Student 1 Verification: Steroid-Induced Osteonecrosis (drugs)

[callahantiff]

Student (GitHub Username): @deringtonc Verification Number: 1

Verification Assignments: Steroid-Induced Osteonecrosis

• Phenotype Definition
• Spreadsheet

[deringtonc]

@trinklek @callahantiff for this phenotype, should we marking medications that can cause steroid-induced osteonecrosis as clinically relevant (i.e., any steroid)? Or do you only want us to mark medications that can be used for treatment of this condition as clinically relevant?

Example row: https://docs.google.com/spreadsheets/d/1Va7cin600gumAUxQ-MT6InADK87iIwpqR0Qmca22dCc/edit#gid=2090244794&range=2573:2573

[callahantiff]

@trinklek @callahantiff for this phenotype, should we marking medications that can cause steroid-induced osteonecrosis as clinically relevant (i.e., any steroid)? Or do you only want us to mark medications that can be used for treatment of this condition as clinically relevant?

Example row: https://docs.google.com/spreadsheets/d/1Va7cin600gumAUxQ-MT6InADK87iIwpqR0Qmca22dCc/edit#gid=2090244794&range=2573:2573

@deringtonc - Good question, in this case, given that the phenotype definition explicitly states that "Presence of ANY corticosteroid (no minimum dose IV, IM or PO ≥ 14 days cumulative) (drug strings)" I would argue that since this is explicitly criteria for identifying case patients that is is probably clinically relevant. @trinklek, what are your thoughts on this?

[deringtonc]

@trinklek @callahantiff for this phenotype, should we marking medications that can cause steroid-induced osteonecrosis as clinically relevant (i.e., any steroid)? Or do you only want us to mark medications that can be used for treatment of this condition as clinically relevant? Example row: https://docs.google.com/spreadsheets/d/1Va7cin600gumAUxQ-MT6InADK87iIwpqR0Qmca22dCc/edit#gid=2090244794&range=2573:2573

@deringtonc - Good question, in this case, given that the phenotype definition explicitly states that "Presence of ANY corticosteroid (no minimum dose IV, IM or PO ≥ 14 days cumulative) (drug strings)" I would argue that since this is explicitly criteria for identifying case patients that is is probably clinically relevant. @trinklek, what are your thoughts on this?

@callahantiff Ah okay, I see. I agree that they need to be clinically relevant after re-reading the cohort criteria.

I've fixed it in the spreadsheet!

[deringtonc]

@trinklek @callahantiff for this phenotype, should we marking medications that can cause steroid-induced osteonecrosis as clinically relevant (i.e., any steroid)? Or do you only want us to mark medications that can be used for treatment of this condition as clinically relevant? Example row: https://docs.google.com/spreadsheets/d/1Va7cin600gumAUxQ-MT6InADK87iIwpqR0Qmca22dCc/edit#gid=2090244794&range=2573:2573

@deringtonc - Good question, in this case, given that the phenotype definition explicitly states that "Presence of ANY corticosteroid (no minimum dose IV, IM or PO ≥ 14 days cumulative) (drug strings)" I would argue that since this is explicitly criteria for identifying case patients that is is probably clinically relevant. @trinklek, what are your thoughts on this?

[trinklek]

@trinklek @callahantiff for this phenotype, should we marking medications that can cause steroid-induced osteonecrosis as clinically relevant (i.e., any steroid)? Or do you only want us to mark medications that can be used for treatment of this condition as clinically relevant? Example row: https://docs.google.com/spreadsheets/d/1Va7cin600gumAUxQ-MT6InADK87iIwpqR0Qmca22dCc/edit#gid=2090244794&range=2573:2573

@deringtonc - Good question, in this case, given that the phenotype definition explicitly states that "Presence of ANY corticosteroid (no minimum dose IV, IM or PO ≥ 14 days cumulative) (drug strings)" I would argue that since this is explicitly criteria for identifying case patients that is is probably clinically relevant. @trinklek, what are your thoughts on this?

I completely agree! Include as clinically relevant because it identifies the phenotype. In this case, the medication you are vetting is more part of the "diagnosis" than "treatment of the diagnosis."

[deringtonc]

Just documenting this while I'm thinking about it so I don't forget. This phenotype is trying to exclude patients with HIV, and there are so many HIV medications (combinations + brand names + classes). This spreadsheet is missing many brand name products for many of these HIV drugs.

[deringtonc]

@trinklek I need your domain expert opinion. :)

We are identifying patients taking HIV medications for this phenotype in order to exclude them, because HIV antiretroviral therapy can be associated with osteonecrosis.

Some HIV medications, like tenofovir, must be administered with other agents when treating HIV because monotherapy is associated with high likelihood of viral resistance. However, some of these entries are for tenofovir-only products, which would indicate hepatitis B treatment and would not be clinically relevant to identify HIV.

However, it is possible that people on a tenofovir-only product are getting other HIV products in another pill. Therefore tenofovir-only products could be clinically relevant to identify HIV.

So my question is, for tenofovir-only entries (tenofovir alafenamide and tenofovir disoproxil fumarate), should we consider these clinically relevant or not clinically relevant?

Unfortunately, the dose for HIV treatment vs. Hepatitis B infection is the same, so I can't discern based on the dose of the product.

https://docs.google.com/spreadsheets/d/1Va7cin600gumAUxQ-MT6InADK87iIwpqR0Qmca22dCc/edit#gid=2090244794&range=37267:37267

[trinklek]

@trinklek I need your domain expert opinion. :) We are identifying patients taking HIV medications for this phenotype in order to exclude them, because HIV antiretroviral therapy can be associated with osteonecrosis. Some HIV medications, like tenofovir, must be administered with other agents when treating HIV because monotherapy is associated with high likelihood of viral resistance. However, some of these entries are for tenofovir-only products, which would indicate hepatitis B treatment and would not be clinically relevant to identify HIV. However, it is possible that people on a tenofovir-only product are getting other HIV products in another pill. Therefore tenofovir-only products could be clinically relevant to identify HIV. So my question is, for tenofovir-only entries (tenofovir alafenamide and tenofovir disoproxil fumarate), should we consider these clinically relevant or not clinically relevant? Unfortunately, the dose for HIV treatment vs. Hepatitis B infection is the same, so I can't discern based on the dose of the product. https://docs.google.com/spreadsheets/d/1Va7cin600gumAUxQ-MT6InADK87iIwpqR0Qmca22dCc/edit#gid=2090244794&range=37267:37267

@deringtonc Great question! I had to look into the available products to confirm it is possible to get the combination drugs in separate pills. I would include tenofovir-only agents as clinically relevant because they could be getting the other agent in a different pill.

[trinklek]

Just documenting this while I'm thinking about it so I don't forget. This phenotype is trying to exclude patients with HIV, and there are so many HIV medications (combinations + brand names + classes). This spreadsheet is missing many brand name products for many of these HIV drugs.

I think this is a really good thing to note, but I am wondering if it actually matters. Once we get to this step of the project, I think it will an interesting discussion.

[deringtonc]

Completed on 6/20/19