just of different data sources from across a single study, but from different sources across multiple studies. Standardization allows us to ingest the data, prepare visualizations across different styles, and carry out analysis between different studies. This breaks down silos, creates efficiencies, and can unlock new insights for the client.

This process can be further enhanced with the use of artificial intelligence (AI)-enhanced analytics. While different naming conventions and structures, combined with the volume of data, can make it challenging to spot similarities across different studies manually, a large language model or machine learning can analyze data and spot the similarities between differently named data across multiple studies.

Customization to drive efficiency

To be truly effective, technologies need to be customizable to the specific needs of each clinical trial or client. Template approaches which do not reflect what sponsors want, or need are not fit for purpose in a modern clinical trial environment. Instead, we need to be able to listen to user needs and respond to the specific requirements of the protocol or business use, allowing sponsors to utilize tools in a way which works for them.

By creating a custom-fit model for a specific study we can drive the best results and improve quality of medical review. Once again, it comes down to getting faster results which are more on point.

We also need to understand key industry challenges and create solutions which respond to them. For example, from a case study, clinical trials run by a sponsor organization frequently failed to meet the target of administering the first treatment to the last participant within the planned time frame. The problem stemmed primarily from an inability to swiftly identify bottlenecks within studies and formulate effective solutions.

A recruitment forecasting system was developed to provide valuable insights into recruitment efforts, predict when the last participant would receive treatment, and propose strategic reallocation scenarios at a country level. It factored in non-started sites, improving forecasts for situations where sites opened later in the recruitment phase, and integrated a novel Kaplan-Meier model to predict participant progression based on their screening duration.

The system featured real-time data updates, accessible to all sponsor personnel. This facilitated swift access and customized filtering of forecasts across various geographical levels. The integration of a write-back solution empowered authorized users to view forecasts and actively reallocate participants between different countries based on real-time data and predictions.

With the ability to accurately reallocate patients across different countries, the company witnessed a significant acceleration in trial completion, ensuring that trials consistently adhered to planned timelines.

Turning complexity into clarity

By putting effective technology-powered tools into the hands of users we can make insights available to all stakeholders, not just data managers. Data visualizations allow a whole range of users to explore the data themselves at a level that works for them—whether that is a top-line overview of the whole clinical trial or digging into granular data at a very detailed patient level.

Adaptive data visualizations enable quick and easy data comprehension and allow users to seamlessly draw connections between clinical events of interest and patient records. They also save resources as insights can be continuously accessed without having to ask colleagues to extract or analyze data.

For example, a new application was developed to provide a user-friendly interface for generating summary tables and graphics from study data tabulation model (SDTM) data. The aim was to help clinicians explore and understand investigational drug performance and safety profiles, enabling faster decision making and exploration of trial data. The data processing workflow included two key steps. The first was the creation of a derived subject-level (DSL) dataset by processing SDTM data. The second was to combine the DSL dataset with additional SDTM domains to generate tailored analysis datasets for specific outputs. Once outputs were created, users could further refine their analysis to generate precise insights, interactive visualizations and analysis outputs. The final outputs could be customized and exported in various formats for reporting, regulatory submission, or presentations.

A key challenge when developing the app was generating rich text format (RTF) outputs for tables and listings. Clinical reporting requires detailed formatting which adheres to regulatory standards. While the existing package offered a starting point, it lacked the flexibility to meet all formatting requirements. To address this, a custom RTF generation tool was developed. This allowed greater control over table structures and formatting.

Another challenge was creating customizable, interactive plots which met the detailed requirements of clinical standards. An additional package was adopted which extended the functionality to include greater interactivity. While this approach enhanced user engagement, it posed new challenges in balancing interactivity with strict formatting and consistency requirements. Extensive customization was required to ensure compliance and maintain user experience.

These challenges highlight the need to balance competing requirements when utilizing advanced tech tools in clinical trials. As user expectations increase, care must be taken to align user experience and usability with regulatory requirements.

Honing the power of AI

In the past, many technological advancements came too late or were too expensive. Now, however, they are feasible, and we are finding new ways to utilize them to, for example, automate data processing and search public data via specifically-trained models which read-in data through API-like connections.

As we move into a new era of digitally-enabled clinical trials, we will need to continue to hone our use of AI. A new capability being explored is the use of AI to understand the context of the data we are working with in medical monitoring.

A medical monitoring analytics app can already be used to enhance the medical review process. By using real-time data to achieve timely detection of events and signals of special interest, an app can provide a holistic view of each patient through consolidated patient profiles, backed by visualizations. This empowers medical monitors to make data-driven decisions, improving the likelihood of successful trial outcomes.

However, hypothetically, if AI has full understanding of the data behind a medical monitoring product, you would be able to ask it, for example: “How many AEs does this have patient have?” and it would be able to give you an answer within the tool. Such an application would have compliance implications which would need to be explored. But it demonstrates how backend data processing and intaking of public domain data can use AI to reduce human effort, shorten data processing time, and ultimately expedite clinical trials.

Using technology to turn data into intelligence

Behind every data point is a patient waiting for treatment. By leveraging technological tools and combining them with deep therapeutic expertise, we can transform complex data into clear, actionable intelligence.

To achieve this, new tools must be usable, customizable, promote a deeper understanding of the data, and empower rapid anomaly resolution. We also need to remember that technology is only as good as the people using it. Organizations need the right people interpreting data and implementing technology to truly drive efficiency.

If we can achieve all the above, and balance user expectations with regulatory requirements, we can then accelerate drug development and bring life-changing therapies to patients faster in a new era of digital solutions.

Christian Schmidt, an Associate Director at Phastar