We've made changes to Bonnie that should make the process of labeling population groups and stratifications during measure import clearer, which will be included in the July 24th release of Bonnie.

I've updated the summary to reflect the remaining open issue described in this ticket.

Email is insufficient to describe. Please provide time for a call.

1) Did the explanation about why the patients you listed are failing successfully clarify how Bonnie works?

2) Can you expand on what you mean by "I think there is a problem with how the strata uploaded to Jira. Patients who apply to the IPP and denominator don’t qualify in Bonnie because of the logic upload - and Bonnie won’t evaluate the numerator without the denominator being met"? Ideally, it would help to be pointed to a specific measure and patient with some detail about what you're seeing and what you would expect to see.

Note - I think there is a problem with how the strata uploaded to Jira. Patients who apply to the IPP and denominator don’t qualify in Bonnie because of the logic upload - and Bonnie won’t evaluate the numerator without the denominator being met.

Please advise.

Smith_Test_Too_Soon Test1 (1) (1) - Test performed before 15th birthday

This patient is failing in Bonnie because he is marked as expecting to qualify for the numerator (the NUM checkbox is checked in the EXPECTED VALUE section of the patient), but does not actually qualify for the numerator (as you say, the test was performed before the patient's 15th birthday).

Test1_No_MP_Visit John (1) (1) - no visit during 2012

This patient is failing in Bonnie because he is marked as expecting to qualify for the initial population, denominator, and numerator (the IPP, DEN, and NUM checkboxes are checked in the EXPECTED VALUE section of the patient), but does not actually qualify for any of these populations.

In all of these cases, it comes back to the EXPECTED VALUE section; this section is intended to allow the creator of the patient to specify which populations the creator intends the patient to qualify for. When building a patient where the patient is intended to qualify for the numerator, check the NUM checkbox. When building a patient where the patient is not intended to qualify for the numerator (as in the first patient mentioned here, where the patient received the test when too young), leave the NUM checkbox unchecked.

Measure: CMS349v0: HIV Screening
Newly created out of range patients:

Smith_Test_Too_Soon Test1 (1) (1) - Test performed before 15th birthday
Test1_No_MP_Visit John (1) (1) - no visit during 2012
Test1_Too-Old John (1) (1) (1) - patient over age 65
Test_very young Nancy (1) (1) - patient under age 15

I also don't understand why the strata defined in the measure come up failures as well since they should accommodate some patients who fit and some who don't.

Can you follow the steps again, and then tell me which measure and the name of the patient? I can take a look at the data loaded in your bonnie account, which let me see exactly what the problem is.

This doesn’t help. I did exactly what you suggested and the patient failed. I reset the age to inside the range and he passed. Similarly I did the same with a lab test performed too early and the patient failed. I moved the test date and the patient passed. This explanation doesn’t get me anywhere. I need to have a one-on-one discussion and review to understand. I already read the User Guide.
Can you suggest times to discuss - Wednesday at noon (7/2), Thursday after 1 PM ET?

Thanks
Floyd

1) Bonnie is intended to be used for testing the edge conditions of measures, ie patients who don't qualify for any of the measure populations. Let's say we want to construct a patient that does not qualify for the initial population because the patient is too old. Here are the steps you might follow to construct such a patient:

a) First, look at the "Expected Value" section, which is where we specify what populations we expect our patient to fall into; we expect this patient to fall outside of the IPP, DENOMINATOR, NUMERATOR, etc, so we leave all of these check boxes unchecked.

b) Second, set the age of the patient to an age outside the range specified in the IPP.

c) Third, save the patient; the patient should be passing.

Some notes: if the measure has multiple populations, expectations should be set for each of the populations (there is a tab at the top of the expected value section). Also, it's important to be clear about the distinction between a patient "passing" in Bonnie (that means the "expected values" set for the patient agree with the populations the patient successfully evaluates against) and a patient evaluating into the numerator of a measure (what you appear to mean by "passing" in your comments above).

2) Can you point to a specific measure and patient that you have loaded, where the patient meets the criteria but still fails, and I will take a look?