33
DATA HANDING AND ANALYSIS
• Data analysis plan, inlcuding statistical methodology and planned tables and figures: Describe the
sampling methods, information collection procedures, methods to maximize response rates, test
procedures and relevant statistical quantities (e.g., variance, confidence intervals and power based on
data from the study) in sufficient detail that the methods are reproducible. This includes calculation of
relevant quantitative measures for methods and reproducible. This includes calculation of relevant
quantitative measures for test and instruments, such as sensitivity and specificity. In outbreak
investigations, it is common to employ an iterative process in the analysis (consisting of developing and
testing hypotheses and planning and evaluating interventions) to identify the source of the outbreak and
control it. For projects establishing or utilizing data from a surveillance system, this could include how
and how often the surveillance system will be evaluated. Describe what tables and figures are planned
to present study results.
• Data collection: Describe data collection procedures, processes and documentation. For data emanating
from a surveillance system, this would include frequency or reports.
• Information management and analysis software: Provide the names of data entry, management and
analysis software packages and computer programming languages to be used for the project.
• Data entry editing and management, including handling of data collection forms, different
versions of data, and data storage and disposition: Describe the overall procedures for management
of the data collected. Include in the description the process for entering and editing data. Describe how
study materials, including questionnaires, statistical analyses, unique reagents, annotated notebooks,
computer programs and other computerized information, whether used for publication or not, will be
maintained to allow ready, future access for analysis and review. Document operating procedures for
managing and accessing different versions of data sets. State who the data belong to and any rights to
and limitations to access for any primary and secondary data analyses and publications. Documents
procedures regarding confidentiality of the data, including how confidentiality will be preserved during
transmission, use and storage of the data and the names of persons or positions responsible for technical
and administrative stewardship responsibilities. Document what the final disposition of records, data,
computer files, and specimens will be, including location for any relevant information to be stored.
• Quality control/assurance: Describe the steps that will ensure no unintended conseqences that could
affect the quality of the data. Those steps might include methods to capture all reported data exactly as
received, assuring logical consistency among all parts of a record and ensuring that manipulation or
transformation of the data (e.g., from audio tape to transcribe text) produces no unintended changes, and
verifying that statistical and arithmetic calculations are performed as proposed in the data analysis plan.
For outbreak investigations, this would include verifying diagnosis and confirming the outbreak.
Describe procedures for ongoing data quality monitoring to assure that information of appropriate
depth, breadth, specificity is collected and remains consistent within and among staff over time, and
acceptable levels of such attributes as validity, reliability, reproducibility, sensitivity and specificity are
achieved.
• Bias in data collection, measurement and analysis: Describe the kinds of bias that may occur in
collecting the data or in the measurement or analysis phases, and the steps that will be taken to avoid,
minimize and compensate for the bias. Include factors in the study population or in study personnel that
could bias results, as well as the steps that will be taken to assure vaild self-reporting or recording of
observations. Include any randomization and blinding procedures that will be used to
eliminate/minimize bias by investigators, other study staff or participants (e.g., in selection of
participants, allocation to treatment groups, providing/receiving treatment).