Lecture: Stratified Covariate Balancing  

Question 3:  These data come from STAR*D experiment conducted by National Institute of Medicine.  Use instructor's last name as password.  Data► Protocol►

The data report the experience of approximately 4,000 patients with various antidepressants: citalopram, bupropion, mirzapine, buspirone, lithium, nortriptyline, sertraline, thyroid, tranylclypromine, and venlafaxine.   The following table shows the distribution of the covariates in the data.

The data are reported for a total of 22,254 visits.  Visits may be 2 week or more apart.  Not every patient shows for every scheduled visit.  Organize the data so there is one row for each patient and each antidepressant trial (known in the data as Concat). Note that this field considers combination of antidepressants as a new antidepressant. Ignore the dose of the medication. Patients received multiple antidepressants during these trials until something worked for them. Include each time a new antidepressant was tried as a separate trial. If the patient has taken the antidepressant at any time during the trial, then mark it as 1, otherwise 0. Notice that some patients have taken the medication and others have not. Patients who have not taken a particular medication have taken other medications, so at any time we are comparing one medication to alternative treatments.  The medication is considered to have caused the remission if the patient is referred to follow up portion of the study, at any point while taking the medication; i.e. the variable "Treatment_plan_equal_3" is set to 1 while taking the medication .

1. Clean the data SQL►
2. For 3 antidepressants, balance the data using SQL and stratified covariate balancing. 
3. If necessary use parents in Markov Blanket of the medication to improve overlap beyound 80%.
4. Describe which of the 3 medications should a patient who has PTSD and neurological disorders take. SQL►

Question 4:  The following data have been taken from nurses rounding in a facility.  The time they spent with patients has been recorded.  In addition, several characteristics of the patients have also been recorded and standardized.  Using stratified covariate balancing indicate if any of the nurses have a significant impact on overall satisfaction in the unit?  Please note that the listed teach one assignment uses the wrong command for stratified covariate balancing package.   Data► Polly's Teach One► New R Package► Zabowski's Answer►

Question 5:  In a nursing home, data were collected on residents' survival and disabilities.  The data are listed in the following order: ID, age, gender (M for male, F for Female), number of assessments completed on the person, number of days followed, days since first assessment, days to last assessment, unable to eat, unable to transfer, unable to groom, unable to toilet, unable to bathe, unable to walk, unable to dress, unable to bowel, unable to urine, dead (1) or alive (0), and assessment number. 

• Does inability to eat increase probability of mortality in 6 months?  Use SQL and stratified covariate balancing to determine if inability to eat contributes to mortality, after controlling for other disabilities of the patient.  Data► Adel's Teach One► Adel's SQL Code►
• Does inability to toilet contribute to survival?  To answer this question several steps must be taken.  First, later events must be ignored.  These are events that occur after the patient is unable to toilet.  Second, the stratification need only occur among parents in Markov Blanket of "Unable to toilet".  A quick way to identify the parents in the Markov Blanket of "Unable to Toilet" is to regress it on all the variables that precede it.  Then the variables that are signficant and have large effect size (procedure to simulate LASSO regression) must be stratified.  SQL and R Code►

Question 6:  The following data show the variation in diabetes in select counties across United States.  Using stratified covariate balancing report the impact of access to supermarkets on diabetes after controlling for other variables. Data► Monotone?►

1. Simplify the database.  Please organize the data so all independent  variables are measured in the next to last available year.  The outcome variable is diabetes and should be measured in the last year.  Ignore all other data, including data on independent variables in prior years.   
2. Describe the data.  Check that all independent variables are positively and monotonely related to prevalence of diabetes in the last year in the county. Split variables that are not monotonely related to prevalence of diabetes into 2 or more variables that are positively and monotonely related to diabetes
3. Calculate impact of variables on diabetes in the last year
   1. Calculate the impact of obesity in prior year on diabetes in the last year, while controlling for other variables in prior year.  Use Stratified Covariate Balancing to calculate the impact.
   2. Calculate the impact of access to food sources in prior year on diabetes in last year, while controlling for other variables.  Use Stratified Covariate Balancing to calculate the impact.