Background

In 1987, the American College of Cardiology (ACC) committed to developing a standardised database to characterise the clinical profiles and outcomes of patients undergoing cardiac catheterisation and percutaneous coronary intervention (PCI) in the USA. This initiative began with the formation of a database committee that explored various strategies for the collection and evaluation of cardiovascular data and a data standards workgroup to develop a standardised cardiovascular vocabulary. In 1990 after this process matured, the ACC contracted with Summit Medical Systems, Inc. to develop and distribute software for the collection of 411 core data elements with annual data submissions to a central repository.

In planning the original dataset, the database committee specified that periodic revisions would be required to reflect changes in clinical practice and evolving science. In 1994, the database committee initiated the first update of data elements. This update also initiated the process of identifying which new clinical variables to properly characterise patients, developing data standards that were necessary to collect and analyse clinical data consistently.

In 1997, the leadership of the ACC approved Cath Lab Module V.1.1 with data collection beginning in November 1998. Simultaneously, by increasing the support services provided to hospitals collecting and submitting data, the ACC entered a new phase of data collection and clinical information storage. This was accomplished with the following actions:

1. Establishing the National Cardiovascular Data Registry registered trademark (NCDR) with management and support for the NCDR located at ACC Headquarters;

2. Allowing commercial vendors to develop data collection software that would be subject to approval and certification by the ACC;

3. Hiring clinically trained personnel to support and recruit NCDR participants;

4. Using the NCDR as a data processing and quality feedback tool;

5. Increasing the frequency of data submissions from annually to quarterly.

In November 1998, the original Data Registry participants were asked to re-enrol with the NCDR CathPCI Registry. A second revision of the CathPCI Registry data collection tool (DCT) (v2.0b) was developed in March 2000. The DCT is now on its fourth version, each with refined variables collected in alignment with the strategic goals of the registry (see online supplementary appendix A). The Society for Cardiovascular Angiography and Interventions (SCAI) joined the ACC as a partner in the CathPCI Registry, which has now grown to include 1577 US hospitals, several international sites and more than 12 million patient records (figure 1). With this growth and increasing use of these data for research and quality purposes, the burden and complexity of data analysis has grown. The ACC contracts with select academic research organisations to manage data and analyses for scientific publications. The primary research analytic centre for the CathPCI Registry is the Duke Cardiovascular Research Institute. A full description of the historical development of the NCDR is presented elsewhere.1 ,2

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Numbers of patient records and participating sites in the National Cardiovascular Data Registry CathPCI Registry.

Organisation and operations of the NCDR and the CathPCI REGISTRY

The mission of the NCDR is to improve the quality of cardiovascular care by providing information, knowledge and tools; implementing quality initiatives; and supporting research that improves patient care and outcomes. The NCDR consists of seven clinical registry programmes, the oldest being the CathPCI Registry (table 1). Each programme represents collaboration between the ACC and partner organisations that share a commitment to measuring and improving cardiovascular care. The SCAI is the partner for the CathPCI Registry, with representation on CathPCI committees as described below. Oversight of all NCDR programmes is provided by a management board, which reports to the ACC Board of Trustees.

CathPCI DATA

Since 1998, the CathPCI Registry has collected data on more than 12 750 000 catheterisation laboratory visits, 8 175 000 cardiac catheterisations and 4 596 000 PCIs from institutions of various sizes and geographic locations representing community and academic centres within the USA. The Registry is designed to include all consecutive adult patients who undergo PCI in participating facilities as the participant contract stipulates that all patients receiving PCI will be reported. To date, the Registry has not mandated the inclusion of all diagnostic catheterisations primarily because the interest has been on improving quality in PCI and because of the burden of data entry given the larger number of diagnostic procedures than coronary interventions performed. Facilities differ on the proportion of patients undergoing diagnostic cardiac catheterisation who are entered into the Registry: approximately two thirds of participants submit data on diagnostic catheterisation and PCI, and a third submit data on PCI only. The facilities that submit all diagnostic cardiac catheterisations are able to use the reports to evaluate care and outcomes in this population.

Data collection, validation and submission are the responsibility of a Registry site manager at each institution. Depending on the procedure performed, data collected include: patient demographics, clinical data on the episode of care, medical history and risk factors, laboratory studies, procedural information, intraprocedural and postprocedural events, and discharge information. Data elements are assessed for inclusion based upon the extent to which they contribute to the calculation of performance measures, risk-adjustment algorithms, procedural appropriateness or inhospital outcomes. The current CathPCI database V.4.4 consists of a total of 252 data elements (see online supplementary appendix B).

Data quality: accuracy, validation and completeness

The NCDR uses several approaches to educate participating facilities about the importance of high-quality data entry. This includes educational sessions on data definitions at the annual user group meeting, participant workshops, webinars, a participant training manual and newsletters as well as access to clinical support staff. To ensure that high quality data are included in the Registry, the NCDR has developed the DQP to ensure the completeness, consistency and accuracy of data submitted. The DQP consists of three main components: (1) a data quality report; (2) internal quality assurance protocols; and (3) a yearly data audit programme. This programme has recently been described in detail.4 In brief, registry-specific algorithms that require predetermined levels of completeness and consistency for submitted data fields are used to examine data submitted by a participant. A report is provided back to the facility after each quarterly data submission. If submitted data do not satisfy the completeness criteria, participants are informed and may clarify their data and resubmit until completeness criteria are met.

Beyond the data quality report, internal quality assurance algorithms are applied at each transactional step to ensure that data used for reporting and research are error free. An independent review of all data output is performed to ensure data are complete and appropriately formatted. Upon creation of the final analytical file, data are again reviewed to confirm that any created variables are coded correctly. In the final step, analytic output tables are compared with original data to ensure they correctly reflect the specified requirements. Only after passing these quality steps are data and reports released.

Finally, annual audits are conducted to assess data validity and reliability. Each audit is focused on identifying inaccurate data entry and opportunities for improvement through training or further documentation at individual facilities. More than 50 fields are audited each year, with some fields rotating in and out in a 3-year cycle. A total of 625 records are audited annually at 25 randomly identified sites. With the implementation of strategies such as remote auditing, the scope of the audit is continually expanding with respect to the number of sites and records included. Michigan, Massachusetts and West Virginia conduct more extensive audits as their data are used for statewide reporting programmes.

Data reports and use

CathPCI Registry participants receive comprehensive feedback about patient care variables and quality measures as quarterly and annual Outcome Reports. These institution-specific reports facilitate continuous monitoring of quality improvement efforts and enable facilities to compare, on a blinded basis, their site's practice patterns to national averages and volume-based peer comparison groups.

The Quarterly and Annual Outcome Report contains two major sections: the Executive Summary metrics and the details of each variable. The Executive Summary metrics have expanded over time to include the processes of care metrics, risk-adjusted outcome metrics and more recently, appropriate use criteria (table 2). These metrics are presented against national benchmarks, allowing participants to understand their data in context (figure 2). The Executive Summary metrics are further stratified in the detail section for specific patient groups (eg, patients with ST segment elevation myocardial infarction).

View this table:

• View inline
• View popup

Table 2

CathPCI Registry executive summary metrics

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

‘Box and Whisker’ plot used in the CathPCI Registry reports to participants, which identify participant performance (Arrow) compared with the rate representing all participants.

To provide more timely and customised data feedback, the CathPCI Registry recently introduced the ‘NCDR-Dashboard’ available to participants on a secure website (figure 3). The dashboard includes interactive reports that allow participants to interrogate each metric of the Executive Summary in more detail, including longitudinal data trends. The dashboard also provides functionality that allows participants to compare their site's metrics with other participant groups (reported anonymously) according to a range of site characteristics, including teaching status, PCI volume, hospital location and bed size. Participants can use this information to track the results of local quality improvement programmes and to perform analysis of individual cases. In some cases, CathPCI Registry participation also satisfies the requirements of regulatory bodies or conditions of participation of payers. The NCDR Analytic and Research Services group contracts with some state governmental agencies and payers to report data on hospital performance for quality improvement (see online supplementary appendix C). The NCDR Management Board reviews these contracts for approval before approval of any activity involving CathPCI data.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

National Cardiovascular Data Registry quality dashboard, which allows participants to customise data reports.

Given the number of procedures included over almost 15 years, the CathPCI Registry is a potent research tool that permits analysis of clinical treatments, procedures and outcomes in patients with coronary artery disease who are subjected to diagnostic cardiac catheterisation or PCI procedures. Data collected through the Registry can be analysed to assess compliance with clinical guideline recommendations, assist in medical decision-making and assess the appropriateness of medical care provided for patients with coronary artery disease.

Publication and presentation of CathPCI DATA

CathPCI data have been used for research by physicians, government and other entities with permission from the NCDR and oversight from the CathPCI R&P Subcommittee. Datasets used in these instances are always aggregated and created subject to HIPAA and HiTECH laws. Once the research is approved the investigator and R&P Committee submit a request for data analysis to one of the NCDR Analytic Centers. To date, more than 450 research proposals have been submitted for scientific review and approval. From these, 179 abstracts have been presented at national scientific meetings and 100 manuscripts have been published in peer-reviewed journals. An updated list of active and completed research projects is available on the NCDR website (http://www.ncdr.com/cathpci). Access to NCDR data for research is by request using a web-based application process. Applications for research proposals must conform to the policies and procedures for conducting research using NCDR data.

Strengths and weaknesses of NCDR data

Although there is no specific mandate for participation, there is strong professional support for CathPCI. A major strength of the registry is that it includes approximately 90% of PCI centre participants and more than 90% of PCI procedures performed in the USA representing the treatment of a broad spectrum of patients in diverse hospital settings. Some US states and insurers have integrated the CathPCI Registry into their cardiovascular quality assessment and improvement programmes, thereby increasing the use of the registry. Second, these are clinical rather than administrative data and thus less subject to the inherent limitations of administrative datasets.5 Third, data are collected using standardised definitions and subjected to rigorous data control actions. These measures enhance the consistency and accuracy of data and thus enhance the extent to which these data can be used to provide valid insights into the care and outcomes of patients undergoing PCI. Finally, these data have been used develop risk-adjusted mortality and bleeding models which are highly useful to clinicians.6 ,7

Limitations of the CathPCI data include:

1. Variable reporting of angiography procedures without PCI across institutions. Thus, the population of patients for whom data on angiography alone are collected may be less representative.

2. Some data elements that could inform risk-adjustment models may not be included in the registry.8 The CathPCI Steering Committee continually evaluates the registry DCT with a consideration including additional variables to enhance risk adjustment models.

3. Linkages to longitudinal data have to this point been probabilistic rather than deterministic, and only include those patients who are represented in the administrative sources from which the longitudinal data derive (eg, linkages with CMS administrative data provide details for patients who are covered by fee-for-service Medicare insurance).

4. Direct audits are only conducted for relatively small proportions of sites and records. However, the scope of the audit continues to expand as described above.

5. Reports typically lag data collection by 6 months, thus limiting rapid assessment of the implementation of quality improvement or other process changes at a hospital.

6. The absence of a unique patient identifier, limits the ability to track repeat procedures and longer term outcomes and to couple the data with payer and other electronic records and databases.

The future of CathPCI

The CathPCI Registry was established as a platform for characterising and improving the quality of care for patients undergoing coronary angiography and PCI. As the registry has expanded to include most US facilities its role has also expanded. Efforts already in process or planned include:

1. Further collaboration with the US Food and Drug Administration to perform postmarket surveillance of devices employed for coronary angiography and PCI. Several projects related to vascular closure devices have been completed.9–11

2. Linkages with other NCDR registries, including PINNACLE, the outpatient registry, which will allow for longitudinal follow-up using clinical data rather than claims data.

3. Opportunities for Maintenance of Certification and Lifelong Learning using CathPCI Registry data for practice improvement activities.

4. Voluntary hospital-level public reporting programmes using CathPCI Registry metrics.

5. Use of the CathPCI Registry as a platform for national quality improvement efforts and recruitment for pragmatic clinical trials and comparative effectiveness studies. One such study has already been completed.12

6. Development of structured catheterisation laboratory reports to serve as a national standard.

7. International collaboration, which will provide important insights into differences in patient characteristics and management strategies among countries, reduce unnecessary duplication and accelerate quality improvement activities in other countries.