No abstract

PURPOSE Medical records contain a wealth of useful, informative data points valuable for clinical research. Most data points are stored in semistructured or unstructured legacy documents and require manual data abstraction into a structured format to render the information more readily accessible, searchable, and generally analysis ready. The substantial labor needed for this can be cost prohibitive, particularly when dealing with large patient cohorts. METHODS To establish a high-throughput approach to data abstraction, we developed a novel framework using natural language processing (NLP) and a decision-rules algorithm to extract, transform, and load (ETL) melanoma primary pathology features from pathology reports in an institutional legacy electronic medical record system into a structured database. We compared a subset of these data with a manually curated data set comprising the same patients and developed a novel scoring system to assess confidence in records generated by the algorithm, thus obviating manual review of high-confidence records while flagging specific, low-confidence records for review. RESULTS The algorithm generated 368,624 individual melanoma data points comprising 16 primary tumor prognostic factors and metadata from 23,039 patients. From these data points, a subset of 147,872 was compared with an existing, manually abstracted data set, demonstrating an exact or synonymous match between 90.4% of all data points. Additionally, the confidence-scoring algorithm demonstrated an error rate of only 3.7%. CONCLUSION Our NLP platform can identify and abstract melanoma primary prognostic factors with accuracy comparable to that of manual abstraction (< 5% error rate), with vastly greater efficiency. Principles used in the development of this algorithm could be expanded to include other melanoma-specific data points as well as disease-agnostic fields and further enhance capture of essential elements from nonstructured data.

The concept of Big Data is changing the way that clinical research can be performed. Cardiothoracic surgeons need to understand the dynamic digital transformation taking place in the healthcare industry. In the last decade, technological advances and Big Data analytics have become powerful tools for businesses. In healthcare, rapid expansion of Big Data infrastructure has occurred in parallel with attempts to reduce cost and improve outcomes. Many hospitals around the country are augmenting traditional relational databases with Big Data infrastructure. Advanced data capture and categorization tools such as natural language processing and optical character recognition are being developed for clinical and research use, while Internet of Things in the form of wearable technology serves as an additional source of data usable for research. As cardiothoracic surgeons seek ways to innovate, novel approaches to data acquisition and analysis enable a more rigorous level of investigatory efforts.

In the first part of this series, we introduced the tools of Big Data, including Not Only Standard Query Language data warehouse, natural language processing (NLP), optical character recognition (OCR), and Internet of Things (IoT). There are nuances to the utilization of these analytics tools, which must be well understood by clinicians seeking to take advantage of these innovative research strategies. One must recognize technical challenges to NLP, such as unintended search outcomes and variability in the expression of human written texts. Other caveats include dealing written texts in image formats, which may ultimately be handled with transformation to text format by OCR, though this technology is still under development. IoT is beginning to be used in cardiac monitoring, medication adherence alerts, lifestyle monitoring, and saving traditional labs from equipment failure catastrophes. These technologies will become more prevalent in the future research landscape, and cardiothoracic surgeons should understand the advantages of these technologies to propel our research to the next level. Experience and understanding of technology are needed in building a robust NLP search result, and effective communication with the data management team is a crucial step in successful utilization of these technologies. In this second installment of the series, we provide examples of published investigations utilizing the advanced analytic tools introduced in Part I. We will explain our processes in developing the research question, barriers to achieving the research goals using traditional research methods, tools used to overcome the barriers, and the research findings.

No abstract