A patient registry is an organized program for the collection of a clearly defined set of data on identifiable individuals for a specific purpose.8
Patient registries generally use observational research designs to capture data from sampled disease populations
to better understand disease etiology or to explore patient variation and experience among different treatment options. Registries can also be used for clinical trials recruitment,9
patient safety (e.g., post-marketing surveillance), and monitoring of provider performance relative to practice guidelines or target measures. Additionally, registries can support quality improvement, specifically with respect to outreaching to patients through letters, phone, or email. The term patient registry implies follow-up over time. While patient registries can support research, not all research data sets emerge from registry programs, and not all registry programs result in research-quality data sets. In type 1 diabetes, there are several independent registry efforts that have supported new understandings of the epidemiology and mechanisms of the disease. These include the Diamond study10
worldwide, the EURODIAB11
in Europe, and the SEARCH study/registry in the United States.12
In addition, there are many independent statewide surveillance programs and registries in the United States organized by state public health agencies.
There are inherent limitations of certain registry designs for certain functions, particularly in the exploration of research questions involving treatment evaluation.13,14
To be useful for research, a registry must have high-quality data. Two fundamental concerns related to gauging the quality of registry data include completeness of case ascertainment and validity of values for each data point.15
The completeness of case ascertainment can be assessed by comparing multiple sources. Authors in the United Kingdom found discrepancies between the prevalence of diabetes derived from epidemiological studies and those reported through the national quality improvement scheme.16
Another study found that linkage of multiple electronic data sources was significantly more sensitive than general practice registers in identifying both diabetic and high-risk subjects.17
Using a random sample of 125 charts of Medicare patients, Tang and associates18
also showed a major discrepancy between diabetes patients identified using clinical data captured in an EHR system and those identified from claims data, resulting in statistically significant differences in the quality measures for frequency of HbA1c testing, control of blood pressure, frequency of testing for urine protein, and frequency of eye exams for diabetes patients. Others have described the problem of different case definitions and EHR query strategies and are developing an electronic medical record standard definition of diabetes that can be used in different settings.19
In addition to standardized approaches to utilizing clinical data, it is critical to ensure valid and reliable data for all registry purposes, standardization of case definitions, data definitions, and clinical diagnostic criteria. Detailed examination of representative subsamples can also be conducted to validate large survey results.20
While EHRs can support both registry recruitment and data generation in a registry, the distinction between registries and EHRs must be kept clearly in mind. Registries, sometimes referred to as chronic disease management systems if they are used for that purpose, can serve a number of functions as previously described, but all are designed to collect data on populations
for predefined purposes, and the collected data are prespecified. A discussion of available registry products for diabetes management, including many variations of the Centers for Disease Control and Prevention-funded diabetes electronic management system, and guidance for different levels of integration with EHRs has been developed by the California Healthcare Foundation.21
In contrast to registries, the primary purpose for EHRs is to capture data on individuals
to support their care. There is some evidence that diabetes care can be improved with EHRs, and subsequent speculation that better EHR implementation models could improve the consistency and impact of these results.22
However, others have shown that widespread implementation of an EHR is not sufficient to improve the quality of diabetes care (as measured by provider compliance with process and treatment guidelines and intermediate diabetes patient outcomes).23
Lester and colleagues22
synthesized the limitations of information technology in diabetes care to date and provided eight helpful “rules” for designing informatics systems that catalyze change in diabetes care.
Institutions may have both an ethics board (institutional review board) and a privacy board; however, many institutions combine the responsibilities of the ethics board with the privacy board. Registries that send general information related to enrolling research studies or that allow the registrant a conduit through which they may opt to share their information with research sponsors are required to have privacy board [Health Insurance Portability and Accountability Act (HIPAA)] review and approval, as the authorization requirement may need to be altered or, more likely, a waiver is sought [45 CFR § 164.512(i)]. If the registry provides enrollment information on specific studies or research programs, then privacy board (HIPAA) review is generally required. (Ethics review for the registry itself is not required, although individual studies might describe the use of the registry in the recruitment plans of ethics-reviewed studies.)
Not all registries require ethics or privacy board review. Contact registries that do not promote or advertise research studies and that collect minimal patient contact information (e.g., name and email address) and minimal health information (e.g., diagnosis and age) can be thought of as information services. These information-sharing registries do not typically require privacy board or institutional review board approval, as these registries do not meet the federal definition of research (45 CFR § 46.102) nor do they typically require a waiver or alteration of the authorization requirement in section 164.512(i) of the privacy rule. However, any patient registry that is designed to support any type of research data collection (prospective systematic investigation) will require ethics review and approval and, therefore, privacy board review [45 CFR §§ 46.101 and 46.111(a)(7)].