Examples: Tables. Descriptive stats with ANOVA, ANCOVA, and t test p values Frequency and percent Correlation coefficients

Similar documents
Frequency Tables Aspects to Consider / PhUSE - Industry Starters / Katja Glaß

CDISC Journey in Solid Tumor using RECIST 1.1. Kevin Lee Statistician/CDISC Consultant/Programmer

Standard Methods for Analysis and Reporting of VAS or NRS Derived Pain Relief Response Scores

ADaM Grouping: Groups, Categories, and Criteria. Which Way Should I Go?

Michigan Neuropathy Screening Instrument (MNSI)

Hospital Anxiety and Depression Scale (HADS)

Clinical Global Impression (CGI)

Epworth Sleepiness Scale (ESS)

Neuropathic Pain Scale (NPS)

Columbia-Suicide Severity Rating Scale Baseline (C-SSRS BASELINE)

Let s Create Standard Value Level Metadata

Modified Hachinski Ischemic Scale: NACC Version (MHIS-NACC) v.1.0

Graph Display of Patient Profile Using SAS

Barnes Akathisia Rating Scale (BARS)

Brief Psychiatric Rating Scale-Anchored (BPRS-A)

Columbia-Suicide Severity Rating Scale Baseline (C-SSRS BASELINE)

Disability Assessment for Dementia (DAD)

Short-Form McGill Pain Questionniare-2 (SHORT-FORM MPQ-2)

Columbia-Suicide Severity Rating Scale Baseline (C-SSRS BASELINE)

Brief Psychiatric Rating Scale-Anchored (BPRS-A)

Pain Relief (PR) Notes to Readers

ADaM Tips for Exposure Summary

PROGRAMMER S SAFETY KIT: Important Points to Remember While Programming or Validating Safety Tables

CDISC Journey in Solid Tumor using RECIST 1.1 Kevin Lee PhUSE conference Oct 14th, 2013

Hamilton Depression Rating Scale 17-Item (HAMD 17)

Analysis Datasets for Baseline Characteristics and Exposure Time for Multi-stage Clinical Trials

Integrated ADSL. PhUSE Nate Freimark/Thomas Clinch Theorem Clinical Research

AE: An Essential Part of Safety Summary Table Creation. Rucha Landge, Inventiv International Pharma Services Pvt. Ltd.

A SAS sy Study of ediary Data

Statistical Analysis Plan RH01649 Version Document Identifier Effective Date eldo_clinical_doc Reason For Issue

Immediate-release Hydrocodone/Acetaminophen M Abbreviated Clinical Study Report R&D/08/1020

Clinical Trial Synopsis TL-OPI-518, NCT#

Imputing Dose Levels for Adverse Events

SYNOPSIS. ER OROS Paliperidone: Clinical Study Report R SCH-301

PFIZER INC. THERAPEUTIC AREA AND FDA APPROVED INDICATIONS: See USPI.

Summary ID# Clinical Study Summary: Study B4Z-MC-LYCL

Design and Construct Efficacy Analysis Datasets in Late Phase Oncology Studies

SYNOPSIS 2/198 CSR_BDY-EFC5825-EN-E02. Name of company: TABULAR FORMAT (For National Authority Use only)

Summary ID#4668 Clinical Study Summary: Study B4Z-MC-LYAQ

Pharmaceutical Applications

OCCDS Creating flags or records. Rob Wartenhorst PhUSE Barcelona 2016

Stat Wk 9: Hypothesis Tests and Analysis

EVERY DAY A GUIDE TO KNOW YOUR NUMBERS

ADaM Considerations and Content in the TA User Guides. Susan J Kenny, PhD Maximum Likelihood, Inc.

Linear Regression in SAS

Summary ID# Clinical Study Summary: Study B4Z-MC-LYAX

Support for Acetaminophen 1000 mg Over-the-Counter Dose:

Information Sheet. Weight. Accessible information about weight for adults with an eating disorder

Individual Study Table Referring to Item of the Submission: Volume: Page:

Using Direct Standardization SAS Macro for a Valid Comparison in Observational Studies

A study of academic stress and its effect on vital parameters in final year medical students at SAIMS Medical College, Indore, Madhya Pradesh

The clinical trial information provided in this public disclosure synopsis is supplied for informational purposes only.

SYNOPSIS. Publications No publications at the time of writing this report.

SYNOPSIS (FOR NATIONAL AUTHORITY USE ONLY) INDIVIDUAL STUDY TABLE REFERRING TO PART OF THE DOSSIER

SUPPLEMENTARY MATERIAL

Summary ID# Clinical Study Summary: Study B4Z-JE-LYBC

Summary ID# Clinical Study Summary: Study B4Z-MC-LYBR

PATIENT INFORMATION. Medicine To Treat: H ypertension. Patient Awareness and Self-Care

NOT-FED Study New Obesity Treatment- Fasting, Exercise, Diet

Sample Resting Metabolic Rate Test Results

2.0 Synopsis. ABT-711 M Clinical Study Report R&D/06/573. (For National Authority Use Only) to Part of Dossier: Volume:

Hungry Mice. NP: Mice in this group ate as much as they pleased of a non-purified, standard diet for laboratory mice.

Digestive Project Part 1

SUPPLEMENTARY DATA. Supplementary Methods

Adalimumab M Clinical Study Report Final R&D/16/0603

Implementing Patient Report Outcome Data in Clinical Trials Analysis

Paliperidone: Clinical Protocol R076477SCH4012, CR Amendment INT-1

Intro to SPSS. Using SPSS through WebFAS

SYNOPSIS. Study center(s) This study was conducted in the United States (128 centers).

Summary ID# Clinical Study Summary: Study B4Z-MC-LYBU

RICHMOND PARK SCHOOL LIFESTYLE SCREENING REPORT Carmarthenshire County Council

Allergan Not Applicable AGN A Multi-Center, Double-Blind, Randomized, Placebo-Controlled, Multiple Dose, Parallel

a) Vitality Compass Life Expectancy

Chapter 6 Measures of Bivariate Association 1

Chapter 9 Vital Signs and SAMPLE History DOT Directory

HIGH ENERGY AND FAT INTAKE BETWEEN DORMITORY AND NON-DORMITORY STUDENTS IN CHULALONGKORN UNIVERSITY, THAILAND

Clinical Trial Study Synopsis

BRL /RSD-101C0F/1/CPMS-716. Report Synopsis

Supplementary Online Content

PharmaSUG 2018 Paper AD-02

Study Center(s): The study was conducted at 39 study sites in Japan.

2.0 Synopsis. ABT-358 M Clinical Study Report R&D/06/099. (For National Authority Use Only) to Item of the Submission: Volume:

Summary ID#7029. Clinical Study Summary: Study F1D-MC-HGKQ

Supplementary Appendix

2.0 Synopsis. ABT-333 M Clinical Study Report R&D/09/956

Lindsey Dorflinger, Ph.D. VA Connecticut Healthcare System Yale School of Medicine

Unified Parkinson's Disease Rating Scale UPDRS

Standards for Analyzable Submission Datasets for Directed Safety Analyses

Paper number: CC02 MACRO %NEWFLOW JIAN HUA (DANIEL) HUANG, FOREST LABORATORIES INC, NJ

Who is Exante Diet suitable for?

INDIVIDUAL STUDY TABLE REFERRING TO PART OF THE DOSSIER Volume: Page:

Home Food Resources for You Consumers. Food

UNIVERSIDAD*DE*COSTA*RICA*

BCH 445 Biochemistry of nutrition Dr. Mohamed Saad Daoud

Clinical Study Report. Arbonne Essentials Products Arbonne Evolution Products

Things you need to know about the Normal Distribution. How to use your statistical calculator to calculate The mean The SD of a set of data points.

Table S2: Anthropometric, clinical, cardiovascular and appetite outcome changes over 8 weeks (baseline-week 8) by snack group

Analysis of Oncology Studies for Programmers and Statisticians

P105 STUDY REPORT A Double Blind Placebo Controlled Trial of the Effect of AppTrim

PharmaSUG Paper SP01

Transcription:

Examples: Tables Descriptive stats with ANOVA, ANCOVA, and t test p values Frequency and percent Correlation coefficients

Descriptive Statistics Table 1. Body Mass Index using Last Observation Carried Forward Visit Baseline Week 6 Statistics N Mean Median (etc.) N Mean Median (etc.) p value (3) Active (N=xx) xx xx.x xx xx.x Observed Placebo (N=xx) xx xx.x xx xx.x Active (N=xx) xx xx.x x.xxxx Change from Baseline Placebo (N=xx) xx xx.x x.xxxx (1) ANOVA model: response = Treatment Group + Study Center. (2) ANCOVA model: change = Baseline + Treatment Group + Study Center. (3) Paired t test for change from baseline within Treatment Group. Treatment Group Comparison (p value) x.xxxx (1) x.xxxx (2)

Incidence Table 2. Incidence of Body Weight Loss, Baseline to Week 12 Visit / Outcome Statistics Active (N=xx) Placebo (N=xx) Week 12 Body Weight Loss No Body Weight Loss N (pct) xx (xxx.x) xx (xxx.x)

Correlations Table 3. Correlation Coefficients for Weight/BMI with Dietary Intake Change from Baseline to End of Study Change from Baseline to End of Study Active Correlation N Placebo Correlation N Weight (kg) Total calories (kcal) Carbohydrate (g) (etc.) Body Mass Index Total calories (kcal) Carbohydrate (g) (etc.) Body Mass Index % of Calories from Carbohydrate (etc.)

SDTM Files Vital Signs: Height and Weight (VS) Food Frequency Questionnaire: dietary intake and analysis (QS)

SDTM Files: VS USUBJID VISIT VSTESTCD VSSTRESN VSSTAT 01 01 BASELINE HEIGHT xx 01 01 BASELINE WEIGHT xx 01 01 WEEK 6 HEIGHT xx 01 01 WEEK 6 WEIGHT xx 01 01 WEEK 12 WEIGHT xx 01 01 WEEK 12 HEIGHT. NOT DONE

SDTM Files: QS QSCAT = [name of questionnaire] QSSCAT QSTESTCD QSTEST QSORRES QSSTRESN VISIT Raw Questionnaire Response COLDCFR Cold cereal, Freq last wk 3 4 DAYS LAST WEEK 4 BASELINE Raw Questionnaire Response COLDCQU Cold cereal, Quan each 1 BOWL 2 BASELINE Dietary Analysis Dietary Analysis TOTCAL Total calories 839.52 839.52 BASELINE CARB Carbohydrate 105.00 105.00 BASELINE

Plan for ADaM File ADDIET Include nonmissing VSSTRESN, QSSTRESN Calculate BMI = [weight (kg)] / [height (m)] 2 Calculate percent of total calories from carbohydrate, fat, protein

Compute BMI AVISIT PARAMCD PARAMTYP SRCDOM SRCVAR SRCSEQ Baseline HEIGHT VS VSSTRESN xx Baseline WEIGHT VS VSSTRESN xx Baseline BMI DERIVED Week 6 HEIGHT VS VSSTRESN xx Week 6 WEIGHT VS VSSTRESN xx Week 6 BMI DERIVED

Compute PCTCARB AVISIT PARAMCD PARAMTYP SRCDOM SRCVAR SRCSEQ Baseline CARB QS QSSTRESN xx Baseline TOTCAL QS QSSTRESN xx Baseline PCTCARB DERIVED Week 6 CARB QS QSSTRESN xx Week 6 TOTCAL QS VSSTRESN xx Week 6 PCTCARB DERIVED

Parameters PARAM PARAMCD PARAMTYP Body Mass Index BMI DERIVED Carbohydrate (grams) CARB Height (cm) HEIGHT % of Calories from Carbohydrate PCTCARB DERIVED % of Calories from Fat PCTFAT DERIVED % of Calories from Protein PCTPROT DERIVED Total calories (kcal) TOTCAL Weight (kg) WEIGHT

Baseline flag AVISIT PARAMCD PARAMTYP ABLFL Baseline BMI DERIVED Y Baseline CARB Y Baseline HEIGHT Y Baseline PCTCARB DERIVED Y Baseline TOTCAL Y Baseline WEIGHT Y

Compute LOCF BMI AVISIT VISIT PARAMCD PARAMTYP DTYPE SRCDOM SRCVAR SRCSEQ Baseline BASELINE BMI DERIVED Week 6 WEEK 6 BMI DERIVED Week 12 WEEK 6 BMI DERIVED LOCF

Average Over Visits AVISIT PARAMCD PARAMTYP DTYPE SRCDOM SRCVAR SRCSEQ Post Baseline CARB AVERAGE Post Baseline PCTCARB DERIVED AVERAGE Post Baseline TOTCAL AVERAGE

Dietary Parameters AVISIT PARAMCD PARAMTYP DTYPE SRCDOM SRCVAR SRCSEQ Baseline CARB QS QSSTRESN xx Baseline TOTCAL QS QSSTRESN xx Baseline PCTCARB DERIVED Week 6 CARB QS QSSTRESN xx Week 6 TOTCAL QS QSSTRESN xx Week 6 PCTCARB DERIVED Week 12 CARB QS QSSTRESN xx Week 12 TOTCAL QS QSSTRESN xx Week 12 PCTCARB DERIVED Post Baseline CARB AVERAGE Post Baseline PCTCARB DERIVED AVERAGE Post Baseline TOTCAL AVERAGE

AVAL, BASE, CHG AVISIT PARAMCD AVAL BASE CHG ABLFL Baseline BMI x.x x.x Y Baseline HEIGHT x.x x.x Y Baseline CARB x.x x.x Y Baseline TOTCAL x.x x.x Y Baseline PCTCARB x.x x.x Y Baseline WEIGHT x.x x.x Y Week 6 BMI x.x x.x x.x Week 6 HEIGHT x.x x.x x.x Week 6 CARB x.x x.x x.x Week 6 TOTCAL x.x x.x x.x Week 6 PCTCARB x.x x.x x.x Week 6 WEIGHT x.x x.x x.x

Descriptive Statistics Table 1. Body Mass Index using Last Observation Carried Forward Visit Baseline Week 6 Statistics N Mean Median (etc.) N Mean Median (etc.) p value (3) Active (N=xx) xx xx.x xx xx.x Observed Placebo (N=xx) xx xx.x xx xx.x Active (N=xx) xx xx.x x.xxxx Change from Baseline Placebo (N=xx) xx xx.x x.xxxx (1) ANOVA model: response = Treatment Group + Study Center. (2) ANCOVA model: change = Baseline + Treatment Group + Study Center. (3) Paired t test for change from baseline within Treatment Group. Treatment Group Comparison (p value) x.xxxx (1) x.xxxx (2)

Results Metadata: Table 1 Display Identifier: Table 1 Display Name: Table 1. Body Mass Index using Last Observation Carried Forward Result Identifier Param Paramcd Analysis Variable Dataset Selection Criteria Documentation Programming Statements (SAS ) Observed baseline Treatment Group Comparison p value Body Mass Index BMI AVAL ADDIET AVISITN = 0 and ITTFL = Y Use Type III sums of squares for all models, and pooled site number proc glm; class trtpn sitegr1n; model aval = trtpn sitegr1n; Observed postbaseline LS means Body Mass Index BMI AVAL ADDIET AVISITN > 0 and ITTFL = Y proc glm; by avisitn; class trtpn sitegr1n; model aval = base trtpn sitegr1n; lsmeans trtpn;

Results Metadata: Table 1 Display Identifier: Table 1 Display Name: Table 1. Body Mass Index using Last Observation Carried Forward Result Identifier Param Paramcd Analysis Variable Dataset Selection Criteria Document ation Programming Statements (SAS ) Change LS means and Treatment Group Comparison p value (footnote 2) Body Mass Index BMI CHG ADDIET AVISITN > 0 and ITTFL = Y Model with no interaction terms proc glm; by avisitn; class trtpn sitegr1n; model chg = base trtpn sitegr1n; lsmeans trtpn; p value (footnote 3) Body Mass Index BMI CHG ADDIET AVISITN > 0 and ITTFL = Y Paired t test proc means prt nway; class avisitn trtpn; var chg;

Categorize Weight Change AVISIT PARAMCD AVAL BASE CHG CHGCAT1 Baseline WEIGHT x.x x.x Week 6 WEIGHT x.x x.x 1.3 Body Weight Loss Week 12 WEIGHT x.x x.x 0.9 No Body Weight Loss

Incidence Table 2. Incidence of Body Weight Loss, Baseline to Week 12 Visit / Outcome Statistics Active (N=xx) Placebo (N=xx) Week 12 Body Weight Loss No Body Weight Loss N (pct) xx (xxx.x) xx (xxx.x)

Results Metadata: Table 2 Display Identifier: Table 2 Display Name: Incidence of Body Weight Loss, Baseline to Week 12 Result Identifier Param Paramcd Analysis Variable Weight (kg) Dataset Selection Criteria Documentation Programming Statements (SAS ) WEIGHT CHGCAT1 ADDIET ITTFL = Y and CHGCAT1N in (1,2) and AVISIT = Week 12 CHGCAT1N = 1 means body weight loss, 2 = no body weight loss.

Correlations Table 3. Correlation Coefficients for Weight/BMI with Dietary Intake Change from Baseline to End of Study Change from Baseline to End of Study Active Correlation N Placebo Correlation N Weight (kg) Total calories (kcal) Carbohydrate (g) (etc.) Body Mass Index Total calories (kcal) Carbohydrate (g) (etc.) Body Mass Index % of Calories from Carbohydrate (etc.)

Correlations: tasks Only visits where all parameters were measured (no LOCF) Change from baseline to last such postbaseline visit

All values measured Flag visits where all parameters measured Easy when denormalized: if nmiss(height,weight,carb,totcal) = 0 then ANL01FL = Y ; AVISIT HEIGHT WEIGHT TOTCAL CARB ANL01FL Baseline x.x x.x x.x x.x Y Week 6 x.x x.x x.x x.x Y

All values measured Then transpose: AVISIT PARAMCD CHG ABLFL ANL01FL Baseline CARB Y Y Baseline HEIGHT Y Y Baseline TOTCAL Y Y Baseline WEIGHT Y Y Week 6 CARB x.x Y Week 6 HEIGHT x.x Y Week 6 TOTCAL x.x Y Week 6 WEIGHT x.x Y

Last visit with all measured AVISIT PARAMCD CHG ABLFL ANL01FL ANL02FL Baseline CARB Y Y Baseline HEIGHT Y Y Baseline TOTCAL Y Y Baseline WEIGHT Y Y Week 6 CARB x.x Y Week 6 HEIGHT x.x Y Week 6 TOTCAL x.x Y Week 6 WEIGHT x.x Y Week 12 CARB x.x Y Y Week 12 HEIGHT x.x Y Y Week 12 TOTCAL x.x Y Y Week 12 WEIGHT x.x Y Y

All measured at baseline AVISIT PARAMCD CHG ABLFL ANL01FL ANL02FL ANL03FL Baseline CARB Y Y Y Baseline HEIGHT Y Y Y Baseline TOTCAL Y Y Y Baseline WEIGHT Y Y Y Week 6 CARB x.x Y Y Week 6 HEIGHT x.x Y Y Week 6 TOTCAL x.x Y Y Week 6 WEIGHT x.x Y Y Week 12 CARB x.x Y Y Y Week 12 HEIGHT x.x Y Y Y Week 12 TOTCAL x.x Y Y Y Week 12 WEIGHT x.x Y Y Y

All not measured at baseline AVISIT PARAMCD CHG ABLFL ANL01FL ANL02FL ANL03FL Baseline CARB Y Baseline TOTCAL Y Baseline WEIGHT Y Week 6 CARB x.x Y Week 6 HEIGHT x.x Y Week 6 TOTCAL x.x Y Week 6 WEIGHT x.x Y Week 12 CARB x.x Y Y Week 12 HEIGHT x.x Y Y Week 12 TOTCAL x.x Y Y Week 12 WEIGHT x.x Y Y

Record Selection ANL02FL = Y AND ANL03FL = Y AVISIT PARAMCD CHG ABLFL ANL01FL ANL02FL ANL03FL Week 12 CARB x.x Y Y Y Week 12 HEIGHT x.x Y Y Y Week 12 TOTCAL x.x Y Y Y Week 12 WEIGHT x.x Y Y Y

In SAS proc sort data=addiet out=tabledata; by trtp usubjid paramcd; where anl02fl = Y and anl03fl = Y ; run; proc transpose data=tabledata out=trandata; by trtp usubjid; id paramcd; var chg; run; proc corr data=trandata pearson; by trtp; var weight bmi; with totcal carb (other parameters); run;

Record Selection (actual) proc sql; create table main as select usubjid, trtpn, paramcd, chg from adam.addiet where ittfl = Y and anl03fl = Y and ( (avisit = Post Baseline and paramcd in ( TOTCAL, CARB, )) or (paramcd in ( BMI, WEIGHT ) and anl02fl = Y ) ) order by trtpn, usubjid, paramcd; quit;

Analysis File Metadata Dataset Description Structure Purpose Keys Class of Dataset ADDIET Diet and weight analysis file One record per subject per parameter per analysis visit Analysis STUDYID, USUBJID, PARAMCD, AVISIT BDS Location addiet.xpt

Bonus Example! VS Toxicity Guidance for Industry: Toxicity Grading Scale for Healthy Adult and Adolescent Volunteers Enrolled in Preventive Vaccine Clinical Trials Fever Vital Signs Tachycardia beats per minute Bradycardia beats per minute Hypertension (systolic) mm Hg Hypertension (diastolic) mm Hg Hypotension (systolic) mm Hg Respiratory rate breaths per minute Mild (Grade 1) 38.0 38.4 C 100.4 101.1 F Moderate (Grade 2) 38.5 38.9 C 101.2 102.0 F 101 115 116 130 >130 50 54 45 49 <45 141 150 151 155 >155 91 95 96 100 >100 85 89 80 84 <80 Severe (Grade 3) 39.0 40 C 102.1 104 F >40 C >104 F 17 20 21 25 >25 Intubation Potentially Life Threatening (Grade 4) ER visit or hospitalization for arrhythmia ER visit or hospitalization for arrhythmia ER visit or hospitalization for malignant hypertension ER visit or hospitalization for malignant hypertension ER visit or hospitalization for hypotensive shock

VS Toxicity Shift Table N (%) with No increase Increase by 1 grade Increase by 2 grades Increase by 3 grades Increase by 4 grades (only possible with Fever based on vital signs alone)

ADVS Design Approach: ATOXGR, BTOXGR: Assign toxicity grade AVALCAT1, BASECAT1: Categorize results into Normal, Fever, Tachycardia, Bradycardia, Hypertension (systolic), etc. Ranges for both dependent on PARAMCD SHIFT1: Assign shift as No increase, Increase by 1 grade, Increase by 2 grades, etc.

ADVS Design For systolic blood pressure: (row) USUBJID PARAMCD AVISIT AVAL ATOXGR AVALCAT1 1 01 123 SYSBP Baseline 120 0 Normal 2 01 123 SYSBP Week 1 145 1 Hypertension (systolic) 3 01 456 SYSBP Baseline 110 0 Normal 4 01 456 SYSBP Week 1 79 3 Hypotension (systolic) (row) BASE BTOXGR BASECAT1 SHIFT1 SHIFT1N ABLFL 1 120 0 Normal Y 2 120 0 Normal Increase by 1 grade 1 3 110 0 Normal Y 4 110 0 Normal Increase by 3 grades 3

ADVS Design The shift is based on ATOXGR, BTOXGR without reference to type of toxicity Use AVALCAT1 to select type of toxicity where PARAMCD = SYSBP and SHIFT1 ne and AVALCAT1 in ( Normal, Hypertension (systolic) ); where PARAMCD = TEMP and SHIFT1 ne and AVALCAT1 in ( Normal, Fever );

Karen G. Malley Malley Research Programming, Inc. Email: kgmalley@comcast.net Tel: 301 424 8670