SI10 Implementation of a Metadata-based Approach to Statistical - PowerPoint PPT Presentation

PhUSE EU CONNECT 2018 SI10 Implementation of a Metadata-based Approach to Statistical Planning, Analysis & Reporting By Frank Freischläger & Hanspeter Schnitzer

INDEX Situation & Purpose Clinical Study Metadata Statistical Programming for Analysis Statistical Reporting Statistical Planning Conclusion Page2 // 7 Nov 2018

SITUATION without METADATA Traditional approach of statistical planning, analysis & reporting of clinical studies TOC & Protocol SAP SP Specs Mock TFLs Templates & Standards eCRF Clinical Programming TFLs Study Report DD Specs Raw Data Derived Data Page3 // 7 Nov 2018

PURPOSE Metadata approach to biostatistics Consider all aspects of processing Set up a system built on and compliant with standards Assure flexibility in support of clients and sponsors Let the system do the routine work Limit needs for programming to a minimum Allow statisticians and programmers to concentrate on advanced methods Enable respective shift in manpower Page4 // 7 Nov 2018

CLINICAL STUDY METADATA Page5 // 7 Nov 2018

METADATA REPOSITORY Overview of data panels with content explicitly required per project General Protocol Deliverables Statistics Programming • Clients • Documents • Stakeholders • IE Criteria • Plans • Populations • LibDirSpecs • Projects • Objectives • Derived • Endpoints • Output • Studies • Analyses datasets structures • Intercurrent • Outputs events • Report structures • Reports • Estimands Design Data • Plan structures • Evaluations • Methods • Epochs • Sources • References • Visits • Groups • Disposition • Variables • Exposure Page6 // 7 Nov 2018

FLEXIBILITY in APPLYING STANDARDS Approach with different layers Project-specific content is mandatory Levels are dynamic Other content may be defined on higher levels Content is guaranteed Metadata entry or import Metadata edit checks are in place Independent double entry as an option Individual Study Project Sponsor Default Page7 // 7 Nov 2018

METADATA REPOSITORY Overview of data panels with content that may be given on any level Resulting in project Statistics Programming relevant metadata • Semantics • InDataset Specs Processed and archived • Tools • Settings per production and • Templates • Styles delivery • Method options • Output options • Derivations • Report options • Descriptors • Plan options • Time definitions • Date imputations Page8 // 7 Nov 2018

STATISTICAL PROGRAMMING for ANALYSIS Page9 // 7 Nov 2018

WORKFLOW From derived datasets to booklets of outputs Settings OutputOptions Outputs OutputControls OutputStructures For development Style programming Control Center TFLs in Booklets With original SAS outputs Macro-free Macro Library With log checking source code summaries Derived Data Results Datasets Output Datasets Page10 // 7 Nov 2018

LAYOUTS and FORMATTING Example of PhUSE CSS Analyses and Code Sharing WG on demographics xx (xx.x) Label, n (%) xx ( xx.x) Label - n (%) xx (xxx.x) Label n (%) If 100%? N (100.0) N (100) Label, Unit Label (Unit) If 0? Label [Unit] 0 ( 0.0) 0 - . Missing category required? Page11 // 7 Nov 2018

STATISTICAL REPORTING Page12 // 7 Nov 2018

DATA-TO-TEXT TECHNOLOGY Utilizing methods of Natural Language Generation Developed by Ehud Reiter and others From structured data to text messages and documents Automatic creation of sentences and documents Content selection and surface realization in computational linguistics Examples include weather forecasts, automated journalism Published successful applications to healthcare data Page13 // 7 Nov 2018

WORKFLOW Auto-generated report elements and post-processing Micro-writing and macro-writing Other Metadata & Reports Results Data Flexible post-processing ReportOptions ReportStructures Updateable with new clinical data Sentences Filled Sentences Template Further QC & peer statistical review writing In-text Revisions Report Outputs Finalization Client review Page14 // 7 Nov 2018

SENTENCES for a REPORT Some examples: opening line & short description of demographics A total of 500 patients were screened, of whom 50 (10.0%) failed screening prior to A total of [DISP_SCR_TOT_N] [SUBJ_TXT] were screened, randomization (see Table 14.1.1.1). of whom [DISP_FAIL_TOT_N] ( [DISP_FAIL_TOT_PCT] %) failed screening prior to randomization ( [REF_TXT1] Table [REF_PT_DISP] ). Most patients were female (62.2%) and white (54.4%). Most [SUBJ_TXT] were [DEMO_GENDER_MOSTFRQ_LBL] ( [DEMO_GENDER_MOSTFRQ_PCT] %) and [DEMO_RACE_MOSTFRQ_LBL] ( [DEMO_RACE_MOSTFRQ_PCT] %). Page15 // 7 Nov 2018

STATISTICAL PLANNING Page16 // 7 Nov 2018

WORKFLOW Auto-generated elements of a planning document, and post-processing Similar to report generation Other Metadata Plans Increased complexity PlanOptions PlanStructures Sentences reusable for reporting Sentences Filled Sentences Template Further QC & peer statistical review writing Revisions Plan Finalization Client review Page17 // 7 Nov 2018

SENTENCES for a PLAN Some examples: definition of TEAE & description of an estimand An adverse event is treatment-emergent, if onset An adverse event is treatment-emergent, if onset or deterioration of the event appear after the first or deterioration of the event appear after the first [TRT_LBL] [TRTAPPL_LBL] and no later than [ TE_LAGDEF] [TE_LAGUNIT] study treatment intake and no later than 45 days after the last [TRT_LBL] [TRTAPPL_LBL] . after the last study treatment intake. Population-level summary Difference in means between treatment conditions in the change from baseline to Week 24 in sBP in Measure Parameter the targeted population regardless of intercurrent events. Patient population Specification of consideration of intercurrent events Page18 // 7 Nov 2018

CONCLUSION Page19 // 7 Nov 2018

METADATA-BASED APPROACH New approach of statistical planning, analysis & reporting of clinical studies Control Center Templates & Standards SAP, Table and Mock Versions of TFLs, Statistical Programming Protocol Specifications Metadata Repository eCRF Auto-generated Programming Clinical Study Report DD Specs Raw Data Derived Data TFLs Page20 // 7 Nov 2018

SUMMARY OUTLOOK Extensive MDR Applicable to quality documents like SOPs etc. Flexible & neural Useful for service proposals Productive & reliable Requires careful metadata Improve MDR frontend, backend management and control center Limitations Consider OpenXML or Linked Data as alternatives Challenging in preparing coded links for sentences Page21 // 7 Nov 2018

Questions Comments THANK YOU Frank Freischläger Hanspeter Schnitzer Estimondo GmbH Estimondo GmbH frank.freischlaeger@estimondo.com hanspeter.schnitzer@estimondo.com www.estimondo.com www.estimondo.com