>> Applied Statistics
Assignment - Root Cause Analysis
The goal of this is to conduct probabilistic root cause analysis using methodology described in the lecture. You need to select an adverse event and construct a Bayesian network model (i.e., using Netica software) that can be used to probabilistically reason about causes that lead to the event and constraints that can be used to prevent it.
1. Interview a colleague at work (if you don't work find someone else) to analyze root causes of an adverse outcome (not necessarily a sentinel event).
a. Make sure that you list at least 3 direct causes or constraints, and some indirect causes.
b. Make sure that you include the categories suggested by JCAHO for healthcare related projects, or categories known in the literature for other projects.
c. Draw a flow chart that represents causes and effects of the adverse event.
2. Indicate what are the direct and root causes of the adverse event.
3. Give an example question that can check the conditional independence assumption associated with root causes. Make sure the question is not awkward.
4. Verify all assumptions of conditional independence in your model by interacting with your expert. Show what assumptions were checked and which assumptions were violated.
5. Estimate marginal and conditional probabilities by interviewing your expert.
6. Use software to estimate the probability of the sentinel event. Is the probability different from your original estimate based on expert interview or literature?
7. Use software to calculate the probability of sentinel event under at least 3 different scenarios (combination of causes occurring or not occurring).
a. Make sure that the scenarios make sense, i.e. do not "block" root cause by setting value of a direct cause.
8. Ask your expert if the various estimates in questions 6 through 7 are within your expert's expectations.
9. Calculate the prevalence of root causes for the sentinel event in your analysis (reverse prediction). Compare these data to JCAHO's reports on prevalence of causes of sentinel events. Report the difference between your model assumptions and the JCAHO's data. If you do not use JACHO's criteria (or your project is not healthcare related), compare to results available in the literature.
10. Discuss validity of your model.
11. Suggest how you would change the causal model to better accommodate your expert's insights. Show how your root cause analysis changed as a consequence of the data you examined.