data management Flashcards
what are CRF’s
case report forms
- collect data consistently across sites
- collect data as speciified in the protocol
- help sites in complying with protocol procesdures
collect data in analysable format
what needs to happen before the CRF is signed off
cross functional review and approval process
includes
- drafting it
reviewing it by the relevant teams
- approval
whole process repeated when amendments are made
what do we use to form the data that is to be collected
the protocol
data nees to be collected as is specified in the protocol, for all specified time points
thats IT! dont collect additional data incase it might be interesting
why do we need the CRF to be easy to follow for patients and sites filling it out
this aids protocol compliance and reduces number of queries
lets say we know we want to collect questions on the asberg depression scale, how might the collection of data differ
how they will actually fill this out: checkbox, coded dropdown list
what is a major source of queries in CRF
ambiguous data - make it very clear what is being requested
how can we avoid collecting ambiguous data
include units of measurment and specify the accuracy required (e.g., 6 spaces for DOB)
avoid blank responses - include option to say the measurment was not done or nt applicable e.g., N/A
use headers to say whether certain questions are applicable e.g., do you have any of the following: diabetes - IF YES –>when were you diagnosed etc. if no ignore
clear questions e.g., death instead of death of baby (n might think death of who?)
clear instructions along side of questions to help reduce number of protocol deviations
is one identifier enough for a patient e.g., initials
no, second identifier needed
- patient initials
- month/year of birth
what are CDISC standards and why do we need them?
Clinical Data Interchange Standards Consortium
standards that ensure consistent and reliable data collection and analysis in clinical trials
facilitates collaboration andd sharing of data
what are some key CDISM standards?
- Standards for data collection – cDASH
- study data tabulation model – SDTM
- data analysis – ADaM
what standards are required to be met for a submission of a trial to FDA
CDISC STANDARDS
why is it good to use CDISC standards alongside a template CRF
speeds up process of building a database
how does using CDISC standards alongside a template CRF speed up the process of building a database?
because in the database design we need to define the variables and the data to be collacted in the study
CDISC standards provide a pre-defined and standardised set of vairables and data
e.g., the SDTM provides set of standardised variable names, formats and valuees that can be used to organise collected data
means we dont need to spend time defining nad creating new vairiable names /data formats since they are already provided by the standard
what are the different data capture methods
- traditional paper CRF’s
- electronic CRF’s
- patient completed questionairs (paper or electronic)
- laboratory data
- other external data e.g., blood pressure monitor
what ways are there to colect data that are not the CRF
lab data
data collected from devices or apps
what do we need to make sure about non-CRF collected data
structure - needs to have consistency
format - way that allows for analysis
data manager needs to make sure that transport of the data is not changed accidentaly. needs to have integrity
what are clinical datamanagment systems
software applcications that help manage data collection, cleaning and analysis of clinical trial data
when do we do data cleaning
from when data is first captured all the way to submitting the data for final analysis
what is data cleaning
Identify missing, inconsistent and out of range
checking the data collected in the clinical trial to ensure that there are no missing data points or values that are inconsistent or outside the expected range. This is important for ensuring the accuracy and completeness of the data.
- Data coding queries - review data that is transfered to the standard format for analysis and make sure the data has been coded correctly.
data cleaning
Reconciliation with external data sources:
This involves comparing the data collected in the clinical trial to external sources of data, such as medical records or laboratory results, to ensure that the data is accurate and complete.
data cleaning
review and identify protocol deviations
eviewing the data to identify any deviations from the study protocol, such as missed visits or medication doses, which can affect the accuracy and reliability of the data.
where would we find the data cleaning plan
data managment plan
what kind of things constitute data discrepencies
- Incorrect source data used
- Transcription errors from source to CRF
- Data entry errors
- Incomplete data
- Study identification number errors
how can we identify data discrepencies
- Programmed database checks
- Batch checks on entered data
- Dataset review
- Reconciliation with external datasets
- Safety data review (e.g., checking AEs) and medical coding (translating medical terms into standardised codes)
- Source data verification
what is database lock?
marks the end of data collection and start of analysis
When cleaning data what do you need to look for?
- Incorrect data - look up the min and max values from the questionnaires - any out of this range?
- protocol deviations
- inconsistencies
- missing data
