Patient Registry Restart: One Step Back, Two Steps Forward

At a glance

Assess the current state

Examine the data that have been collected, how well they've supported the registry's research aims, and the operational factors that may explain the findings. Account for any interruption in data collection.

Reevaluate the methods

Identify opportunities to improve collection procedures, simplify workflows, reduce burden, strengthen participant engagement, or introduce new capabilities.

Prepare to restart

Determine what needs to change in the database, collection procedures, and schedules, then prepare participating sites and staff to resume operations.

Sometimes patient registries go through a reset. These situations generally fall into two categories:

Restarting after an interruption. Registries occasionally suspend operations for various reasons, such as organizational changes, funding interruptions, or administrative reviews.

Reevaluating an active registry. A registry may undergo a planned reassessment as research questions evolve, data collection for certain variables or domains comes to an end, new technologies become available, or opportunities arise to improve operations.

In either case, the restart challenge is determining what's involved in moving the registry forward.

One Step Back: A Little Sleuth Work

Restarting or reevaluating a patient registry can feel a little overwhelming. A good place to begin is by taking a step back and doing a little sleuth work.

The goal is to develop a fair and accurate picture of the registry's current state, recognizing both what's working well and what could use improvement. That assessment provides the foundation for informed decisions about what needs to happen next.

Examine Existing Data

Data are the foundation of any patient registry, so getting your arms around what's been collected is the best place to start. This involves examining both the quality of the collected data and how those data have been used to address the registry's research questions.

Assess data quality. Examine variable values using descriptive statistics such as ranges, frequencies, variance, skewness, and kurtosis. Look at these characteristics both overall and across key grouping variables, such as study sites and participant groups.

Look for:

• Invalid values. Values outside expected ranges or inconsistent with other recorded information.
• Outliers. Unusual observations that may reflect errors or meaningful differences.
• Restricted variance. Variables showing little variation, including floor or ceiling effects that may limit their usefulness.
• Missing values. Patterns of incomplete data, including high rates of missing or "prefer not to say" responses.

Evaluate the empirical aims. Revisit the registry's original research questions and planned analyses to assess how well the collected data have supported its research objectives.

• Completed aims. Which research questions have been addressed through completed analyses or publications?
• Remaining aims. Which research questions and planned analyses have yet to be addressed?
• Data sufficiency. Are the accumulated data sufficient to answer the remaining questions, including adequate statistical power, or is additional collection needed?
• Continued collection. Which variables or domains have served their intended purpose, and which remain important for ongoing analyses or longitudinal follow-up?

Operational Review

The operational review is intended to explain patterns revealed by the data evaluation. The goal is to understand which operational factors may be contributing to those findings.

Consider several areas:

• Variable-level factors. Are definitions or collection instructions unclear? Could differences in interpretation, response options, or assessment timing explain unexpected values, missingness, or unusual distributions?
• Site-level factors. Are there differences in procedures, resources, or how assessments fit into clinical workflows?
• Staff-related factors. Could differences in training, experience, staffing, or adherence to collection procedures explain some of the observed patterns?
• Collection-burden factors. Are certain variables or assessments difficult, time-consuming, or inconvenient for participants or staff to complete?

Understand How Collection Was Affected

If registry operations were interrupted, determine how the interruption affected data collection. The impact may range from no interruption to a complete stoppage, with varying levels in between.

The goal is to understand the impact on the registry's data and what will need to be accounted for when operations resume. Depending on the circumstances, this may involve:

• Backfilling existing information. Clinical observations, laboratory results, or other information may have continued to be collected outside the registry and can potentially be entered retrospectively.
• Documenting missing observations. Assessments that were never performed, such as time-specific participant-reported outcomes, may need to be identified and recorded as missing rather than recreated.
• Adjusting collection intervals. Some assessments may need an extended collection window or additional time for retrospective entry to account for information accumulated during the interruption.
• Accounting for differences across sites or domains. Collection may have continued in some areas but stopped in others, requiring different approaches to preparing the registry for resumed operations.

Reevaluate the Methods

The goal is to determine whether existing methods could be improved or new capabilities could make the registry more effective.

Consider several areas:

• Data collection. Could current approaches to collecting information be simplified or improved?
• Workflows. Are there opportunities to reduce unnecessary steps, duplicate entry, or administrative burden?
• Participant engagement. Could communication, education, follow-up, or other approaches improve participation and retention?
• New functions and capabilities. Talk with the people doing the day-to-day work. What additional features or functions would help them address recurring challenges, simplify their work, or accomplish things they currently can't?
• Technology. Could existing or emerging technologies support better methods, automate routine activities, or enable approaches that weren't previously practical?

Two Steps Forward

With a clear understanding of the registry's current state, the next challenge is determining what needs to be done to prepare it for opening up data collection again.

Get to the Starting Line

Before restarting data collection, determine what needs to be done to bring the database up to date and prepare it for ongoing operations. Depending on the findings from the review, several activities may be necessary:

• Historical data recovery. Determine what data remain available from the interruption period. Information generated through ongoing clinical care, such as medical records and laboratory results, may still be available even though registry collection activities stopped. Identify what can be imported electronically and what staff will need to retrieve and enter retrospectively.
• Unrecoverable observations. Identify data that would have been generated through registry-specific collection activities but were never obtained. For example, participant-reported symptoms or functional status assessed at a particular time cannot reliably be recreated months later. Determine how these observations should be documented as missing.
• Database configuration. Identify changes needed to variables, definitions, forms, validation rules, or other aspects of the database to accommodate retrospective data entry. For example, a form may need a missing-value option for assessments that could not be performed during the interruption, or validation rules may need to accommodate entries outside the original collection window.
• Collection schedules. Determine how to handle outstanding assessments, extended collection windows, or adjustments needed to resume regular follow-up schedules.
• Collection methods. Identify procedures, workflows, or capabilities that need to be updated before collection resumes.

Another consideration is whether new data collection will be introduced. If so, determine how it will fit into the registry:

• Existing collection schedule. Can new variables or assessments be incorporated into collection activities already taking place?
• New collection schedule. Will the additional information require separate collection methods, assessment intervals, or follow-up procedures?

Onboarding and Training

Preparing the registry for data collection also means making sure participating sites and staff are ready to resume operations.

Consider several areas:

• Staffing and responsibilities. Have personnel or responsibilities changed? Are new staff members joining, and do existing users have the appropriate roles and permissions?
• Onboarding. What needs to be done to bring new staff or participating sites into the registry? Are accounts, access permissions, and required resources in place?
• Training. What training or refresher sessions are needed, particularly for changes in data collection procedures, forms, schedules, or system functionality?
• Communication. How will changes be communicated to participating sites and staff? Do they know when collection will resume and what is expected of them?
• Rollout scope. Will data collection resume across all sites simultaneously or through a phased rollout? This will depend largely on the extent of changes to collection methods and the number of new or returning staff who need training.

Moving Forward

Break the work into smaller, actionable steps, establish priorities, and tackle them one at a time. With a little patience and steady progress, you'll come out the other end with a stronger, more viable patient registry.