The Internet has caused us all to be our own researchers. It seems like picking a restaurant for dinner involves more research than we put into our college thesis papers. When making almost any decision, we actively gather real world data and evidence on the Internet. In the case of a restaurant, this often comes from diner reviews, social media mentions, and forum postings. Now, imagine if we could gather the same level of real-time data in regards to health across the population and apply the information to developments in treatments, medications, and predicted outcomes.
The 21st Century Cures Act emphasizes a renewed focus on using real world data to power the decisions behind clinical trials. The FDA and Congress define real world data as, “data regarding the usage, or the potential benefits or risks, of a drug derived from sources other than traditional clinical trials.” This idea was once unfathomable due to excessive data blocking. The 21st Century Cures Act seeks to remedy the practice of information blocking, which can be simply defined as healthcare technology companies restricting the ability to share information outside of their platforms. While progress remains to be made, the widespread adoption of FHIR as a new standard for information exchange has pushed vendors towards greater interoperability.
Lack of Diversity in Real-World Data
Let’s go back to the restaurant example. Imagine that you only saw reviews from one type of diner – you may read insights from individuals that have nothing in common with you. Researchers must be cautious that their “real-world data” is not only applicable to a specific age group, gender, or race. In a research piece by senior author Dr. Daniel Rubin of Stanford University and lead author Dr. Eli Cahan of Stanford and New York University, researchers found that the majority of data shared today is based on European ancestral genotypes. The authors suggested that greater transparency be presented in regards to the populations that data was collected from.
This isn’t the first time that researchers have documented a lack of diversity in studies.
In the article, Social Factors Matter in Cancer and Survivorship, researchers launch into an in-depth study of why the mortality rate for breast cancer remains so much higher for black women than white. Their conclusion is that genetic variants and social factors must play a much greater role in the research data used by scientists. Social factors can include data regarding income, education, social status, ease of access to good healthcare, and life events, like childhood trauma. These can all have a significant outcome on the efficacy of a treatment on a certain population. The authors suggest that “When possible, findings should stratify these data by race, SEP, country of origin, sexual orientation, health insurance, or other social factors that theory or previous research suggests are important to the exposures or outcomes being studied.”
The popular asthma inhaler, albuterol, is one of the most cited examples of how a lack of diverse data can negatively impact populations. For years, physicians had noticed that albuterol was much more effective in children of European or Mexican descent rather than African American or Puerto Rican descent. Researchers at USCF found that 95 percent of previous studies had focused on people of European descent. A new study with wider genetic variation was able to pinpoint the genetic markers that are more prevalent in people of African ancestry that caused less effective treatment. Today, physicians can use this information to prescribe the right inhaler type.
A Deep-dive Into Specialty Pharma
A specialty drug is a class of prescription medications used to treat complex, chronic or rare medical conditions. Although this classification was originally intended to define the treatment of rare, also termed “orphan” diseases, affecting fewer than 200,000 people in the US, more recently, specialty drugs have emerged as the cornerstone of treatment for chronic and complex diseases such as cancer, autoimmune conditions, diabetes, hepatitis C, and HIV/AIDS.
This data can be collected from a wide variety of sources ranging from electronic health records, billing data, surveys, and even mobile applications. Previously, the ability to collect and organize huge amounts of data was impossible, but today, machine learning and artificial intelligence tools make it possible.
Using existing and diverse data can save clinical trials the costs, time, and negative patient outcomes associated with traditional randomized trials.
Lack of Data Types
Let’s turn to that restaurant example one last time. Pictures are a huge part of reviews today. Could you imagine not even looking at one image? Yet, imaging is often the forgotten piece of the puzzle in establishing a holistic patient health record. It is not possible to establish strong research programs without imaging, much of which may already be available in the real world due to screening programs. However, the organization and anonymization of imaging often act as a pitfall. Traditionally, medical imaging is still often exchanged on CDs leading to time wasted mailing CDs, uploading and burning, and manually entering patient information. The process is dangerously error-prone and often delays progress in trials due to the slow input of data. Variations in imaging viewing systems and radiology reporting can also cause inconsistent data.
Rapidly indexing through billions of images and data requires complex IT infrastructure. When looking for solutions, a facility must consider tooling that provides solutions around searching for all relevant data and then appropriately anonymizing and de-identifying data, including the meta-data at the pixel level in the case of medical images. Researchers must determine if tooling is able to properly de-identify data while allowing them to maintain a record of information they’ve gathered based on certain conditions. For example, researchers may want to index MRI’s of a specific tumor type based on gender or age. At a leading academic research center, researchers sought to incorporate increased data types but were frustrated by CD related delays. A partnership with a cloud vendor allowed the facility to securely share imaging over the web and remove patient health information from the DICOM tags client-side before the study leaves the sending facility, eliminating the risk of accidentally leaving PHI (personal health information) tags in place.
Working together, vendors and facilities can establish strategies for gathering real-world evidence, including imaging and diverse data, successfully. The data can then be securely anonymized, shared, and even run through additional processing for the acceleration of clinical trials. The future of trials is bright, and we look forward to the patient care improvements that will be made along the way.
Photo: metamorworks, Getty Images
