This document discusses analyzing patient data to predict mortality risk and describes the steps taken:
1) A training and validation set were created from over 17 million patient cases to calculate likelihood ratios and predict 6-month mortality.
2) The most and least deadly diseases were identified based on their likelihood ratios, with things like intracranial hemorrhage being very deadly and disorders like uterine myomas being less so.
3) Sensitivity and specificity of the predictions were calculated to determine the usefulness of the model.
2. The objective of the work
Analysis to Predict Who Will Die When.
HOW ?
Create Training and Validation set .
Use the training set to calculate likelihood ratio.
It’s important because it gives forecast information regarding health outcomes.
this assignment teach us to explore data and locate exact information among big
data.
3. Data source
Number of cases
What is the distribution of the data
• Data source from the Assignment
Select count (*) from dbo.final
• The total number of cases
( 17,443,442 number of cases)
• distribution of the data
The average is -59.5318
And the Standard deviation- 4.2931
Average AgeAtDx: 59.53186
Standard Deviation of AgeAtDx: 4.293136
Start Dataset (hap464.dbo.final): 17,443,442 Cases
and 829,827 IDs
Zombies Removed: 17,432,694 Cases and 829,659
IDs
>365 Dx/Yr Removed: 17,379,218 Cases and
829,603 IDs This is your clean data.
80% Training Set From Clean Data: 13,760,416 Cases
and 657,905 IDs
20% Validation Set From Clean Data: 3,619,297
Cases and 171,698 IDs
4. Preparation of the data
17,443,442
10,748
diagnoses
removed
53,476
diagnoses
removed
829,827 distinct
IDs
Remove
Zombies: 168
distinct IDs
5. Calculating Likelihood Ratios
(The number of patient who will died within 6 months /Dead Patients)
(The number of patient who will died within 6 months/ Alive Patients)
7. The name of Least deadly icd9 diagnose
• 1 I218.9 : eiomyoma of uterus,
• 2 I626.2 : Disorders of menstruation
• 3 I478.0 : Hypertophy of nasal turbnates
• 4 I599.7 : Hekaturia
• 5 I620.2 : Other and unspecified ovarian cyst
• 6 I717.83 : Old disruption of anterior cruciate ligament
• 7 I474.00 : Effusion, right root
• 8 I296.42 : Bipolar i disorder , most recent episode
• 9 I716.17 : Traumatic arthropathy, ankle and foot
• 10 IV57.22 : Operation On urinary Bladeder
8. The name of Most deadly icd9 diagnose
• 1 I853.05 : Unspecified Intracranial hemorrhage following injury without mention of open interacranian wound
• 2 I798.2 : Death occurring in less than 24 hours from onest of symptoms
• 3 I183.2 : Malingant neoplasm of fallopain tube
• 4 I798.9 : Unspecified interacranial homerrhage following
• 5 I194.8 : Malignant neoplasm of other endocrine gland and related Structure
• 6 I960.7 : posing by antineoplastic antibiotics
• 7 I862.21 : injury to Bronchus without mention of open wound into cavity
• 8 I852.05 : subarachniod hermorrhage folowing injury without mention of open intracranial wound
• 9 I718.59 : Ankylosis of joint ,Multiples sites
• 10 IV57.21 :Acute gastric Ulcer with hemorrhage and perforation with obstruction
9. Calculate sensitivity and specificity of the
predictions
Posterior Odds
Alive Dead
True
Condition
Alive True Positive False Negative
Dead False Positive True Negative
10. Usefulness of the project
• The usefulness of the project is to practice doing SQL in a large data set by using the skills of
codes, Also to figure out Selecting appropriate method of data analysis and removal of
confounding in the data, Visually present complex multivariate data and Interpret
quantitative findings and relate it to specific policy issues or management decisions.
• In fact, It’s important in our future work filed