1 / 16

Utilize Dummy Datasets in Clinical Statistical Programming

Utilize Dummy Datasets in Clinical Statistical Programming. Amos Shu. Introduction. Tables such as PE, Demo, and some Efficacy, etc. need to be partially made up, e.g. data collectability Dummy datasets – an effective way to improve programming efficiency on these situations

phiala
Download Presentation

Utilize Dummy Datasets in Clinical Statistical Programming

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Utilize Dummy Datasets in Clinical Statistical Programming Amos Shu

  2. Introduction • Tables such as PE, Demo, and some Efficacy, etc. need to be partially made up, e.g. data collectability • Dummy datasets – an effective way to improve programming efficiency on these situations • Six ways to utilize dummy datasets • Make up PE tables • Make up Demographic Char. Tables • Optimize datasets in efficacy analysis • Stack datasets in macro • Impute LOCF • Add extra lines in RTF output

  3. 1. MAKE UP PE (Physical Examination) TABLES • Each organ should have all medical scenarios - ‘Normal’, ‘Abnormal’, and ‘Not Done’ displayed Placebo Treat1 Treat2 (N=45) (N=43) (N=43) ___________________________________________ Heart Normal 8 9 11 Abnormal 0 0 0 Not Done 0 0 0 Lungs/Chest Normal 7 8 10 Abnormal 0 1 0 Not Done 1 0 1 ...... ____________________________________________

  4. 1. MAKE UP PE TABLES (cont.) • Not all subjects have all findings for each organ in real practice • Create a dummy dataset: DATA dummy; DO x= 1 to 9; DO y= 1 to 3; Output; END; END; RUN; X – organ y – finding

  5. 1. MAKE UP PE TABLES (cont.)

  6. 2. MAKE UP DEMOGRAPHIC CHAR. TABLES • Most common issue - not all races are available in real data Placebo Treat1 Treat2 (N=45) (N=43) (N=43) ____________________________________________________________ Race Caucasian 28 29 27 African American 0 3 1 Asian 10 11 9 American Indian or Alaska native 0 0 0 Native Hawaiian or Pacific Islander 0 0 0 Other 1 0 1 ...... ____________________________________________________________

  7. 3. OPTIMIZE DATASETS IN EFFICACY ANALYSIS • Various subgroup efficacy analysis – handle using one macro • Geographic subgroups – US, Europe, and Asia • Sex subgroups – Male and Female • Race subgroups – Asian and Non-Asian • Smoking status subgroups – Never smoker and Current smoker • Age subgroups – Age <=65 and age >65 • Etc. • Some subgroup may not have all treatment values, e.g. • Subgroup A – treatment 1 and 2, not 3 • Subgroup B – treatment 1 and 3, not 2

  8. 3. OPTIMIZE DATASETS IN EFFICACY ANALYSIS (cont.) • Treatment dummy dataset - make the macro work DATA dummytrt; treat1 = . ; treat2 = . ; treat3 = . ; RUN; DATA SubEff_d; MERGE dummytrtSubEff; RUN; * Dummytrt must be the left

  9. 4. STACK DATASETS IN MACRO • The number of subgroup datasets are unknown • It is impossible to use a single SET statement to stack all subgroup datasets in a DO loop macro • Need a dummy dataset to avoid macro failure DATA comb ; x = 0 ; RUN;

  10. 4. STACK DATASETS IN MACRO (cont.) %DO i= 1 %TO &sqlobs.; ... DATA sub&i. ; ... ; RUN; DATA comb ; SET comb sub&i. ; RUN ; %END; DATA final; SET comb; IF x = 0 THEN DELETE ; RUN;

  11. 5. IMPUTE LOCF (Last Observation Carried Forward) • Two types of LOCF: • No data for a visit number • Both the visit number and data are missing

  12. 5. IMPUTE LOCF (Last Observation Carried Forward) (cont.) DATA dummy ; SET lockbase; DO AVISITN = 9, 10, 11 ; output ; END; RUN; DATA lockdata2 ; MERGE dummylockdata ; BY USUBJID AVISITN ; RUN;

  13. 6. ADD EXTRA LINES IN RTF OUTPUT • Improve the appearance of a table

  14. 6. ADD EXTRA LINES IN RTF OUTPUT (cont.) Placebo Treat1 Treat2 (N=45) (N=43) (N=43) ________________________________________________________ Week 1 N 103 103 101 Mean 36.2 37.4 36.6 SD 10.25 10.43 9.77 Median 36.0 37.0 36.0 Minimum 2 11 6 Maximum 59 64 60 Week 2 N 103 102 101 Mean 35.9 37.1 36.3 SD 10.15 10.03 10.07 Median 36.0 37.0 36.0 Minimum 3 10 6 Maximum 58 64 59 ...... __________________________________________________________

  15. CONCLUSION • Creating a dummy dataset is easy and simple • It’s extremely helpful to programming • It’s crucial to understand when and how to create a dummy dataset

  16. Thank you! Amos Shu Endo Pharmaceuticals

More Related