The validity of dementia diagnoses in routinely collected electronic health records in the United Kingdom: A systematic review.

While heterogenous across studies, reported validity estimates were generally high. However, the credibility of these estimates is limited by the methodological quality of studies, primarily resulting from insufficient blinding of researchers interpreting the reference test. Inadequate reporting, particularly of the specific codes validated, hindered comparison of estimates across studies. Future validation studies should make use of more robust reference tests, follow established reporting guidelines, and calculate all measures of validity.