Review
Application of Clinical Databases to Contemporary Cardiac Surgery Practice: Where are We now?

https://doi.org/10.1016/j.hlc.2015.01.006Get rights and content

Cardiac surgery has embraced and encouraged the use of large, multi-institutional datasets in clinical practice. From a research perspective, database studies have facilitated an increased understanding of cardiac surgery. Among other uses, they have allowed an investigation of disease incidence and mortality, high risk groups, disparities in health care delivery and the impact of new devices and techniques. Databases are also important tools for clinical governance and quality improvement. Despite their obvious utility, clinical databases have limitations; they are subject to treatment bias, contain missing data and cannot establish causality. Moreover, the ongoing maintenance of the database requires significant human and financial resources. In the future, inclusion of more detailed follow-up data and integration with other datasets will improve the utility of clinical databases.

Section snippets

Background

Whilst prospective and well-conducted randomised controlled trials (RCTs) are regarded as the highest level of research evidence in clinical medicine they often cannot be performed due to ethical, financial or practical reasons [1], [2]. This is particularly true in surgery where only 20-30% of primary treatment interventions are supported by randomised evidence [3]. For thoracic surgery only 14% of treatments are supported by randomised evidence [4]. Consequently, the effects of new devices

Advantages of Clinical Databases

The research utility of large databases is indisputable. One particular advantage is the ready availability of data on a large patient sample. This generally represents a larger proportion of the actual patient population and subsequently reduces sampling error and improves external validity [14]. Large databases also capture data on patients with rare diseases or those undergoing an infrequently performed procedure. Often, in clinical medicine, RCTs cannot be performed because low disease

Disadvantages for Clinical Databases

Despite their utility, databases have several disadvantages. Firstly, they are subject to treatment bias. The observed treatments were not randomly assigned, but rather the decisions were made based on patient, physician and institutional characteristics and preferences. Procedural selection is problematic as it confounds direct comparisons between groups. For research studies, several statistical techniques are employed to improve the ability to reduce this confounding and to help establish

Clinical Databases in Current Cardiothoracic Surgery Practice

Clinical databases capturing data on patients undergoing cardiothoracic surgery have been present for several decades. The largest, and arguably, most robust, cardiac surgery database is the Society of Thoracic Surgery (STS) Adult Cardiac Surgery Database (ACSD) from the United States which was established in 1989. The dataset has unique records on more than five million patients and represents 95% of programs in the United States [32]. The latest STS-ACSD edition collects information on >600

Future Endeavours

Cardiac surgery has increasingly used databases in clinical practice over the last 20 years. Moreover, clinical databases have been refined and improved through the process of continuous quality improvement. Nevertheless, there is considerable room for improvement. Many databases have not been linked to provide access to long-term survival and those that do often lack data on the cause of mortality. Moreover, angiographic follow-up is absent in all large clinical databases. The possible

Conclusions

Clinical databases are important for the purposes of clinical governance, quality control and research, among others. They have facilitated an increased understanding of the impact of cardiac surgery on patient outcomes. Important findings regarding the incidence or prevalence of cardiac risk factors in a population, disease and procedure-specific mortality, volume-outcome relationships, national trends, disparities in healthcare delivery, and other important insights will continue to be

References (40)

  • D.M. Shahian et al.

    Cardiac surgery risk models: a position article

    Ann Thorac Surg.

    (2004)
  • E.D. Peterson et al.

    The effects of New York's bypass surgery provider profiling on access to care and patient outcomes in the elderly

    J Am Coll Cardiol.

    (1998)
  • A. Saxena et al.

    Impact of smoking status on early and late outcomes after isolated coronary artery bypass graft surgery

    J Cardiol.

    (2013)
  • W.Y. Shi et al.

    Impact of socioeconomic status and rurality on early outcomes and mid-term survival after CABG: insights from a multicentre registry

    Heart Lung Circ.

    (2014)
  • W. Chan et al.

    Progress towards a National Cardiac Procedure Database–development of the Australasian Society of Cardiac and Thoracic Surgeons (ASCTS) and Melbourne Interventional Group (MIG) registries

    Heart Lung Circ

    (2011)
  • J.A. Cook

    The challenges faced in the design, conduct and analysis of surgical randomised controlled trials

    Trials

    (2009)
  • P. McCulloch et al.

    Randomised trials in surgery: problems and possible solutions

    BMJ.

    (2002)
  • N. Howes et al.

    Surgical practice is evidence based

    Br J Surg.

    (1997)
  • J.P. Vandenbroucke

    Observational research, randomised trials, and two views of medical science

    PLoS Med.

    (2008)
  • N.M. Thalji et al.

    Risk assessment methods for cardiac surgery and intervention

    Nat Rev Cardiol

    (2014)
  • View full text