Difference between revisions of "Obtaining and Maintaining Registry Data"

Latest revision as of 17:04, 26 February 2014

Evaluating Information

With any complex endeavor, such as a registry or biobank, we often gather information from multiple sources, including the Internet, journal articles, our advisors, and others in the biobanking field. But how do we know that the information provided is accurate? How do we know the information is trustworthy?

One strategy is to use Genetic Alliance’s Trust it or Trash it? tool to ask the following questions:

Who said it? – If it’s a written piece, it’s important to consider who wrote it, where the facts came from, and who paid for it.
When did they say it? – It’s also important to know when something was written or updated, especially in fields that change very quickly.
How did they know? – It’s also relevant to determine if the information pertains to you, and if it seems reasonable based on what you already know.

While the Trust it or Trash it? tool was designed to evaluate written information, it can also be modified to evaluate conversations you have with others. Remember, if it sounds too good to be true, it probably is.

Processes and Procedures for Registry Data

It is important to develop processes and procedures for the use of your registry. This includes designating personnel to enter data, assure data integrity, and maintain the registry. The Institute of Healthcare Improvement recommends the following:

Develop tools for collecting data, using electronic systems when possible
Train people who will be collecting and entering data into the registry (use test cases)
Create a process for data entry
Schedule and produce reports
Ensure system security
Make sure that people who need information from the system can get it in a timely manner
Make sure staff are trained on data entry, backup, and security procedures
Establish system backups

Strategies for Data Entry

Many registries use online data capture forms where the participant, a parent, or a provider enters the data. In special cases (e.g. lack of internet access), the participant will complete a paper questionnaire, and someone within your organization must enter the data manually. It is important to develop protocols for data entry that include the information collected, how it is collected, and how it will be entered. These protocols should be written down and included as part of a formal training for all who encounter data.

When entering data, be consistent with both the data and the process. A checklist is a useful tool to ensure data are entered the same way regardless of who completes the entry. You should also periodically check data you are entering against the paper questionnaire for accuracy. When you have finished entering data, verify all information is complete and review the entered data for errors.

It is good practice to have one individual enter data, and another review data. Be sure to capture who entered the data and when it was entered, as well as who reviewed it. It is also a good exercise to have two team members enter the same data and check its reliability. Finally, be sure all data entry personnel are given adequate time to enter data.

Audit Trails and Data Access

In any registry, it is vital to ensure that only authorized users can access your data. One strategy is to include a fine-grained hierarchy of authentication with users, groups, permissions and roles. A site administrator can create users and groups, and establish specific roles and permissions for each user or group. In addition, each user´s privileges should be verified by the system prior to displaying or changing a data point. Authentication logs should also be maintained to show changes to permissions and access, as well as changes to all data fields. Your software vendor should have protocols for chain of custody information and audit tracking in place to monitor data access, and activity reports should be available upon request.

Data Curation

The usefulness of a registry is directly related to the quality of data within the registry. Data quality can be improved through appropriate questionnaire design, data standardization, and data curation. Data curation is the “the active and ongoing management of data through their lifecycle of interest to science.”1 Once data are entered, it is essential that data are curated.

One strategy for data curation is to review and approve each completed questionnaire. Many registries provide the opportunity for a completed questionnaire to be temporarily stored and then migrated to the database once it has been reviewed and approved. This provides an opportunity to correct any ambiguous or erroneously entered data. Each questionnaire is unique, but there are some common elements that warrant examination.

Have all questions been answered?
Do answers have the appropriate format? (e.g. Are dates entered correctly?)
Do the answers make sense? (e.g. Are questions about pregnancy only completed by women?)
Are there specific questions vital to the study that should be reviewed?

If necessary, the registry coordinator can contact the individual who completed the questionnaire for clarification. If there are particular questions that users consistently have difficulty answering, you may want to modify those questions. Starting with high quality data is an important step for building a better registry.

View the whole science daily release.

Backing up data

It is important to develop policies and procedures for backing up data in your registry. If using a registry vendor, a reputable vendor will have backup policies and procedures in place. Additionally, you will need to develop an adequate back-up strategy to protect information that is accumulated prior to being entered in the vendor’s software as well as important internal files about your registry.

Some things to consider when backing up data:

Determine when and how frequently backups are done. You will want to schedule backups based on how frequently data changes. For example, data that changes weekly may not need to be backed up every day. Data that changes hourly will need to be backed up several times a day. You can stagger full back-ups with incremental backups to capture those files that have changed since the last complete backup.

Determine who is responsible for backing up data. Integrate backing up data into workflow by scheduling it as a required task. You may also want to keep a manual log to verify that backups have been completed as scheduled.

Determine where you will store your backup data. It is usually standard protocol to store backup data offsite. Consider the location. For example, if you live in an area where flooding is common, consider storing back-ups on a higher floor and not in the basement.

Test your backups periodically to ensure your procedures are working properly. By restoring your data in an alternate location, you will be able to determine if your backups are complete. If they are not complete, you will be able to adjust your protocol.

Experts suggest having three different backups of your data. Using the cloud is an additional option for backing up important data.