Phenotype Submission - Symptoms and signs

1

The OHDSI Phenotype library has 29 cohort definitions that are either symptoms or signs – all in pending peer review status. These cohort definitions need to go through the phenotype submission and evaluation (peer review) cycle.

Because these are presentations of disease, and not a disease by itself itself - I think, they don’t easily fit the process we have been following recently. So, i think a single post like this may be sufficient to discuss these phenotypes.

I will try to provide a commentary on what I see in Cohort diagnostics. It would be great if other clinicians in the forum have a look at them and provide a commentary of their opinions.

Output of cohort diagnostics is available at data.ohdsi.org/PhenotypeLibrary and these cohort definitions below are at atlas-phenotype.ohdsi.org

These cohort definitions are useful for many reasons:

  1. They may be used as feature cohorts. e.g. when characterizing a disease such as Myocardial Infarction - it is more informative to use a cohort definition for nausea or vomiting, chest pain - instead of individual codes. Feature/covariate cohorts are handled by OHDSI software like https://github.com/OHDSI/FeatureExtraction/pull/167 it may also be used in propensity score calculation.

  2. Outcome cohorts: it may be used in comparative effectiveness or safety studies were the outcome maybe such signs or symptoms.

Id Name
3 Cough or Sputum
4 Diarrhea
5 Dyspnea
6 Fever
7 Headache or Headache disorder
8 Anosmia OR Hyposmia OR Dysgeusia
9 Sore throat
10 Nausea or Vomiting
11 Malaise or fatigue
12 Rhinitis or common cold
13 Myalgia (not explained by injury, ischemia or systemic inflammation)
14 Myalgia
20 Bronchitis
32 Obesity
54 Febrile seizure
57 Bleeding
62 Generalized Seizure
64 Flu-like symptoms fever, cough, malaise, fatigue, dyspnea, myalgia
95 Delirium
189 Right Upper Quadrant Pain
190 Abdominal Distension
191 Fatigue, Malaise, Lethargy, Anorexia
193 Jaundice or Itching
194 Encephalopathy or its presentations
232 Paresthesia
241 Urticaria
243 Tinnitus
244 Dizziness
2

Thanks @Gowtham_Rao for contributing these phenotypes to the OHDSI Phenotype Library. I am re-using the Obesity phenotype for purposes of creating a nesting indication for GLP1 inhibitors.

A methodological note: the definition for obesity uses all events (conditions, observations, measurements) with a cohort era gap of 0 days, meaning that a person can have multiple cohort episodes. Obesity seems like a tricky cohort in this regard in that its generally a chronic disease and often can be presumed to persist in absence of evidence to refute it. At the same time, it is possible that a person could satisfy ‘obesity’ requirements, then ‘exit the cohort’ through weight loss via a multitude of interventions, but then ‘re-enter the cohort’ if they experience weight gain. Clearly there will be index date misspecification involved in both the obesity start and end dates. As I plan to use it, I think it might be useful to consider some compromise approach whereby I will presume that obesity is ‘chronic’ for some duration (which could be implemented by using a fixed duration offset and by using the era collapse gap size), so this would allow for event recurrence but would also stich together periods that look like continuous period of obesity. For nesting indication, this will manifest as looking back from the index date and requiring observing some obesity event within the 365d prior.