A Database is not what you think and different designs serve different purposes.
Download as PPT, PDF1 like483 views
Different types of longitudinal observational studies like databases, registries, and cohort studies serve different purposes. Registries relate to the pace of patient entry and events and are used to estimate disease incidence or identify adverse drug events. Cohort studies are set up to test hypotheses through periodic, pre-established data collection from a probabilistic sample. Administrative databases allow assessment of resource allocation and costs but have selection bias, while clinical databases monitor practice but lack representativeness. Different database types should be used appropriately based on their purpose and design.
A Database is not what you think and different designs serve different purposes.
- 1. Different designs of longitudinal observational studies (databases, regist ries , cohort studies) serve different purposes Loreto Carmona Research Unit Spanish Society of Rheumatology
- 2. Why this talk? Sometimes the only point in which investigators agree in a given project is about the need to collect data . T here is indeed much misunderstanding about databases. T he first principle of data collection : We should not collect data just for the sake of it, without explicit and specified objectives .
- 3. What is a database? NOT: Any set of data entered into a computer YES: A tool to organize the data in a study
- 4. What is not a database? A dataset
- 5. Other uses
- 6. What is a database? Organizing Hierarchical Relational
- 7. Structure patient treatment treatment treatment treatment AE AE AE AE AE AE
- 8. Structure patient visit visit visit visit outcome outcome outcome outcome outcome outcome
- 9. Time and databases Medical databases time Longitudinal Observational Studies
- 10. Can we analyse data directly from a DB? wide long
- 11. wide
- 12. long
- 13. Types of databases Depend on the purpose and design of the study they serve .
- 14. Types of databases Registries Cohorts Administrative databases Clinical databases
- 15. Registries The introduction of new elements is not pre-planned (no schedule). relate to the pace of patients entry events entry The purpose of the registry defines the types of registry: Disease registry Drug registry
- 16. Registries Disease registry : to estimate incidence of diseases pace of new entries = occurrence of new cases Drug registry : The main purpose of such registries is the identification of adverse events pace of entry = initiation of new treatments, of a target drug or therapeutic group, in patients who meet the inclusion criteria
- 17. Cohorts Studies set up for testing hypotheses . Sensu strictu aetiological hypotheses they can also be assembled to test prognosis, or to test resource use hypothesis Subjects in a cohort must be sampled in a probabilistic way ( to be representative ) . Timing of data entry pre-established in visits or examinations at a given interval ( periodical ).
- 18. Cohorts (concepts) Inception cohort all new cases or early Nested case-control studies advantage: cases and controls selected equally
- 19. Cohorts (concepts) Open cohorts Permanent incorporation allow the study of the simultaneous influence of calendar time, age (duration) and cohort (onset) in demography, epidemiology, and clinical follow-up
- 21. Cohorts (concepts) Closed cohorts specific research objective ( sample size ) recruitment period subjects become older with follow-up Permanent surveys examination over time is established in repeated cross-sectional surveys large open cohorts which are monitored over such a long time ( Framingham study )
- 22. Cohorts (concepts) Randomised controlled trials cohorts in which the hypothesis is the efficacy of an intervention Not a LOS ( planned intervention )
- 23. Cohorts (concepts) Aren’t prognosis or resource use studies cohorts? OK if the sample is selected in a probabilistic way and the hypothesis is pre-established before the launch of the cohort : link between exposure to prognostic factors (i.e. determinants) and rate of occurrence of outcome link between determinants and rate of use or cost
- 24. Types of databases Registries Cohorts Administrative databases Clinical databases
- 25. Administrative databases S pecific purpose : managing the economic and organization aspects They allow to assess and preview : the need of workforce allocation of resources logistic needs insurance monitoring costs, etc.
- 26. Administrative databases as surrogates PROS rapid response (availability) large sets of population ( increases statistical power ) CONS severe selection bias
- 27. Administrative databases ’ best use To cross-check data in clinical databases : to confirm death , work status , drug consumption ...
- 28. Clinical databases Quite often physicians record patients characteristic and variables directly from their practice in an intent to monitor health care practices. The resulting clinical databases are useful to : retrieve patients’ reports Retrieve individual clinical histories assess the individual practice composition of a department or a clinic
- 29. Clinical databases ’ bias The patients included have not been selected in a probabilistic way , what precludes the use of formal hypothesis testing statistical instruments. Representativity D ifficult y to record all variables with the expected quality of a main outcome variable if one does not know ahead of time what the main outcome variable will be ( study protocol )
- 30. Clinical databases ’ best use To monitor practice T o identify patients with a given characteristic, and to use them to draw a random sample, better if multicentre, and to test prospectively a research hypothesis
- 31. Conclusions We should bear in mind the meaning and use of databases. Different types of databases serve very different purposes . A clear understanding of the different investigation designs sh ould avoid many of the databases we launch be just a lot of work and very little of science.