Read PearlDiver FAQs For General Queries and Information

PearlDiver Mariner Patient Claims Database (PearlDiver Technologies, Colorado Springs, CO, USA).

Mariner currently includes 165 million patients in total. 

All “Age” fields are based upon the year the patient was born. Age is handled in three different options.

“Age” can be used as a filter, or statistical variable, and is based upon the year the patient was born. “Age” represents the patient’s age in years at the time of event, as opposed to a range of years.

“Age Range” is also based upon the year the patient was born and captures the specified range of patient ages at the time of the claim. This is available as a variable in statistical queries, cross tab reporting, and the breakdown report.

“Age Code” allows you to use the code that corresponds to the “Age Range”. This can be found in filtering.

Yes. Older age groups have significantly lower counts. To protect patient privacy, these groups are rolled into the lower age group of 70-74 for the years 2010-2013, 75-79 for the years 2014-2017, or 80-84 for 2018-2023. This is done so patient privacy is maintained without losing the ability to perform research on these patients.

Each row of the breakdown report will report on the claims of the referenced population that meet the demographic criteria of the respective row. Patients with more than one claim may meet the criteria of multiple rows and will be reported in each row. Use of the ‘first instance’ or ‘last instance’ filter(s) will by definition limit the patient claims to a single claim. Patient counts by demographic will then add up to the total patient count.

Mariner contains all payer types including commercially insured, Medicare, Medicaid, self-pay, and government plans.

Clinic – Services provided to patients in an outpatient setting that is often a free standing or satellite facility. Urgent care facilities, renal dialysis centers and federally qualified health centers are all examples of clinics.

Inpatient - Services provided during admitted stay in hospital or long-term facility.

Office - Services provided to patients in office setting where one or more medical doctors, usually general practitioners (GP), receive and treat patients.

Outpatient - Services provided to patients in outpatient rehabilitation facility, hospital - outpatient surgery, skilled nursing – outpatient, ambulatory surgical center, or hospital - emergency room.

We are always happy to assist with your questions. The best method to send us a question is to select “Report Issue” from the query progress screen or from the Help menu while creating your query in Bellwether.

Using “Report Issue” allows you to write a message to us, and also automatically sends us a copy of your query, any results generated from that query, and information regarding the data set you are running the query in. This allows us to get started troubleshooting right away and eliminates most back and forth via email to collect all necessary information.

For procedures that can occur inpatient, include both the ICD procedure code as well as the CPT in your population definition. In a perfect coding world, both of these codes should show up in the form of a facility claim and a physician clam. But in the coding world, the reality is that it is common for there to be only one of these. To maximize the sample size, include both.

Example:
Population1 = {CPT-12345, ICD-9-P-1234, ICD-10-P-12345}

While CPT codes do not have modifiers in Mariner data, searching for laterality in the new data set is possible using ICD 10 code definitions for inpatient procedures, as well as ICD 10 diagnosis coding for diagnosis codes accompanying the procedure.

The Mariner data sets do not contain information on mortality. The suggested approach currently is to use the sudden/unexpected death ICD 9 & 10 diagnosis codes. While these codes may not capture all patients passing outside of provider care, these patients have definitively expired. These codes include: ICD-9-D-7980, ICD-9-D-7981, ICD-9-D-7982, ICD-9-D-7989, ICD-10-D-R99.

One additional option is to include hospice treatment codes to capture patients who have entered end-of-life care. This is a close proxy to expiration. These codes include: CPT-T2042, CPT-T2043, CPT-T2044, CPT-T2045, CPT-T2046, CPT-G9751, CPT-G9812, CPT-G9852, CPT-G9855, CPT-99377, CPT-99378, CPT-G9473, CPT-G9475, CPT-G9476, CPT-G9477, CPT-G9478, CPT-G9479, CPT-G9524, CPT-G9525, CPT-G9687, CPT-G9688, CPT-G9690, CPT-G9691, CPT-G9692, CPT-G9693, CPT-G9694, CPT-G9700, CPT-G9702, CPT-G9707, CPT-G9709, CPT-G9710, CPT-G9714, CPT-G9715, CPT-G9718, CPT-G9720, CPT-G9723, CPT-G9725, CPT-G9740, CPT-G9741, CPT-G9761, CPT-G9768, CPT-G9857, CPT-G9809, CPT-G9858, CPT-M1022, CPT-Q5004, CPT-Q5005, CPT-Q5006, CPT-Q5007, CPT-Q5008, CPT-Q5010, CPT-S0255, CPT-S0271, CPT-S9126.

Just as in previous data sets, LOS is only on the inpatient claim (ICD 9/10 procedure code). To ensure an accurate LOS, you will want your population that is to be filtered or reported on to contain only the inpatient claim. Should you have a mix of code types in your population bucket, it is recommended you first filter for the inpatient claims only. This would look like:

LOSgroup = filter base codetype = {ICD-9-P, ICD-10-P}

When using the ‘Totalcost’ query, many researchers like to break out the results based upon the source of the claims. Typically, this is separated into Rx, inpatient claims, outpatient/physician claims. To separate the reimbursement amounts out in this manner, you can run a ‘Totalcost’ query and then filter based upon these groupings.

Example: Total reimbursement within 90 days following a procedure

Kneetcwithrequired = totalcost Requireboth start = 1 day after {Requireboth} end = 90 days afte
Kneein = filter Kneetcwithrequired chargetype = I
Kneerx = filter Kneetcwithrequired chargetype = R
Kneeother = filter out Kneetcwithrequired chargetype = {I,R}

Yes, this can be done using the ‘min’ and ‘max’ commands. This would look like:

allmin = MIN bucketname
allmax = MAX bucketname
tbortho = TIMEBETWEEN bucketname {allmin}{allmax}