This document discusses causal inference methods for uncertain time series data. It presents a method for analyzing time-stamped medical sensor data to determine causal relationships between variables like glucose levels, exercise, sleep, and insulin dosage. The method is applied to data from 17 subjects with Type 1 diabetes over 72 hours. The results suggest that very vigorous exercise leads to short-term hyperglycemia, which is supported by literature. Open problems discussed include handling latent variables, non-stationary data, real-time prediction and explanation, and simulation.
10. Logic-based causal inference
• Complex, temporal relationships
• Potential causes raise probability of and are
earlier than effects
Kleinberg, S. (2012) Causality, Probability, and Time.
11. Which causes are significant?
• Main idea: looking for better explanations for
the effect
• Assess average difference cause makes to
probability of effect
• Determine which ε are statistically significant
12. Causes from EHR data
8
9
17
17
18
11
10
22
22
21
0510152025
dx_urinary_symptoms
first_antihtn_combo
dx_overweight
dx_diabetes
dx_hypothyroidism
Months before CHF diagnosis
S. Kleinberg and N. Elhadad (2013) Lessons Learned in Replicating Data-Driven Experiments in
Multiple Medical Systems and Patient Populations. AMIA Annual Symposium.
20. Causes of changes in glucose
Cohort: 17 subjects with T1DM
Sensor data (collected for >72 hours)
– Glucose values
– Insulin dosage
– Activity
– Sleep stage
– Heart rate
– Temperature
With N. Heintzman (UCSD,Dexcom) http://dial.ucsd.edu/what-we-do.php
21. Results
very vigorous exercise leads to hyperglycemia
(fdr <.01) in 5-30 minutes
– Found using both HR (anaerobic activity zone) and
METs
• Not found with strict discretization
• Supported by literature
(Marliss and Vranic, 2002; Riddell and
Perkins, 2006)
S. Kleinberg and N. Heintzman. Inference of Causal Relationships from Uncertain Data and
Application to Type 1 Diabetes. (under review)
22. Open problems
• Latent variables
• Nonstationary time series
• Real-time prediction/explanation
• Simulation