View source for The rise of big clinical databases

This is a review of a descriptive study by Cook and Collins (2015), “The rise of big clinical databases.” <ref name="Cook"> Cook JA & Collins GS (2015). British Journal of Surgery. The rise of big clinical databases. http://www.ncbi.nlm.nih.gov/pubmed/25627139. </ref>

==Introduction== 

Use of EHRs and big data are becoming more common, and global use promises to usher in a new era of analysis. This review addresses the properties of such data, benefits, use, pitfalls, and challenges for the future. The author provides a panoramic view of recent experience, and stresses improved and rigorous conclusions, but without compromising data quality in the process.    

==Purpose==

The paper discusses the concept of managing substantial amounts of data from multiple clinical databases, and how the data can best be employed in different research applications.

==Methods== 

This meta-analysis is divided into three sections: a) the types of data sources and benefits, b) usage, with respect to research study designs; and c) challenges relating to data quality, applications, and later interpretation of the data. 

==Results== 

The two chief sources of large databases are administrative and clinical. Main uses include i) describing population characteristics, such as diagnoses, the treatment process, and outcomes; (ii) identification of risk factors and their incorporation into predictive models; (iv) assessing outcomes of different interventions as determined by observational studies; (iv) comparing providers and health utilization, and (v) as a source of supplementary data for future study.  Advantages of big data are their obvious size, and potential for comparing providers.  Main challenges encountered are data accuracy and completeness, and the validity and reliability of results.

==Conclusion==

After discussing different databases and research applications, the author notes that despite the evolution of IT infrastructure to generate big data, this by itself does not translate to improved patient care. For example, at the EHR level, inaccurate, inconsistent, and absent data may impair quality of later analysis. While the potential to improve patient care is great, ensuring data quality is needed. Second, implying causality when it does not exist is another pitfall. Evidence-based data is the basis for generating valid [[Clinical Decision Support]] (CDS) policy, and the place to start is at the EHR.

==Comments==

This review was chosen because of its relevance, timeliness, and clarity. The use of large clinical databases in the healthcare industry worldwide has led to enormous amounts of data, which has potential to guide providers' in decision-making and problem-solving skills, thereby increasing quality of patient care. Big promises from big data come with big challenges such as, maintaining data quality and interpretation of the data. Therefore, one must be careful when translating results to health policies.

== Related articles ==

*[[Big Data to Knowledge (BD2K)]]
*[[With "big data" comes big responsibility: outreach to North Carolina Medicaid patients with 10 or more computed tomography scans in 12 months.]]

==References==
<references/>

[[Category: Reviews]]
[[Category: EHR]]
[[Category: Databases, Big Data, and Data Quality]]
[[Category:HI5313-2015-FALL]]