Lab 03 Codebook: ED Visits

ED Visits Probability Lab Dataset

This is a simplified teaching version of the prepared NHAMCS emergency department data.

Each row is one emergency department visit record. The row includes patient characteristics for that visit, but the file should not be treated as a longitudinal patient file with unique patients tracked over time.

This dataset is for probability and data-science teaching. It is not a clinical research dataset, and disease_group is a simplified teaching label rather than an ICD diagnosis.

Files

Core Columns

column meaning
age_years patient age in years
age_group simplified age group: 0-17, 18-34, 35-49, 50-64, or 65+
sex recorded patient sex
payer primary expected source of payment
arrival_by_ambulance whether the patient arrived by ambulance
triage_level simplified triage category
wait_time_minutes recorded waiting time before being seen; some values are missing
reason_for_visit_count number of reasons for visit recorded
chronic_condition_count number of listed chronic conditions; some values are missing
disease_group broad teaching label for the visit diagnosis

Symptom Columns

The symptom columns are 0/1 indicators. A value of 1 means that reason for visit was recorded somewhere among the visit’s reason-for-visit fields.

column meaning
abdominal_pain abdominal pain, cramps, or spasms
cough cough
vomiting vomiting
fever fever
nausea nausea
chest_pain chest pain
shortness_of_breath shortness of breath
headache headache or pain in head
accident accident, not otherwise specified
dizziness vertigo or dizziness
back_pain back pain, ache, soreness, or discomfort
anxiety anxiety or nervousness
skin_rash skin rash

Suggested Probability Events

These columns make it easy to define events:

cough = df["cough"] == 1
respiratory = df["disease_group"] == "respiratory"
urgent = df["triage_level"] == "Urgent"
ambulance = df["arrival_by_ambulance"] == "Yes"

Then empirical probabilities are just proportions:

cough.mean()
(cough & respiratory).mean()
respiratory[cough].mean()

For this lab, interpret probabilities as proportions in this prepared dataset, not as national estimates.