Clinical Study Teams Find Value in High-Fidelity Data Sources

Harnessing High-Fidelity Data in Clinical Trials

With the largest attendance in its 25-year history, the Society for Clinical Data Management (SCDM) Annual Conference was another sign of growing digital innovation in clinical trials. Called "The Festival of Clinical Data Science," the event filled the week of Oct. 8-11, 2023, in San Diego, CA, with fresh perspectives, deep insights, and good engagement among more than 1,000 professional colleagues.

AiCure was honored to serve on the panel for the topic, How to Overcome the Ever-Increasing Diversity of Data Sources in Support of Analytics, AI and Machine Learning. From where we sit, we see great opportunities for collecting and integrating high-fidelity data, delivering valuable insights in ways that have never before been possible.

In a clinical trial, high-fidelity data is information with a high capture rate or frequency that enables researchers to better understand complex patterns of behavior and treatment response. The video and audio captured by AiCure technology on patient smartphones and analyzed by our machine learning algorithms are examples of high-fidelity data. These novel data sources have high fidelity to the real-world patient experience and offer broader insights than what’s possible through traditional measures conducted in a clinic at single points in time.

High-fidelity data offer a wealth of learning opportunities to pharmaceutical research teams. This blog touches on the potential of novel high-fidelity visual and audio data captured via patients' cell phones, and addresses the work needed to ensure these data make a difference in clinical trials.1

Novel Data Collection: Challenges and Opportunities

Before the emergence of high-fidelity data collection methods, the basic process of data management in clinical research had not changed much in the last 20 years. Historically, research was conducted in a fit-for-purpose manner with isolated variables that were collected as specific timepoints. Clinical data managers would typically identify the format of the electronic data and try to map it onto existing processes. What’s happening today, and what resonated for me at SCDM 2023, is the growing challenge of integrating novel data sources into existing processes. There’s no quick and easy answer, but I’m seeing real progress across a wide variety of innovative companies, both small and large.

Today, visual and audio data collected via patients' cell phones or wearable sensors provide a precise and sensitive method of monitoring, offering critical understanding of the day-to-day patient experience. These insights can help determine when a patient may need more support to stay compliant with the protocol and their treatment, discover unreported adverse effects, and even provide predictive insights such as a patient's potential response to a therapy or likelihood of dropping out of the study.

Digital Biomarkers: Personalizing Care and Enhancing Conclusions

The development of digital biomarkers, which are therapeutically tailored endpoints derived from novel data sources, is another significant advancement in clinical trials. Digital biomarkers provide unique insights that can be used to personalize care and derive more comprehensive conclusions about a therapy's impact. By leveraging high-fidelity data, clinical study teams can better understand patient experiences, target high-impact interventions, and introduce more objective measures of treatment performance that ultimately improve the overall success of clinical trials.

Overcoming Challenges in High-Fidelity Data Integration

To fully harness the potential of high-fidelity data in clinical trials, several challenges must be addressed:

  • Data integration and standardization: Ensuring consistency and compatibility across datasets from various sources and formats2

  • Data diversity and heterogeneity: Managing the complexity of diverse data types, structures, and formats, which can complicate the data integration process2

  • Data privacy and security: Protecting sensitive information and complying with data protection regulations while integrating data from different sources3

  • Data cleaning and preprocessing: Addressing data quality issues, such as missing, inconsistent, or duplicate data, before using it in AI and machine learning models4

By overcoming these challenges, often through the use of innovative data integration and analytics tools that empower AI and machine learning, clinical data managers can effectively leverage high-fidelity data sources to derive insights that will improve clinical trials and advance healthcare.

A Call for Continued Collaboration

High-fidelity data sources, such as visual and audio data collected via patients' cell phones, offer a wealth of opportunities for pharmaceutical research organizations to gain new and unique insights into the patient experience. By harnessing the potential of novel data collection methods and addressing the challenges of data integration, clinical study teams can improve patient experiences, personalize care, and accelerate successful clinical research programs.

The collaboration among AiCure and my fellow panelists at SCDM 2023 promises to be a prelude to greater expansion of this conversation in the future. I look forward to connecting with others who share our enthusiasm for the opportunities that high-fidelity data are bringing to clinical research. Let’s keep the conversation going to optimize the value of novel data and improve patient experience and outcomes in clinical trials and beyond.

1Content powered by
2Data Integration Challenges for Machine Learning in Precision Medicine
3Guiding Principles for Sharing Clinical Trial Data
44 Common Data Integration Challenges - DATAVERSITY