1

Creating a Library of Patient Registry Metrics for Quality Improvement

Justin Tran

Health Care Informatics - University of San Diego

HCIN 548: Health Care Informatics Seminar

Professor Dorothy O’Hagan

May 20, 2022

Creating a Library of Patient Registry Metrics for Quality Improvement

Abstract

A patient registry is a system of organization that collects and aggregates data from a

patient population defined by a particular condition, known as a metric. My current healthcare

organization has just over 170 patient registries, each with its own set of registry metrics, but

does not currently have a system for grouping and displaying relevant metric data. The goal of

this project is to create a library that displays each of the system’s patient registry metrics with

pertinent information. A SQL query is to be created that extracts data from the organization’s

Epic records and placed into Collibra, a data cataloging and governance application. The catalog

will allow providers to verify the accuracy, validity, and cleanliness of available data in patient

registries. It will also permit users to revise registry metric data by removing or updating

metrics that are old or outdated. This project will be leveraged for use with direct patient care,

population health reporting, and clinical research.

Introduction

As the healthcare industry continues to shift to more technological means and the sheer

amount of health data exponentially rises, it has become increasingly more important for

healthcare organizations to be able to maintain and organize this information. One method,

often leveraged by hospitals to store and categorize groups of related patients, is known as a

registry. A clinical or patient registry is an structured dataset of an organization’s patients that

are grouped based on a particular diagnosis, medication, chronic condition, etc., known as a

metric. Examples of registries include patients with type II diabetes or patients that have tested

positive for COIVD-19, whereas examples of registry metrics would include a specific ICD code
3

or performed procedure. Registries can be leveraged by clinicians for several different

purposes, ranging from direct patient care to population health reporting to clinical research.

Any number of metrics can be used for consideration into a registry, and as a result, large

healthcare systems can have hundreds or even thousands of unique patient registries, each

with their own set of registry metrics.

The basis of my Capstone project is to create a catalog that organizes the registry

metrics found in patient registries at my healthcare organization. The organization currently has

over 170 patient registries that contain data for certain patient groups, such as those with

chronic conditions or specific comorbidities. Each registry is defined by metrics that searches

through patients’ electronic health records (EHRs) in their Epic records to place them in certain

groups. The library of metrics will allow users and researchers to find particular registry metrics

more easily, make edits or updates to them, or eliminate ones that yield null or missing values.

The catalogue will be used for both population health researchers as well as daily patient care.

Problem

My healthcare organization has seen huge growth in the number of patients and thus

the amount of patient data in recent years, now with over 170 unique registries that group

patients together based on certain metrics. They do not currently have a tool or catalog that

organizes, tracks, and displays important information pertaining to each of its registry metrics.

It is now necessary to have one in place to ensure each registry is drawing the correct patients

and information from its EHRs, each metric is updated and revised, and to create metrics that

make data retrieval more streamlined.

Goals/Purpose

The goal of this project is to create a library/catalogue to organize and track each metric

in over 170 patient registries. The project will help ensure the accuracy, validity, and cleanliness

of available patient data. The catalog will also allow users to perform more efficient revision of

registry metrics. The first example of this would be removal of old or inactive metrics. As

healthcare organization are emphasizing multiple, incremental system changes throughout the

year as opposed to one or two large updates to their systems, it is more likely that standards

can be changed or updated, rendering some registries obsolete and in need of removal. The

second revision method for metrics is updating those that are outdated, for instance, a smoking

registry that does not yet account for e-cigarette use and must be changed to reflect something

that is more common in society. The last revision method would be to review and correct

metrics that are misplaced in a particular registry or that need to be added to one.

Literature Review

With the proliferation of available healthcare data and the increased popularity of

patient registries, a method for standardization of registry metrics is just as important as the

registries themselves. This brief literature review will dive into some of the published literature

pertaining to patient/clinical registries, their metrics, and the standards by which they should

be organized. Some of these concepts will be applied in the Capstone project.

Gliklich et al. (2014) defines a patient registry as “an organized system that uses

observational study methods to collect uniform data (clinical and other) to evaluate specified

outcomes for a population defined by a particular disease, condition, or exposure and that

serves one or more pre-determined scientific, clinical, or policy purposes”. Patient registries
5

offer the potential to collect deep, nuanced data on large numbers of patients at reduced costs

and support population health, clinical decision making, and quality improvement. The article

explains that while registries and electronic health records are different in terms of purpose

(registries are patient-centered and purpose driven, while EHRs are visit-centered and

transactional), the EHR supplies a myriad of data for registries to tailor toward patient care

from both proactive and retroactive perspectives.

In terms of organization and standards, the EHR contains many different data points and

types that can be considered metrics for inclusion in a patient registry, including demographics,

particular diagnoses or medications, procedures, etc. These common data elements, or CDEs,

can be used in registries to support linkages and aggregations with data from other studies or

registries, essentially increasing the data pool for a particular group of patients. General or even

disease-specific CDEs are often grouped together to create a minimum set of data elements for

inclusion into a patient registry. Having a standard group of metrics across multiple registries

can make the data more easily understood as well as more clearly define inclusion into a certain

group of patients. Registries should also be able to capture patient outcome measures, though

the standardization of those metrics are less common. Being able to track patient outcomes

amongst those with a particular condition will make it easier for users to determine trends and

aggregate results.

Similarly, the Improving Healthcare Interoperability - Project Report, written by the Duke

Clinical Research Institute (2018), suggested recommendations toward the optimization of

clinical registries for the overall improvement of care. The writers acknowledge that registries
6

are useful in that they produce actionable data for quality improvement and patient safety,

thus they should be a point of emphasis when it comes to healthcare interoperability.

The article found that the community of patient registry users has not yet benefitted

from and is not aligned with interoperability efforts. For instance, 11 patient registries had

captured information about tobacco use, but each registry used a different set of questions,

and only one of them used a standardized set of questions regarding tobacco use. It was found

that the predominant model for obtaining registry data was still overwhelmingly forms-based

manual chart abstraction and data re-entry, rather than electronic data capture at the point of

care. It highlights the necessity for not only a shift to more electronic and streamlined means of

data collection, but also a standard process for organizations to collect similar data points and

types. Their recommendations to improve the usefulness of patient registries within and

amongst organizations include local standardization of data collection which includes “pre-

cleaning” of datasets, improved capture of a set of CDEs at the point of care in the EHR system,

packaging of data specific to a registry for ease of transmission, and more widespread adoption

of electronic collection formats. These recommendations are all easier said than done, but the

framework is in place for healthcare organizations to be able to be to refine their patient

registry organization and ultimately improve care.

Project Recommendations/Scope

The recommended plan for this project is fairly simple in nature but more complex in its

execution. A SQL query will be developed that extracts the following information pertaining to a

registry metric: Rule ID, Rule Name, Description, Registry Metric Groupers (connections to

other registries), Registry ID, Registry Name, Metric Active Status, and Latest Revision Date. This
7

data will be taken from available records in the organization’s Epic system and placed into an

Excel spreadsheet to verify that all the needed information is present, is correct, the right data

types, etc. The SQL query presented below is currently under revision.

Figure 1. SQL query to extract information from Epic records

The completed SQL query containing all the necessary registry metric data will then be

uploaded into Collibra, which is a data catalog application that “empowers businesses to find

meaning in their data and improve business decisions” (UC Davis Health, 2022). Collibra is a tool

that can automatically extract data from Epic applications and display relevant information in

an understandable format. Collibra allows for regular checking and maintenance of metrics and

has a number of capabilities, including the organized display of a data point’s metadata and

lineage to other relevant information for use by stakeholders and providers. Collibra is also

based on structured data governance and privacy regulations to remain compliant and ensure

the safety of large amounts of patient information. The layout of data in Collibra is like that of a

relational database (see below), with a registry metric’s data in the leftmost box and its

connections to other registries moving to the right.

Figure 2. Collibra example layout, retrieved from

https://www.collibra.com/wp-content/uploads/Diagrams1. jpg

Following data extraction and uploading into Collibra, an extensive review must be done

that checks the presented registry metric data back to the Epic system to ensure its validity and

accuracy. The completed registry metric catalog will be accessible for use by providers, analysts,

researchers, etc.

Deliverables and measurables for success in this project include the creation of a

holistic, structured framework displaying all relevant registry metric information, the ability to

clean up metrics that present an error, such as a null or zero value, and increased user

satisfaction for those accessing the data for patient care or clinical research.
9

Timeline/Work breakdown structure

Project Challenges

The first challenge in this project is the fact that Collibra is a new application to the

healthcare organization, so there is no previous example or work upon which to base the

project’s vision. As a result, there may be some trial and error or some troubleshooting in order

to achieve the project’s goals. Another challenge that has come up in early discussion has been

where and how to extract information from Epic. The metric information that is wanted for this

project do not lie in one particular Epic application, so while some of the data can automatically

populate from Epic directly to Collibra, data elements such as latest revision date may need to

be manually entered into the application. The alternative to this would be the creation of a new

column in Epic that can populate and report revision date automatically, but criteria for revision

may need to be established.

Conclusion/Next Steps

There are several recommended next steps for this project. The first would be a manual

review of each of the extracted registry metrics from Collibra back to its location in Epic to

validate the accuracy of the data. This will be a time-consuming step but is necessary in order to

verify that the data that providers will be using for a variety of projects and work is accurate

and current. The second step would be the establishment of a system for regular maintenance

of individual metrics or for groups of related metrics to ensure they are up to date and the

connections are correct. As mentioned previously, organizations are shifting towards more

regular updates to their systems, which include changes to codes, protocols, regulations, etc. By

having scheduled checks on groups of registries throughout the year, the organization can more

efficiently and effectively address changes that must be made to a registry or its metrics. This

will ensure that the organization remains compliant with standards and improve patient safety.

Going further, this project falls in line with the organization’s increased emphasis on

Healthy Planet, which is a software module from Epic that uses aggregated patient data to

optimize the delivery of health care (Cummins, 2020). Using Epic features such as its patient

registries allow for providers to better coordinate care and personalize treatment based on the

data provided by the organization’s patients. Healthy Planet will also allow an organization to

more closely and accurate monitor care quality and costs. Increased use of these population

health tools will improve patient outcomes and the overall success of the organization.

Conclusion

In the end, this project will enhance the overall operations of my healthcare

organization by improving the quality and accuracy of patient data for direct care, population
11

health reporting, and clinical research. The library of registry metrics will make it easier for

providers and stakeholders to view relevant information on certain groups of patients and

assist with the decision-making process to ultimately increase the number of positive patient

outcomes and have a beneficial impact on the community. This Capstone project will be just

the start of a process to improve the organization’s operations to be more efficient, patient-

oriented, and useful for all involved.

References

Cummins, R. (2020). Epic’s Healthy Planet helps caregivers fill patient care gaps. The University

of Mississippi Medical Center.

https://www.umc.edu/news/News_Articles/2020/08/Healthy-Planet.html

Datameer. (2022). What is Collibra?. https://www.datameer.com/what-is-collibra/

Duke Clinical Research Institute. (2018). Improving Healthcare Interoperability – Project Report.

Pew Charitable Trusts.

Gliklich, R. E., Dreyer, N. A., & Leavy, M. B. (Eds.). (2014). Registries for Evaluating Patient

Outcomes: A User's Guide. (3rd ed.). Agency for Healthcare Research and Quality (US).

Lyu, H., Cooper, M., Patel, K., Daniel, M., & Makary, M. A. (2016). Prevalence and Data

Transparency of National Clinical Registries in the United States. Journal for healthcare

quality : official publication of the National Association for Healthcare Quality, 38(4),

223–234. 

UC Davis Health. (2022). Collibra. University of California, Davis.

https://health.ucdavis.edu/data/collibra.html
13

Appendix

The skills, tools, and expertise expected of this Capstone project meets several of the

competencies set forth by the Hahn School of Nursing Health Care Informatics program, three

of them being Health Science Knowledge and Skills, Data and Knowledge Management, and

Quality and Regulatory. The project required an understanding of terminology and language

related to the healthcare field, such as registries, metrics, and the different conditions that my

define inclusion in a particular patient population. In learning about the project’s topic and

current problem, it was extremely useful to have a grasp on the systems involved in this type of

analysis, including population health reporting, Epic systems, and chronic disease drivers.

Taking an understanding of health science into account also made it easier to conduct

and evaluate the technical skills necessary for the project. The members of the team assisting in

the project were able to demonstrate for me the capabilities of data management tools such as

Excel, SQL, and Collibra to more efficiently work toward the project’s goal. The analytical

techniques that this project calls for helped build upon my previous knowledge from

throughout the program to better assist me in future projects. Lastly, the project addresses an

understanding of quality and regulatory competencies when it comes to the use and

governance of large amounts of patient data, creating a system that is compliant and secure,

and understanding how private health information is displayed, used, and shared safely within

a healthcare organization. While the Capstone project likely meets each program competency

in some way, these three are the proficiencies that stand out to me as the ones that guided the

scope and direction of the project as a whole.