Cost estimates are mainly driven by the number of fields desired, inclusion/exclusion criteria, and the accessibility of these fields in our databases. Sample size normally does not affect cost for cohorts. We quote our best estimate based on the information given, but this may not take into account time needed for questions and revisions.
If we exceed the original cost estimate, we will notify you and give you the opportunity to continue your project with a higher estimate or give you the option to scale down the scope of your project. Unfortunately, we are unable to provide fixed-cost estimates due to inevitable changes in scope throughout the course of a project.
Research projects evolve over time and we all need to remain flexible and understand that your original expectations may not match the actual data available. In order to effectively manage our queue, we enforce a one-month time limit on follow-up questions and requests for changes after the data has been delivered.
After the one-month time limit has ended, you can still come back with questions and requests for changes but your project will be placed at the back of the queue and may require additional charges for labor.
For this reason, we encourage you to provide a comprehensive list of all variables desired before we begin work on your project. More variables can be added after your data is delivered, but your project will be placed back in the queue and additional charges may be incurred.
EHR data is messy and it is not possible to know in advance exactly what the data desired will look like. EHR data entry practices vary and guidelines and structure of the EHR have changed over time, so the data that you request may not look as expected.
Sometimes researchers request certain data and are dissatisfied with what is available. For example, sometimes inpatient procedures are done at outside clinics, or the standard phrase used in a pathology report is altered and difficult to extract. Please understand that there will be a normal period of data discovery that occurs with every project and that this will take time to resolve.
It is straightforward to extract data structured into fields, but researchers sometimes ask that we pull out references to certain conditions or disease scores from clinical notes or reports. We can do simple text matching, but this is normally less effective. For example, a search for “spinal cord ischemia” may return “discussed risk of spinal cord ischemia” or “no spinal cord ischemia.”
One workaround is for us to extract a short amount of text around the keyword, but you will need to manually review these text stubs to determine the utility of the data provided.
For IRB Chart Review studies that cannot be fulfilled using the STARR tools and services, you may obtain a Research Use Exceeds STARR Capabilities form (fka “EHR Access Privilege for Research” or "STARR Insufficient") here.