CDISC Related Services
- 1. Statistical Programing and CDISC-related Services
- 2. Planimeter’s (Hi)story 2 First years Planimeter was founded in 1997 by private individuals to provide statistical services in the pharmaceutical industry especially to support clinical trial design and evaluation. The team was formed by 4 people at the beginning and we primarily concentrated on design and analysis of Phase III-IV clinical trials, post- marketing studies and epidemiological surveys. E-business solutions We started our e-business activity in 2005 with preparation of our first Patient Registry. This was followed by further registries and the first e-CRF in 2006. Our e-CRF solution became internally well known and well accepted within a reasonably short time pf period. Probably the main reason of our success is that all our experience on conduction and evaluation of clinical trials were incorporated during design and implementation of the system. Our solution – equipped with a high level help-desk service – currently operates in 8 countries. Besides registries and eCRFs we also developed e-learning, dating and special CRM systems specifically build for health-care system support. Our today activities Meantime Planimeter has became an internally acclaimed, full service CRO of a size of 16 employees and a similar bunch of contractors. Our today activity covers – beyond the original targets – full conduction of clinical trials and reporting - including the medical writing activities. Our strength lays in the strong mathematical background of our staff. We are experienced in both classical (parallel, cross-over, etc.) and modern (Bayesian, Adaptive, etc.) designs. Recently we are focusing on support of translational research and implementation of bioinformatics solution. Main therapeutic areas are: vascular and nervous system disorders, ophthalmology, infections, oncology, renal and urinary disorders, immunology and respiratory disorders. In 2014 Planimeter became an ISO qualified company.
- 3. Mission and Vision 3 Vision Our vision is to be an internationally known and accepted service provider in Pharmaceutical industry. One day we would like to turn our rich knowledge and experience in biostatistical / bioinformatical innovation. What we do not aim at any circumstances is to be ‘leader’ in business or measurements aspects. We wish to provide top-level services with a manufactoral taste and comfort. Mission Our mission is to serve as a source of expertise in delivery classical and up-to-date biostatistical services in pharmaceutical, biotechnology, medical device industries world-wide, and to promote the use of rigorous quantitative methods in the biomedical and public health sciences. Our approach is to capitalise on our presence and network in Hungary and use information technology to assist the life sciences industry in the development of quality products, on time and on budget, every time.
- 4. Current activities 4 DM Statisticians PV Programmers Help-desk eCRF Patient Registry Randomisation Data Management Data Entry Systems Query management Coding Statistical Programing CDASH / CDISC TLG – design and implementation Post-hoc analysis Web-based solutions eCRF Patient registry Support of PV E-Learning Post-marketing support Patient- support program eCRF Dashboard Statistical Services and Data Management for Healthcare
- 5. Our Team 5 István Jánosi Founder, Managing Director István graduated at KLTE, in Debrecen, Hungary as a mathematician in 1990. He has been working as biostatistician since then. He also manages – besides Planimeter - WEB2 Research Kft. Member of ISCB and DIA, author or co- author of several publications. Acting as statistician, data manager – on demand – and primarily as the manager of the company. László Szakács Head of Statistical Programming and Medical Writing László has been working for Planimeter since his graduaton as a mathematician (2003). All of our recent systems (automated report generation, QA, web-based solutions, etc.) were developed or designed by him. He took full responsibility of complete management of DM and stat activities of several Phase I-IV studies. Bíbor Balázs Head of Administration, Key Contact Manager Bíbor joined Planimeter in 2002. She is an active member of SOP-develeper team. She takes full responsibility for all logistic issues regading data management. She is also a member of our pharmacovigilance team, she is responsible for monitoring of medical literature.
- 6. References 6
- 7. Statistical Programing and CDISC-related Services
- 8. Business Idea 8 Statistical Programing Statistical Programing primarily means derivation of Tables, Listings and Graphs (TLGs). This activity is performed with the help of SAS, so Statistical programing practically means programing in SAS. In wider sense statistical programing covers the majority of CDISC-related tasks because SDTM and ADaM data tables are also prepared with the help of SAS. Quality Assurance with Chili A sophisticated Quality Assurance procedure was developed and introduced to support the preparation of data models and the process of statistical programing. QA can be performed with different depth depending on the Sponsor’s requests. The whole procedure is documented within Chili, a free web-based tool, which was configured specifically for QA-purposes internally. Data Models (CDISC) CDISC SDTM (Study Data Tabulation Model) is applied to re-structure the raw clinical data into the generally used domains with the help of variable naming conventions. In this approach the Submission Metadata Model is followed as far as possible. ADaM (Analysis Data Model) is used to prepare datasets for direct reporting purposes. The activity is extended with preparation of the Define.xml and Study Data Reviewer’s Guide. Management support with Chili Chili is not only suitable for documentation of the QA- procedure, it also provides environment for delegation of tasks and supervision of both the development and QA- procedures. With granting access to the Sponsor the whole procedure can be monitored with full transparency.
- 9. Detailed Services - Menu 9 CDASH Clinical Data Acquisition Standards Harmonisation – further details SDTM More detailed description of our services ADaM Data models – as we apply this approach in our everyday routine Define.xml Description of our interpretation of generation of define.xml TLG Preparation of Tables, Listings and Graphs QA-procedures Brief description of our QA- procedures QA in Chili QA implemented within Chili framework Management with Chili Management support (development and QA) within Chili framework
- 10. The Clinical Data Acquisition Standards Harmonization (CDASH) model standardizes the way data is collected to facilitate the generation of SDTM tables. As the primary aim of following CDASH Guidance is to make easier and more comfortable to transform raw study data into SDTM domains, CDASH is taken into consideration during CRF (and eCRF) design. CDASH approach is primarily applied by our Data Management Department, but being a small company there are not strict boundaries between DM and statisticians or DM and programmers. Consequently majority of our staff including data managers, statistical programmers and statisticians are familiar with the CDASH standards. Similarly to the CDASH / CDISC community we are also continuously working on building our CDASH libraries and improvement of already implemented modules. CDASH 10
- 11. SDTM 11 Preparation of SDTM domains is quite a well defined procedure, which primarily means the proper application of the guidance in case of the standard variables and domains (e.g. sex, age or blood pressure). Due to the nature of SDTM even the study or therapeutic area specific variables can easily been transformed into SDTM- compliant variables with following the naming and formatting conventions. As SDTM is a living concept in itself , we pay great attention to its own development: we intensively follow the achievements regarding development of Therapeutic Area Data Standards (TAUGs), Metadata Submission Guidelines (MSG) or SDTM implementation Guide for medical devices (SDTMIG-MD). Our services cover: • Creation of Data Standards Library including a CRF Library, and CDASH/SDTM/ADaM Libraries • Set-up Study Specific aggregates mapping visit information into the appropriate visit window structure • Development of a metadata repository to enable data model compliance checking and define.xml generation • Comparison against Study Specifications • Validation against the Data Standards Library • Generation of pooled analysis based on CDISC data (for ISS or ISE purposes) • Creation of macros for ADaM dataset creation.
- 12. SDTM and ADaM data sets are programmed according to the associated specifications, and validated against a series of electronic integrity checks to ensure compliance to the models. Additional QC includes independent verification of results. In the Subject-Level Analysis Dataset (ADSL) only one record is created for all subjects. In the Basis Data Structures (BDs) one or more records per subject are generated per analysis parameter, per analysis time point (conditionally required). The so called „Other” section of ADaM specifically contains – among others – the Basic Data Structure for Time-to-Event Analysis and Data Structure for Adverse Event Analysis. ADaM 12
- 13. Define.XML 13 Information sharing on the content of SDTM in a more or less human-readable manner led to introduction of Define.xml, which – in short – contains all SDTM data AND metadata in XML- format. The Standard content of define.XML is : • Data Metadata (TOC) • Variable Metadata • Variable Value Level Metadata • (Computational Algorithms) • Controlled Terminology/Code Lists • Annotated CRF • Optional: Supplemental Data Definition Document. Computational algorithms can also be stored within define.xml framework. Further advantage of Define.XML: Content can be checked (validated) with define.XML checker programs (e.g. http://www.pharmasug.org/proceedings/2013/AD /PharmaSUG-2013-AD19.pdf).
- 14. TLG 14 Tables, Listings and Graphs are defined in the Statistical Analysis Plan with the help of Table Shells. Theoretically table shells defines the outputs with character precision (in case of Tables and Listings). Planimeter developed a highly sophisticated SAS-macro set which is able to deliver the required outputs with very high flexibility. What is the best in our solution that the macro set can be adjusted through an interface which does not require programing knowledge. So if no new template is required, the outputs can be put together simply in a text-editor. While each study requires some further development of our SAS macro set, greater and greater part of tables, listings and graphs can be used in off-the-shelf manner. Today we are generally cover a typical clinical study with pre-defined outputs in 65-85%. The advantage of this high coverage is that our TLG generation costs can be decreased reasonably proportionally with the decrease in work-load.
- 15. Quality Assurance Procedures 15 Steps establishing quality during development phase • Application of programing conventions • Commenting and segmenting • Application of continuously developed macros in TLG-derivation • Double programming on demand • TLG-derivation in a three-stage method: development, testing, production • Complete and transparent documentation of the whole activity (code development and QA) • Continuous development of standardization according to international standards and guidelines. Steps establishing quality during QA phase • Code and output review • Checking of the applied filtering and sorting in the database s • Checking of program codes for programming conventions • Checking log files for errors, notes, messages • Comparison of the results against the table shells • Checking of titles, notes, remarks, and spelling. In case of any discrepancy notification of the programmer and detailed description of the finding. Response: programmer’s action: fixing the bug or explanation of the discrepancy. In case of data error notification of project manager. In case of specification issue notification of the study statistician.
- 16. QA in Chili 16 Chili (https://www.chiliproject.org/) is a free, web-based tool which perfectly fits the demand of managing and documenting of QA-procedures of statistical programing. Adjustment of the application starts with role definition and set-up of user authorizations. This is followed by the description of all the potential tasks (the program codes themselves) and finally with the assignment of the tasks to the users (programmers). When a task is indicated as „ready for validation” the assignments skips to the dedicated QA-person responsible for the specific task. Not only the work-flow can be commanded within Chili, it also provides area for showing the actual status of a specific program.
- 17. Management with Chili 17 Chili is a perfect tool of monitoring both the program development procedure and the QA-activities. Monitoring can be applied at task-level, when all the activities related with one specific task are tabulated. Summary tables of arbitrary sets of tasks can also be created. With suitable authorization the Sponsor can follow the development and QA-procedure without any additional effort in real-time.
- 18. E-business solutions 18 Automated Business Data Reporting There are many data sources (e.g. IMS) which make possible of preparation of regular reports. Although the same task should be repeated, sometimes this repetition is not supported within the original software framework. We are able to transform regular data into regular reports with a maximal flexibility. Patient Support Program Patient Support Programs are to incorporate information collected by usual eCRFs extended with further study-specific activities of an enrolled subject, like communication with call-center, patient organisations, dietitians, running a Diary of rating the different services on regular basis. E-learning System Our e-Learning system has two roles. First, trains the materials as the name says. If we consider the SOP- system as a potential material to train/learn, then we arrive to our second task: to provide a job- or role-dependent delegation of SOPs to be studied to specific 'students'. Medical Information System Aim of Medical Information System is to collect, document and manage all potential AE-related information reaching a pharma headquarter independently of the communication channel (web, telephone, CRA, trial, etc.). The system delegates the tasks, logs and reports the activities.
- 19. Statistical Programing / CDISC references 19 Vifor Ferinject® in Patient With Thrombocytosis Secondary to Inflammatory Bowel Disease (IBD) Dignity Science A Randomised, Double-blind, Placebo-Controlled, Single- Ascending Dose and Multiple Dose Phase I Study to Assess the Safety, Pharmacokinetics and Effect of Food on Orally Administered DS107G in Healthy Subjects Boehringer Ingelheim ADESPI: Adherence to Spiriva® in Patients With COPD (Chronic Obstructive Pulmonary Disease), Measured by Morisky-8 (MMAS-8) Scale, in Routine Medical Practice Apellis A Phase I Study to Assess the Safety APL-2 as an Add-On to Standard of Care in Subjects With PNH ABBOTT / Abbvie Observational Study to Evaluate the Time to Achieving the Maintenance Dose of Zemplar® (Paricalcitol Injection) in the Treatment of Patients Suffering From End-stage Renal Disease and Severe Over-reactivity of the Parathyroid Glands AstraZeneca CorRELation Between PatIent PErception and Findings on Clinical Examination (RELIEF)
- 20. Characterisation of the staff 20 Role # Availability for 2015-2016 Q3 Q4 Q1 Q2 Head László Szakács On-demand 60% 70% 80% Senior Programmer (8+ years of exp.) Hourly rate (€) 1 empl’s +2 contr’s 65 0+2 1+1 1+1 1+2 Advanced Programmer (5+ years of exp.) Hourly rate (€) 2 empl’s +3 contr’s 48 1+2 1+1 1+2 2+3 Novice Programmer (2+ years of exp.) Hourly rate (€) 2 empl’s +3 contr’s 36 1+0 0+1 2+0 2+2 Total 3+4 2+3 4+3 5+7
- 21. Vaci út 95 Budapest Hungary, H-1139 +36 30 933 89 20 +36 30 429 88 05 Contact Details 21 www.planimeter.hu janosi@planimeter.net bibor@planimeter.net