| Abstract: | This thesis presents insights on integrating evidence-based medicine research methodology and core methods of clinical epidemiology into population-based disease informatics (PbDI) forming a novel research methodology to facilitate administrative healthcare database observational epidemiologic research.
Background: A group of interdisciplinary researchers that consist of medical informatics specialists, information technology specialists, and physician informaticians focus their work and electronic research on public health and population-wide disease informatics by using Taiwan’s national administrative claim database, commonly known as the National Health Insurance Research Database (NHIRD). PbDI is defined as the field of information science within the scope of biomedical informatics that deals with patient information at the level of an entire community or certain groups of a population whose treatment records, in the form of computable electronic data, are shared. The data are obtained through data-mining and/or other information science technology and analyzed to better understand the disease and improve its treatment outcome. PbDI in electronic research (e-research) is abbreviated as PbDIR in the dissertation. This thesis would like to recognize and promote population-wide disease informatics as a branch of knowledge in the rapidly evolving field of medical informatics.
Motivation of Research: 1. Population-based observational epidemiologic research involves physicians, clinicians, academicians, epidemiologists, information scientists and informaticians. 2. To establish a novel methodology in the population-based informatics research that incorporates evidence-based methods so that the research outcomes (manuscripts) contain high-quality results.
The Rationale for Integrating EBM and PbDIR: Scientifically, after a thorough literature review, the development of a novel methodological approach combing EBM and PbDIR using healthcare databases is required to ascertain a quality research outcome. Ethical-legally, we should responsibly conduct epidemiologic research.
Methods: PbDIR can be implemented by using a healthcare or claim database. After full literature search using MeSH terms and Boolean logics, an answerable research question can then be selected. Research-based PECOTS framework was introduced. After gaining an approval (mostly exemption from full review) from an IRB, PbDIR can be carried out. Evidence-based methods to incorporate into PbDIR include at least the followings: Search strategy, Study design (descriptive & analytic) (observational designs: cohort, case-control, hybrid), Calculation of risk, odds, and rate (eg., incidence rate), Minimizing bias, Matching (propensity score matching), Immortal time exclusion, Regression coefficient (ß1) interpretation. Data management involves utilization of coding book, data tables (flat files), data dictionary (meta-data), normalization (referential integrity) and validation rules for value, relational database, table-lookup function and complex programming may sometimes be needed. Linkage databases research such as merging data between NHIRD and National Death Registry or Taiwan Cancer Database (TCDB) shall be planned at the outset. Early self-application and self-evaluation with STROBE checklist shall be performed. The outcomes (endpoints) adopted specifically for this integration research were determined as follows: 1. Quality research outcome which is measured by the acceptance of the PbDIR manuscript submitted to a quarter 1 ranked biomedical journal or cited by a quarter 1 ranked journal. 2. Steady pace of manuscript completion and acceptance gained from the successful establishment of the integrated research methodology in PbDIR.
Results: Realization of the implementation of the incorporation of evidence-based medicine into PbDIR and the depth of this integration was demonstrated by three successful research projects that were association studies.
Conclusion and Recommendation: A novel research methodology in population-based disease informatics incorporating evidence-based methods can help population-based investigators to produce high-quality research outcome. For a multidisciplinary PbDIR collaborative team, adoption of this novel research methodology may further improve the evidence-based foundation of a quality research. |
