When it comes to healthcare, clinical research is vitally important. Through research, we can find the answer to tough questions, and refine the way healthcare professionals work. We can discover new ways to treat patients, and establish which of these treatments are most effective.<\/p>\n\n\n\n
But in what is an increasingly digital world, managing the plethora of available health data should be a key part of any research effort.<\/p>\n\n\n\n
Health data management involves how to best house and use digital health data, ranging from electronic medical records to hand-written medical notes. But what constitutes good health data management?<\/p>\n\n\n\n
To answer this question, we\u2019ll follow the journey of Dr Jade, a clinical data science researcher who\u2019s facing challenges with data needed to develop a localised readmission risk model \u2013 and show how a modern data platform can address these challenges.<\/p>\n\n\n\n
Applying predictive modelling to identify patients with high risk of hospital readmission, and applying effective interventions to mitigate that risk requires well-managed health data.<\/p>\n\n\n\n
In creating a localised readmission risk model for this, Dr Jade\u2019s primary responsibilities include enabling data-driven clinical research approaches, protecting the privacy of the data, and validating the reliability and credibility of research results.<\/p>\n\n\n\n
Her research to create this model commences with conversations with local clinicians to understand the factors that contribute to hospital readmissions, and conducting a detailed literature review on causes of readmissions, globally.<\/p>\n\n\n\n
The next step is to analyse clinical data collected from a range of sources, including (but not limited to) Electronic Health Records (EHR), health registries, clinical trials, genetic information, wearable data, and care management databases.<\/p>\n\n\n\n
While this may sound straight-forward, Dr Jade\u2019s challenge is a big one: she needs to get access to relevant data; collect, structure and process it; and when that\u2019s completed, she has to know what to look for in the data itself.<\/p>\n\n\n\n
Having proper access to clinical data is important for research. While the level of health data being generated has skyrocketed in recent years, much of this is held in \u2018siloed\u2019 systems that aren\u2019t interoperable (able to exchange and use health information from one another to benefit research).<\/p>\n\n\n\n
In the United States alone for example, approximately 1.2 billion clinical documents are produced each year \u2013 but there is no single repository of information to rely on. For researchers like Dr Jade, access to data beyond that contained in their own health system is required to achieve the best research results.<\/p>\n\n\n\n
Currently, Dr Jade will have to make multiple queries to pull data from various sources. Assembling this unstructured data into structured cohorts, forming a standardised set of data from data types such as medications, lab tests and procedures can be particularly time consuming.<\/p>\n\n\n\n
Dr Jade will also need to consider different types of data. Social determinants of health\u2014which are non-medical factors that impact a person\u2019s health\u2014play a significant role in an individual’s health status.<\/p>\n\n\n\n
Dr Jade will also need to request information from another health system on current readmission risk rates for certain conditions. In some cases, these health systems won\u2019t be sufficiently resourced to facilitate this.<\/p>\n\n\n\n
Patient-generated data (such as data generated by medical devices), and genomic data are also important types of data Dr Jade may wish to analyse, but currently there\u2019s limited flexibility for ad hoc data sets and new data types such as those outlined above to be added to her research programme.<\/p>\n\n\n\n
In short, the sheer volume and scope of data Dr Jade needs to assess is overwhelming. Without the right healthcare technology and processes in place, it\u2019s difficult to validate models and scale-up adoption of these models into regular use.<\/p>\n\n\n\n
Given the significant roadblocks Dr Jade faces, it\u2019s clear that a solution is needed to remove current barriers and allow data to be accessed and used in research. Modern data platforms are one such solution.<\/p>\n\n\n\n
Modern data platforms possess the capability to address fragmented health data; aggregate and curate disparate sources of health data; and add new data types as they\u2019re created. They provide a single point of access for research, such as that which Dr Jade is undertaking.<\/p>\n\n\n\n
Research shows that effective readmission interventions that are poorly targeted can reduce hospital readmission rates by over 28%. Imagine the potential of using a modern data platform with machine learning capability to improve targeting.<\/p>\n\n\n\n
Health systems can reasonably expect an even greater reduction in the rate of hospital readmissions. This will save researchers like Dr. Jade time, effort, and cost and ensure that the right data is in the right hands at the right time.<\/p>\n\n\n\n