The following are the general activities required for clinical research informatics. These are used as a framework for the ways in which our Data Management services add value.

We assist with making sure study personnel are familiar with the system they are using. This can be done through in person training, video conferences, and documentation. We also provide help line support for end users that have issues with browsers, plugins, and other customizations that users may have made to their environment. This extends out to supporting barcode printing for sample logistics.

We manage user level access and passwords. We also keep internal records of human subject training, send reminders when training is about to expire, and will lock accounts if an individual user is not current on their training.

Data managers can support enrollment by importing batches of subjects, generating enrollment packets, and monitoring enrollment numbers. Specifics will be tailored to the study and how enrollment will happen.

Studies often have a hybrid approach to enrollment and gathering data. This can be a result of a grant continuing a study under a new environment or a data capture system changing over time. When this happens, it is necessary to operationally move subjects between groups, classes, and events.

While data managers do not do standard data entry, there are a few ways they impact this part of a study. First, they are responsible for the construction of individual forms used in the data entry process. This includes skip logic, bounds checking, and general structure. They will also play a role in merging data from other sources into the unified data set.

Samples need to move through a chain of control to ensure valid data are captured from them. Data managers monitor these samples to ensure correspondence between the study events and the samples collected. They will detect samples that have not completed processing to their final storage destination.

As studies begin to access their data, the managers will support access mechanisms the run the spectrum from de-identified csv downloads, to direct ODBC connections, on up to supporting a secure analysis server that would never move subject data onto the individual machines of study personnel. The data manager will configure data access permissions that will allow customized access on an individual level. They also train on the storage and transmission of PHI and sensitive subject data.

In the event that the original source data does not directly correspond to the final analytical variables, the data managers will perform coding. This can be coding of plain text, where the manager will provide a mechanism to reliably perform that coding. Another option is preparing additional variables that are deterministically coded based on other variables. By having a data manager do this task prior to extraction, we ensure all references to the new variables use consistent logic rather than distributing the logic among all the points of access.

We will monitor for data quality through standard rules such as checking for missing values, data types, bounds checking, outliers, values in hidden fields, and confirming responses match coded values. This is expanded with custom rules on a per study basis to include multivariate date validation, other/please specify pairs, changes across events, and other logical inconsistencies.