Neurological Disorders Database Repository
What is the NDDR?
The health care environment is in the
midst of profound changes necessitating a sharper focus on outcomes
assessment as a positive strategy for the evaluation of clinical protocols
(Joint Commission on Accreditation of Healthcare Organizations, 1994). The
'Clinical Database Repository' is an innovative technological tool used to
integrate data from diverse clinical systems for outcome assessments such as
evaluating benefits of new technology, validating practice guidelines, and
analyzing the effectiveness of care (Larson & Straub, 1996). This study
expands the clinical database repository concept to integrate research with
clinical data in a relational database (Neurological Disorders Database
Repository) that can be accessed by a multidisciplinary team of clinical and
basic science researchers interested in aging and neurological disorders.
This novel approach to disease state management is notable and could be used
to investigate additional aged population-based health risks.
Our Purpose
The purpose of this study is to establish the Neurological Disorders
Database Repository (NDDR) that provides opportunity for longitudinal and
vertical (cross-sectional) outcome analyses in individuals with and without
the diagnosis of Alzheimer's Disease and related disorders (ADRD).
In this effort, the research team has develop a tripartite infrastructure of clinical, basic science, and information processing to allow the testing of different apriori investigator- developed hypotheses as well as new hypotheses generated from statistical database mining techniques. Clinical and basic science researchers and information specialists work as a team to develop the NDDR and implement investigative protocols. This will provide researchers a resource to cite in extramural grant proposals that demonstrates access to broad-based patient information. Awards from such grant applications are used to support maintenance and future expansion of the NDDR.
Initially, the focus was on supporting specific research protocols related to Alzheimer's disease and related disorders. However, the NDDR is deliberately designed conceptually to permit the addition of research protocols focused on aspects of other neurological disorders or issues in an aging population. The establishment of the core of the NDDR supports the design and implementation of the 'Initial Encounter Form' for the collection of comprehensive medical history and current health status parameters. The core database will expand (after appropriate human assurance review) to include advanced algorithms with protocols addressing complicating co-morbidity of dementia, Alzheimer's disease and related behavioral disorders, family caregiver health, and caregiver-patient interrelationships. The core database has been structured such that new research questions can be added as the need arises.
| Specific aims 1: To establish procedures and programming for a NDDR that will be used as a core research vehicle by a multidisciplinary team initially for the investigation of ADRD, and specifically, to develop the “Initial Encounter Form” and subject population recruitment protocols. |
Collaborative Effort With The
Center for Senior Health
The impetus for clinical input for this
project initiates from the MCG Center for Senior Health. Services at the
Center for Senior Health include primary care, delivered by physicians
certified in geriatrics and/or internal medicine, at the Center and in
nursing homes and satellite clinics in senior housing communities. The
Center coordinates conferences to help families determine the needs of loved
ones in long-term care planning; patient education and counseling tailored
to meet individual patient needs; and seminars on such topics as living with
Alzheimer's disease, multi-medication management, preventive medicine and
health promotion, depression in the older person, and normal aging.
Significant amounts of clinical data from Center patients already reside in
electronic database format. This database serves as the platform for
expansion to the NDDR for those patients choosing to participate through
informed consent in accordance with MCG Human Assurance guidelines.
Collaborative Design
Figure 1 shows the integration of research efforts by members of the ARC
and the Center for Senior Health made possible by the Neurological Disorders
Database Repository (NDDR). The core NDDR will support the design and
implementation of the Initial Encounter Form for the collection of
comprehensive medical history and current health status parameters.
Figure 1. Integration of Research
Subjects include healthy individuals and those with diagnoses of ADRD. Both groups are followed longitudinally and analyzed cross-sectionally by established and putative risk factors in advanced protocols.
| Specific aims 2: To expand the NDDR to support the investigation of specific apriori hypothesis driven protocols as well as hypotheses generated by data mining utilizing correlations among clinical and research findings warehoused in the repository and to develop “advanced protocols” based on the investigator initiated problems delineated in Appendix C. These protocols will include clinical data and histories, various blood values, putative and established genetic risk factors as well as behavioral, personality, and caregiver-patient factors. |
Plans for Expansion
The core database will be expanded (after appropriate human
assurance review) to include advanced algorithms with linking clinical
protocols.
Recently, it has been reported that the world's information load doubles every 20 months (McDonald, Brossette, & Moser, 1998). As the body of data expands in both size and complexity, new methodologies are required for analysis. Traditional methods of data analysis are limited by the fact that they are manual and confirmatory (hypothesis-driven). Therefore, traditional statistical analyses cannot lead to discovering patterns that are not already suspected. Knowledge discovery through data mining is an emerging, interdisciplinary research field that represents the intersection of the computer sciences, statistics, and any number of application domains (Fayyad, Piatetsky-Shapiro, & Smyth, 1996).
In the domain of health services delivery and outcomes assessment, these new methodologies can overcome the shortcomings of traditional statistical constructs by discovering previously unrecognized patterns of interest in databases related to health care surveillance, practice guidelines, and molecular sequence analysis.
The patterns generated by data mining can then serve to generate or suggest hypotheses that can be independently tested with traditional confirmatory methods (McDonald et al, 1998).
The advantages of routinely collected laboratory data in outcomes assessments are confirmed (Buaki & Moreau, 1995; Tierney, Miller, James, & McDonald, 1995. However, widespread use of these parameters in assessing clinical outcomes has been hindered not only by the lack of appropriate data systems organized for data warehousing and analysis, but also by lack of accessibility to data, and limitations in data presentation (Larson & Straub, 1996). The NDDR proposed in this study will address these deficiencies by providing the appropriate data and analytic tools to pursue the value of assessment of routine laboratory data in prediction, classification, and staging of ADRD.
Epidemiological studies suggest that health and environmental factors, among them diabetes, anemia, exposure to organic solvents, expression of certain gene variants, and elevated levels of beta amyloid peptide, are associated with ADRD (Beard, Kokmen, O'Brien, AnRa, & Melton, 1997; Leibson et al., 1997). Blood samples are collected from healthy geriatric individuals and from those suffering from ADRD, in addition to blood factor analyses DNA and mRNA from white blood cells are collected. The NDDR is used to catalogue and analyze not only routine laboratory data, necessary for the documentation of diabetes and anemia, but the association of these patient values with established and suspected risk factors for ADRD as well. Thus, the NDDR provides a unique opportunity to analyze both clinical and research data by cohort or for multiple care episodes for a single patient over time. Alternatively, patterns of predictors and risk factors can be assessed across many patients for a particular diagnosis. For example, current research suggests that cardiovascular factors may play a part in the development of Alzheimer’s disease (de la Torre & Hachinski, 1997). Therefore, if in analyzing the database, specific cardiovascular risk factors are identified that associate with not merely cognitive decline but Alzheimer’s disease, hypotheses can be developed to test the specifics of such an association to include assaying for circulating regulatory factors or genetic markers ( from the banked DNA) that might segregate with Alzheimer’s disease.
This system, the first iteration of which focuses on the establishment of the database repository, represents to our knowledge the first application of data mining techniques to a data warehouse of patient-linked research and clinical data. This holistic approach will generate new methodologies that are likely to impact the future of health system delivery research.