Pittsburgh Claude D. Pepper Older Americans Independence Center

Data Management, Analysis and Informatics Core (DMAIC)

The overarching goal of the DMAIC is to provide a central source of expertise and services by a team of faculty and staff familiar with the theme and methods used in the Pepper Center. Services are most effective when they are provided by personnel intimately familiar with the unique issues of the theme, its special research questions, methods, populations and measures. We continue to achieve increased efficiencies due to standardized data entry and management and quality control processes across studies. In addition, we provide special expertise required to address the unique issues involved in studying balance and mobility in older adults, such as methods for falls surveillance, informative censoring and management of related missing data, and novel application of complex techniques for quantifying subtle features of gait.


DMAIC Organizational Structure

Subashan Perera, PhD, Associate Professor of Geriatrics, is the core co-leader Dr. Perera has 14 years of experience in providing statistical support to aging-related research at major academic medical institutions, with the vast majority of projects focusing on some aspect of balance and mobility in older adults consistent with the theme of the Pittsburgh OAIC. Dr. Perera has held Core Leadership positions in multiple OAICs over this period, including Kansas and Pittsburgh Claude D. Pepper OAICs. In addition, he has been the principal statistician in 6 other P01, K07 and T-series program/training grants; and over 12 R-series projects. He is currently the principal source of statistical support to 4 large intervention and patient-centered studies funded by PCORI, AHRQ, CMS and VA. He has also played a statistical mentorship role in over 6 funded K-awards.

Doris Rubio, PhD, Professor of Medicine, Biostatistics, Nursing, and Clinical and Translational Science, is the core co-leader. She is also the Director of the Center for Research on Health Care Data Center that provides data management, web design and development, qualitative research services, and statistical analysis across the University and Medical Center. Dr. Rubio directs Evaluation Core and the Co-Director of the Education and Career Development Core for the Clinical and Translational Science Institute (CTSI) since the CTSI was funded in 2006. In addition, she is the co-director of the Institute for Clinical Research Education which serves as the education arm of the CTSI. Dr. Rubio is the director of the office of evaluation for the ICRE and in this role she evaluates all of the degree programs and career development programs. Dr. Rubio manages the day to day operations for the KL2 and serves in an advisory role for the Scholars.

Steve Handler, MD, PhD, Assistant Professor of Medicine and Medical Director for Long-Term Care Health Information Technology at UPMC has been instrumental in facilitating biomedical informatics technology in the LTC setting and will enrich and expand our capacity for clinical research in the renewal. Both Drs. Nace and Handler have been key investigators in developing the renewal LTC registry.

Robert (Bob) Boudreau, PhD, Assistant Professor of Epidemiology. is a PhD Biostatistician and active member of the Pepper Center’s Data Management, Analysis and Informatics Core (DMAIC). In addition to the DMAIC, he serves as the Core Director for Biostatistics in the Center for Aging and Population Health as well as lead biostatistician in the Health Promotion and Disease Prevention Research Center.

Ervin Sejdic, PhD, is the director of the Innovative Medical Engineering Developments (iMED) Lab at the University of Pittsburgh and the associate director of the RFID Center of Excellence at the University of Pittsburgh. Dr. Sejdić and his lab aim to develop dynamical biomarkers indicative of age- and disease-related changes and their contributions to functional decline under normal and pathological conditions by fostering innovation in computational approaches and instrumentation that can be translated to bedside care. Important aspects of this research are Dr. Sejdić’s efforts to understand age- and disease-related changes in human gait, especially when considering gait in ideal and real-life conditions.

Dr. Sejdić and his lab will focus on the computational and instrumentation aspects of the proposed developmental project for the DMAIC.

(Return to Top)



Data Management

We have the expertise to develop an electronic data management system for both paperless and paper-based studies.

  • For a paperless study, we create an electronic System for Data Management (eSYSDM), based on detailed study protocols and requirements that includes an electronic case report form and a tracking system, with the capability to incorporate EHR. The eSYSDM is developed using .NET 4.0 to create the interface and SQL for the database. The system is designed so that the recruiter can bring an electronic device (computer, iPad, tablet) bedside and record the data directly into the system. We work closely with investigative team and other study personnel to ensure that protocols are being followed, data integrity and confidentiality are maintained, and that the data contains a minimum amount of missing data. We also have expertise to provide web-based data entry, including from mobile devices and smart phones.
  • For paper-based studies, a data entry system will be created using .NET 4.0 or Microsoft Access®. The database will be created in SQL. We will create a tracking database that will be linked to the data entry database that will monitor enrollment, and track follow-up rates and the data entry process, providing up to date status reports. All completed forms will be forwarded to the DC where the data will be double-key entered. The database will include routine data edit checks for consistency both within and between forms.

To get help with your data management, please contact Dr. Doris Rubio.

(Return to Top)



Study Design and Statistical Analysis

We meet with Pepper investigators early in study conception to assist with:

  • ensuring hypotheses are relevant and testable
  • analysis of preliminary data and sample size and power computations
  • development of the analysis plan, randomization (if applicable)
  • analysis of data
  • interpretation of results
  • generating relevant graphics and text for dissemination of results in presentations and manuscripts

We use SAS as our main general purpose statistical software, and Salford Predictive Miner, Stata, SUDAAN, Solas, StatXact, LogXact, PASS, NQuery as needed for specific requirements involving classification/regression trees, frailty models, complex survey data, missing data, exact inference and sample size/power computations.

DMAIC is proficient in providing all standard analytic support such as general (ANOVA, ANCOVA, regression) and generalized (logistic/Poisson/negative binomial variations) linear models, generalized estimating equations and linear mixed model extensions of the above, survival analysis (Cox & frailty models), time series analysis (Box-Jenkings, ARIMA, ARCH) and multivariate methods (factor, cluster, principal component, discriminant, multivariate ANOVA analyses).

To get help with study design or analysis, please contact Dr. Subashan Perera.

(Return to Top)



Informatics

DMAIC is now able to provide informatics support to Pepper Investigators. We assist with capabilities of various databases of the UPMC Health System, MARS, and falls data warehouse. We will assist in merging data across different databases using multiple identifiers; and de-identify the data for research using appropriate methods. Through access to these databases, we provide capabilities for real-time event surveillance such as falls, adverse drug event monitoring and adverse event monitoring in clinical trials.

To get help with informatics, please contact Dr. Steve Handler.

(Return to Top)

This website was created by the Center for Research on Health Care (CRHC) Data Center. Contact the webmaster if you are experiencing any problems with this site.