Dementia ascertainment using existing data in UK longitudinal and cohort studies: a systematic review of methodology

Abstract

BackgroundStudies investigating the risk factors for or causation of dementia must consider subjects prior to disease onset. To overcome the limitations of prospective studies and self-reported recall of information, the use of existing data is key. This review provides a narrative account of dementia ascertainment methods using sources of existing data.MethodsThe literature search was performed using: MEDLINE, EMBASE, PsychInfo and Web of Science. Included articles reported a UK-based study of dementia in which cases were ascertained using existing data. Existing data included that which was routinely collected and that which was collected for previous research. After removing duplicates, abstracts were screened and the remaining articles were included for full-text review. A quality tool was used to evaluate the description of the ascertainment methodology.ResultsOf the 3545 abstracts screened, 360 articles were selected for full-text review. 47 articles were included for final consideration. Data sources for ascertainment included: death records, national datasets, research databases and hospital records among others. 36 articles used existing data alone for ascertainment, of which 27 used only a single data source. The most frequently used source was a research database. Quality scores ranged from 7/16 to 16/16. Quality scores were better for articles with dementia ascertainment as an outcome. Some papers performed validation studies of dementia ascertainment and most indicated that observed rates of dementia were lower than expected.ConclusionsWe identified a lack of consistency in dementia ascertainment methodology using existing data. With no data source identified as a “gold-standard”, we suggest the use of multiple sources. Where possible, studies should access records with evidence to confirm the diagnosis. Studies should also calculate the dementia ascertainment rate for the population being studied to enable a comparison with an expected rate.Electronic supplementary materialThe online version of this article (doi:10.1186/s12888-017-1401-4) contains supplementary material, which is available to authorized users.

GWAS on family history of Alzheimer’s disease

“…One difficulty in traditional studies of AD is case ascertainment 4 —either directly for prevalent cases or indirectly through prospective cohort studies for incident cases. A recent GWAS study on a subset of the UK Biobank cohort used information from family history (parent or first-degree relative with AD or dementia) as a proxy-phenotype for the participants 5 .…”

Ascertainment of Dementia Cases in the Spanish European Prospective Investigation into Cancer and Nutrition-Murcia Cohort

“…The same method was used to identify other neurological diseases, such as migraine [35] or stroke [36]. However, so far there is no "gold standard" methodology to ascertain the incidence of dementia using existing data in cohort studies [37].…”

Association of Lifestyle and Genetic Risk With Incidence of Dementia

“…Fourth, lifestyle factors were self-reported and some cases of dementia are not recorded in medical records or death registers. 34 However, misclassification errors are likely to have biased these findings toward the null. Additional variants associated with dementia are likely to be identified in future genome-wide association studies, and these variants may prove useful in further refining estimates of genetic risk.…”