1
DataCenter Team, Center for Research on Health Care
For each project, the DC brings a team of experts in data entry, data management, systems analyst/programmer, and statistics with a data management director to oversee and supervise the team. Each service was calculated by an hourly rate. This amount was derived as a result of the cost center model in which the DataCenter operates. Operating similarly to a non profit organization, the cost center model enables the DataCenter to bill an hourly rate for services that directly reflects the costs for those services. This model incorporates supervision, training, quality assurance, and data integrity as part of the cost for each service provided by the DataCenter. The hourly rate incorporates facilities and administration costs since the hourly rate is billed as a direct service. No indirect rates are charged.
The services consist of data entry, data manager, systems analyst, data analyst and statistician. The data entry personnel are responsible for data entry and tracking and are billed at $29 per hour. The data managers develop the data entry screens and the paper forms as well as run verification and are billed at $41 per hour. The rate for systems analysts is $65 per hour. The systems analyst develops the databases for tracking and the study data and generates reports as needed for the study. The data analyst assists the statistician with the analyses and data cleaning at $56 per hour. The statistician is responsible for conducting data analysis; our PhD level statisticians are billed at $98 per hour, and our Masters level statisticians are billed at $77 per hour. The director of the DataCenter serves as a liaison between the data team and the PI. She supervises the team and conducts all data team meetings for the project. The cost for this service is already incorporated into the hourly rate.
Data Entry – Primary
Each data entry person has extensive research experience in data entry and tracking research subjects. This person will be responsible for the primary data entry as well as assisting the project coordinator with tracking subjects. As a member of the DataCenter, data entry personnel will be an integral part of the data management team.
In order to calculate the amount of time required for data entry, we counted the number of subjects, questions, and data collection time points in the study. We followed the guidelines published in Statistics in Medicine (Lesser, 1996). Estimating that it would take approximately 0.10 minute to enter the response to one question, times the number of questions in a protocol and number of subjects and follow-ups, it was calculated that data entry for the study protocol would require XX% of a data entry personnel’s time. Given that the study also involves chart abstraction, an additional XX% was estimated based on it taking 45 minutes per chart for each subject in the study. Therefore, XXX hours are needed for the first year and XX% for subsequent years or XXX hours at a rate of $29 per hour.
Therefore, XX% of a data entry personnel or XX hours will be devoted to the project during the course of the study where the person will contribute to the development of the data entry program, test that program, key enter the data and assist with tracking.
Data Entry – Secondary
All data are double entered in order to generate a database that is resolved of data entry errors. From our experience, single data entry results in 6-20% error rate. Double entry eliminates this error. Two databases are created, a primary and secondary and verification is run between the two databases. Another data entry person is needed to enter the data in the secondary database. Given a more limited scope of responsibilities (solely data entry), XX% or XXX hours are needed for the duration of the study. This was calculated based on 75% of the primary data entry effort.
Data Manager
Each data manager has extensive research experience in data management. In order to calculate the amount of time required for data management, we counted the number of subjects, questions, and data collection time points in the study. We followed the guidelines published in Statistics in Medicine (Lesser, 1996). The data manager would be responsible for creating the paper form and data entry screens for the study, creating the tables for the database, running the verification program, and resolving discrepancies between the primary and secondary databases. Comprehensively, it would take approximately 31.2 minutes to complete these tasks per variable in the study. Taking into account the number of questions in a protocol and number of subjects and follow-ups, it was calculated that XX% time for the first year or XXX hoursare needed to assist the systems analyst in developing the databases and XX% in subsequent years of the study or XXX hours at $41/hour. Given that the study also involves chart abstraction, an additional XX% for the first year was estimated or XXX hours and XX% for subsequent years or XXX hours, based on an estimate of 50 minutes per chart being required to generate the data entry screen, paper form, and conduct verification for each subject in the study.
Systems Analyst
The systems analysts have extensive research experience in computer systems, database development, and programming. In order to calculate the amount of time required for the systems analyst, we counted the number of subjects, questions, and data collection time points in the study. We followed the guidelines published in Statistics in Medicine (Lesser, 1996). The systems analyst would be responsible for developing detailed study protocol specifications and data management specifications. It is estimated that each of these takes 10 minutes per variable. The systems analyst also specifies coding for each data item, including validation specifications as well as programming skip patterns for the forms, for a total of 20 minutes per variable. Creating queries and reports for data safety monitoring as well as other needs in the study and programming the databases are additional responsibilities of the systems analyst for 25 minutes per variable. The systems analyst also spends an additional 10 minutes per variable backing up the databases and generating datasets for the statistician. All of these tasks take approximately 75 minutes to complete per variable in the study. Taking into account the number of questions in a protocol and number of subjects and follow-ups, it was calculated that XX% time is needed for the first year or XXX hours and XX% for subsequent years or XXX hours at a rate of $65/hour. Given that the study also involves chart abstraction, an additional XX% was estimated for the first year or XXX hours and XX% for subsequent years or XXX hours based on an estimate of 50 minutes per chart required to complete the previously specified tasks for each subject in the study.
Data Analyst Justification
The data analysts have extensive research experience in analyzing and cleaning data. They work under the direction of a statistician and create datasets and variables, as well as conduct brief analyses. XX% of a data analyst time during year one is needed where the analyst will work closely with the Principal Investigator and statistician for a total of XXX hours at a rate of $56/hour. In years two – four, the analyst will generate interim analyses for our Data Safety Monitoring Board and will conduct preliminary analyses on the data for presentations at XX% effort or XXX hours. In the final year of the grant, XX% effort is needed for analyses and manuscript and report preparation or XXX hours.
Statistician Justification
Each statistician holds a Ph.D. or equivalent level degree. All statisticianshave extensive research experience in clinical trialsand methods for health services research and has statistical expertise in biostatistical methods such as survival analysis and random effects models. The statistician will be responsible for developing the randomization assignments and all statistical analyses associated with this research. Fifteen percent of a statistician’s time is needed during year one to work closely with the Principal Investigator and will be part of the investigative team or 293 hours. Regular meetings are necessary with the data management team on the development of a relational database structure that will facilitate data analyses at study completion. In years two – four, the statistician will generate interim analyses for our Data Safety Monitoring Board and will conduct preliminary analyses on the data for presentations at 10% effort or 195 hours each year. In the final year of the grant, 25% effort will be spent on analyses and manuscript and report preparation for a total of 488 hours at $98/hour. The statistician is responsible for all final analyses of data.
Data Management Director
As part of quality assurance and research integrity, Doris M. Rubio, Ph.D., Director, DataCenter, Center for Research on Health Care, University of Pittsburghserves as the Data Management Director. With extensive research experience, Dr. Rubio will be responsible for directing the data management team and ensuring the integrity of the management process. Dr. Rubio will devote approximately 5% of her time during the course of the study where she will work closely with the Principal Investigator, Co-Investigators, and data team to assure that the data management is being implemented correctly and efficiently while maintaining high quality. She will meet regularly with the data management team and the Principal Investigator, as well as participate in the investigators meetings to assure that the procedures are being followed and will oversee the quality assurance plan for the study. Dr. Rubio will conduct all data team meetings and will monitor the effort spent on the project by the data management team to ensure that time is being used efficiently. The cost for this service is built into the hourly rates for each service.