Additional assignment on lifetables February 25, 1999
(To be completed with Question 3 of Problem Set 3)
1. Attached to an e-mail sent to you is a a STATA data file: lifetable2.dta
If you wish, you may drop by my office and I will provide it on a 3.5 disk.
2. Produce a lifetable from these data that includes both the cumulative survival (or failure) estimates and the hazard rates. You cannot do this with STATA. You'll have to pull some numbers from STATA and construct a with a lifetable. with a spreadsheet or copy down some numbers.
3. Describe what the lifetable (produced in 2) shows. Calculate useful statistics for describing how long the review process takes. Produce a relevant figure (if useful).
4. Suggest additional data that could be added to the data set that would help describe/understand the process.
5. Are we getting any better (i.e. faster)? Draw a lexis diagram that indicates an appropriate test. Carry out the test if you think it useful. If not useful, tell me why.
--------------------------
STATA do file .... Already incorporated in the file you have: lifetable2.dta
/* file "pmrec.do" create lifetable vars */
#delimit;
/* comment: recodes below assign current date for censored obs */
recode dy .=99;
recode dm .=2;
recode dd .=20;
/* calcualte days of exposure... let 1/1/95= day 1: exp=decdt-recdt */
gen recdt=((ry-95)*365)+(rm*30)+(rd);
gen decdt=((dy-95)*365)+(dm*30)+(dd);
gen exp=decdt-recdt;
/* divide days of exposure into months */
gen recexp=int(exp/30.5);
lab var recdt "day of receipt 1/1/95=1";
lab var decdt "day of decision, 1/1/95=1";
lab var exp "days under review";
lab var recexp "months under review";
lab var sub "submission new r1 r2";
/* recode "no decision" to 0 to indicate censoring */
recode decision .=0;
do file that runs a lifetable (cumulative survival) and produces a plot of cumulative survival
/* file "pmlt.do" calculate lifetable and plot survival function */
#delimit;
ltable recexp decision;
ltable recexp decision, graph notab xlab (0,2,4,6,8,10);
exit;
Contains data from lifetable2.dta
obs: 137
vars: 12 24 Feb 1999 16:46
size: 7,124 (99.2% of memory free)
-------------------------------------------------------------------------------
1. sub float %9.0g submission new r1 r2
2. decision float %9.0g
3. rm float %9.0g
4. rd float %9.0g
5. ry float %9.0g
6. dm float %9.0g
7. dd float %9.0g
8. dy float %9.0g
9. recdt float %9.0g day of receipt 1/1/95=1
10. decdt float %9.0g day of decision, 1/1/95=1
11. exp float %9.0g days under review
12. recexp float %9.0g months under review
-------------------------------------------------------------------------------
. ltable recexp decision;
Beg. Std.
Interval Total Deaths Lost Survival Error [95% Conf. Int.]
-------------------------------------------------------------------------------
0 1 137 3 6 0.9776 0.0128 0.9322 0.9927
1 2 128 5 6 0.9385 0.0211 0.8808 0.9688
2 3 117 8 17 0.8693 0.0306 0.7952 0.9180
3 4 92 12 6 0.7521 0.0411 0.6604 0.8224
4 5 74 17 8 0.5694 0.0495 0.4666 0.6597
5 6 49 14 8 0.3923 0.0521 0.2907 0.4923
6 7 27 9 5 0.2482 0.0504 0.1565 0.3509
7 8 13 1 6 0.2234 0.0511 0.1324 0.3292
8 9 6 1 2 0.1787 0.0572 0.0838 0.3023
9 10 3 1 1 0.1072 0.0651 0.0222 0.2698
10 11 1 1 0 0.0000 . . .
-------------------------------------------------------------------------------
. ltable recexp decision, hazard
Beg. Cum. Std. Std.
Interval Total Failure Error Hazard Error [95% Conf. Int.]
-------------------------------------------------------------------------------
0 1 137 0.0224 0.0128 0.0226 0.0131 0.0000 0.0483
1 2 128 0.0615 0.0211 0.0408 0.0182 0.0050 0.0766
2 3 117 0.1307 0.0306 0.0766 0.0270 0.0235 0.1296
3 4 92 0.2479 0.0411 0.1446 0.0416 0.0630 0.2262
4 5 74 0.4306 0.0495 0.2764 0.0664 0.1463 0.4066
5 6 49 0.6077 0.0521 0.3684 0.0968 0.1787 0.5581
6 7 27 0.7518 0.0504 0.4500 0.1462 0.1635 0.7365
7 8 13 0.7766 0.0511 0.1053 0.1051 0.0000 0.3113
8 9 6 0.8213 0.0572 0.2222 0.2208 0.0000 0.6551
9 10 3 0.8928 0.0651 0.5000 0.4841 0.0000 1.4489
10 11 1 1.0000 . 2.0000 0.0000 2.0000 2.0000
-------------------------------------------------------------------------------