TODO - Manuscript Task: Verify Sleep Apnea Phenotype

[callahantiff]

Creating as a separate issue to break-up issue #2

Finalizing the Sleep Apnea Phenotype

Case Criteria:

• Presence of specific sleep apnea codes (ICD9 code)
• Presence of >1 procedures (CPT or HCPCS code)

Control Criteria:

• No control group defined

---

Questions:

• No control group defined, should we try to create one?

• The phenotype definition states that, "The computable phenotype can be cross-referenced against the following to increase specificity: AHI (sleep lab report) >= 5 or HCPCS code E0601 (CPAP) or E0470 or E0471 (BIPAP) or A7034 (CPAP mask) or HCPCS A9270 or E)486 (Mandibular devices) or CPT 95811 (CPAP titration) or CPT 95811 (Split night CPAP titration)." I think we should include these in the sense that if a patient meets the diagnosis code criteria, they should also have had one of these procedures. What do you think?

---

@tdbennett, can you please advise on these questions?

[tdbennett]

1. No new control groups. Perfect is the enemy of done.

2. Because the phenotype is vague ("can be cross-referenced...") let's use the phenotype as is and then plan on a sensitivity analysis looking at requiring the HCPCS codes with the goal of improving the phenotype for subsequent users.

[callahantiff]

Great, works for me! I will make a note in the Google Doc.

[callahantiff]

We have addressed all open questions; closing this issue for now.