From observation to innovation: Natural history studies in the era of AI

Natural history studies play a pivotal role in advancing our knowledge of diseases. These observational studies, which track the progression of diseases in their natural state without medical intervention, are particularly crucial in the realm of rare disease research. They serve as fundamental benchmarks against which the effectiveness of new treatments can be measured. In conducting these studies, extensive data is collected over an individual's lifetime, including initial diagnosis, clinical observations, and patient insights on quality of life. This rich tapestry of information provides a deep and nuanced understanding of how a disease progresses and its impact on those affected.

The long-term nature of these studies results in an abundance of data, which requires thorough analysis. Here, artificial intelligence (AI) emerges as a powerful ally. AI's capacity to process and interpret vast datasets efficiently can significantly aid researchers in conducting natural history studies. In this blog, we’ll further define natural history studies and explore the various ways AI can bolster them, mapping a promising intersection of technology and healthcare.

What are natural history studies?

The primary goal of natural history studies is to chart the unaltered trajectory of a disease, capturing its progression from onset to resolution, or until the end of a participant’s life. Ideally, to understand a disease's “natural history,” the study would observe a population not influenced by medical interventions. However, from an ethical standpoint, withholding treatment can have detrimental effects on participants' wellbeing and it is therefore rare for a study to observe a completely untreated population. In practice, natural history studies  tend to involve individuals who have exhausted available treatments. 

Natural history studies aim to capture various factors—demographic, genetic, environmental—that correlate with disease development and they are vital for a few key reasons: 

Understanding disease progression: Rare diseases in particular often have limited existing research and data. Natural history studies provide crucial insights into how these diseases progress over time, which is essential for developing effective treatments and management strategies.

Identifying variability and subtypes: Many diseases exhibit a wide range of symptoms and progression rates among different individuals. Through natural history studies, researchers can identify different subtypes and variability within a disease, leading to more personalised and effective treatment approaches.

Designing clinical trials: In rare diseases, the small number of patients makes traditional clinical trial designs challenging. Natural history studies can inform the design of these trials, helping to establish appropriate outcome measures, patient selection criteria, and other key design elements.

Improving patient care: Understanding the progression and impact of a disease through natural history studies can lead to better patient care. It helps healthcare providers anticipate and manage symptoms, improving the quality of life for patients.

Regulatory approval and drug development: Data from natural history studies can be crucial in the drug development process, especially for rare diseases. This data can help in demonstrating the drug's effectiveness and safety, which is often a requirement for regulatory approval. For example, natural history studies have been instrumental in researching rare diseases like Duchenne Muscular Dystrophy (DMD), Spinal Muscular Atrophy, and Huntington's Disease. They have provided critical insights into these diseases' origins and mechanisms, aiding treatment development. A notable example is the Cooperative International Neuromuscular Research Group's Duchenne Natural History Study, a comprehensive, decade-long study involving 440 patients from 20 centres across 9 countries.

Regulatory authorities are increasingly recognizing the value of real-world evidence, with natural history studies gaining prominence in drug development. In fact, the U.S. Food and Drug Administration (FDA) has issued guidelines for incorporating natural history study data into drug development, emphasising four key areas:

• Patient population: Understanding the variability in diseases related to genotypes and/or phenotypes.
• Clinical outcome assessments: Using natural history data to validate new or existing assessments in detecting disease progression, including clinic-reported, observer-reported, patient-reported, and performance outcomes.
• Development of biomarkers: Leveraging studies for biomarker strategies that can diagnose diseases, predict disease courses, and guide treatment responses.
• Use of study data: Encouraging the creation of external control arms from patient registries or natural history studies.

How AI is supporting data collection and analysis in natural history studies

AI is revolutionising the way natural history studies are conducted, particularly in terms of data collection and analysis. The role of AI in these studies can be broadly categorised into several key areas:

Enhanced data collection: AI technologies, particularly in the form of wearable devices and mobile apps, enable continuous and real-time collection of a wide range of health metrics. These include physiological data such as heart rate, blood pressure, and glucose levels, as well as behavioural data like physical activity, sleep patterns, and medication adherence. This rich, continuous stream of data provides a more accurate and comprehensive picture of the patient’s condition, as opposed to traditional periodic assessments.

Sophisticated data analysis: AI excels in analysing large datasets, identifying patterns and correlations that might be missed by human researchers. Machine learning algorithms can sift through complex, multidimensional data to uncover insights about disease progression, potential biomarkers, and patient response to treatment. This helps in forming a more nuanced understanding of the natural course of diseases.

Predictive modelling: AI can be used to create predictive models that forecast disease progression, potential complications, or treatment outcomes. These models can help in identifying high-risk patients, tailoring interventions, and even guiding clinical trial design by predicting which patients are most likely to benefit from a particular treatment.

Enhancing patient recruitment: AI algorithms can scan through vast electronic health records (EHRs) to identify potential study participants who meet specific inclusion criteria. This not only speeds up the recruitment process but also ensures a more accurate and diverse participant pool, which is crucial for the validity of natural history studies.

Improving patient engagement and compliance: AI-driven tools can be used to engage patients more effectively in a study. For instance, AI-powered chatbots and apps can provide personalised reminders for medication intake, appointments, and data recording. They can also offer support and information, enhancing patient adherence to study protocols.

Real-time monitoring and intervention: With AI's capability for real-time data processing, any significant changes in a patient’s condition can be flagged immediately. This allows for timely interventions, which is particularly key in studies involving progressive or rapidly changing conditions.

Data standardisation and integration: AI can assist in standardising data from diverse sources, ensuring that it is consistent and compatible for analysis. This is especially important in multi-centre studies, where data is collected from various locations and potentially in different formats.

In essence, AI is a game-changer in natural history studies, significantly enhancing every aspect from data collection to analysis, thereby paving the way for more accurate, efficient, and impactful medical research.

Conclusion

Natural history studies, essential in understanding disease progression, are being transformed by the integration of AI. This combination is streamlining research processes, from data collection to analysis, making studies more efficient and insights more accurate. As AI continues to evolve, its role in these studies promises to further enhance our understanding of diseases and improve outcomes in drug development and patient care. This marks a significant step forward in medical research, where the blend of traditional study methods and modern technology leads to more informed and effective healthcare solutions.

To learn more about how Sano is leveraging AI in its work to drive precision medicine forward, get in touch below!