Innovation in behavioral health has long been limited by a lack of available data to support evidence-based decision-making. This lack of data can be attributed to many factors. Some of these factors are impossible to change, such as the heterogeneity of how the same diagnosis presents differently from patient to patient. Others could be changed but not without cooperation across behavioral health stakeholders; for example, a lack of consensus regarding how to define patient improvement, in addition to which tools to use to measure improvement, has resulted in a lack of standardization unseen in other areas of health.
Despite these challenges, it is not entirely true that behavioral health lacks data; instead, improvement has been hindered because only a very small percentage of the data is accessible and usable. At Holmusk, we are working to overcome this challenge, making more data available and fit-for-purpose through collation, harmonization, and enrichment. The result: Our NeuroBlu Database, the richest, densest source of real-world data for behavioral health.
But what does it mean to have rich, dense data? In this blog, we’ll tackle what data density means, and the initiatives that we at Holmusk are undertaking to achieve it.
For behavioral health real-world data to be considered dense, it must contain data points specific to each and every clinical visit for a patient. Ideally, this information will be more detailed than simply recording that a visit occurred—it will be accompanied by structured measures, when taken, or will be associated with clinical notes giving further information about the appointment and how the patient was doing.
Dense data means we have more information per patient across the entirety of their diagnosis and time in treatment, enabling researchers to understand disease progression, treatment effectiveness, and more. Once we have this dense data for enough patients (thousands), we are able to power analyses at scale. This is our primary aim in building the NeuroBlu Database.
With over 37 million visits from more than 1.5 million patients currently represented in the NeuroBlu Database, we are well on our way to having multiple visits for most of our patients, bringing us closer to the goal of data density. But we also understand there are other initiatives we can undertake to extract even more value from our available data, to support future research, and to bring behavioral health closer to standardization.
As mentioned earlier in this post, behavioral health faces a lack of standardization. One challenge commonly seen in care delivery is that the implementation of measurement tools varies widely across different states and health systems. When providers collect measurements at inconsistent time points, or even use completely different scales to measure the same conditions, it becomes difficult to make comparisons. Our team at Holmusk is analyzing different scales to determine if they can be reliably converted. This conversion would reduce the challenges of comparing apples to oranges, facilitate seamless interactivity among the different scales collected, and enable more effective research.
Another way that Holmusk introduces standardization and makes research accessible is through the development of natural language processing (NLP) models. These models extract structured variables from unstructured data, which was previously inaccessible and impossible to analyze without time-consuming manual processes. This scalable solution reveals more information about patients and their symptoms than ever before, and this information is then made available in NeuroBlu, Holmusk’s powerful data analytics software.
By using NeuroBlu, researchers will gain access to the world’s densest, richest data for behavioral health, and will be able to quickly and easily draw insights from analysis of this valuable information. In the face of the current realities surrounding behavioral health, enriched, dense data is likely the most viable solution until stakeholders can come together to move the needle toward consensus and standardization.