Table 1.
Summary of characteristics of studies included in the review (stratified by setting, database, and dementia type)
| Study | Country | Setting | Study Design | Study Period | Database | Coding System | Characteristics of Population to Which Index Test Was Applied | Diagnosis Validated | Validation Method | ||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Description | Age | Sex | |||||||||
| Walker, 201816 | England | Primary care | Prescribing trend analysis | 1 January 1987‐31 December 2016 | CPRD | READ | Patients with at least 12 consecutive months of records classified as “acceptable” by the CPRD from an “up to standard” practice. | >40 | M/F | AD + dementia | External database comparison (HES + ONS) |
| Whitelaw, 199617 | Scotland | Primary care | Validation | April 1992 | GPASS | READ | Random sample of 250 patients from 41 practices (those having more than 50% of patients with a clinical Read code) that volunteered to take part in the study | 45‐64 | M/F | Dementia | Case note review |
| Seshadri, 200118 | UK | Primary care | Population‐based nested case‐control | 1 January 1990‐31 October 1998 | GPRD | READ | Women in the population who had received ≥1 prescription for a systemic (oral or transdermal) oestrogen preparation during study period. | >40 | F | AD | Case note review |
| Imfeld, 201219 | UK | Primary care | Population‐based nested case‐control | January 1998‐November 2008 | GPRD | READ | Patients with a diagnosis of AD during study period | >65 | M/F | AD | GP confirmationb |
| Dunn, 200520 (a)a | UK | Primary care | Population‐based nested case‐control | 1 January 1992‐1 January 2002 | GPRD | READ | Patients included in GPRD during study period | >60 | M/F | Dementia | GP confirmationb |
| Dunn, 200521 (b)a | UK | Primary care | Population‐based nested case‐control | 1 January 1992‐1 January 2002 | GPRD | READ | Patients included in GPRD during study period | >60 | M/F | Dementia | GP confirmationb |
| Imfeld, 201322 | UK | Primary care | Population‐based nested case‐control | January 1998‐November 2009 | GPRD | READ | Patients with a diagnosis of VaD during study period | >65 | M/F | VaD | GP confirmationb |
| Heath, 201523 | Scotland | Primary care | Population‐based cross‐sectional study | March 2007 | SPICE | READ | Patients with code for dementia at time of study from 314 general practices in Scotland that agreed to take part in the study. | 40‐64 | M/F | Dementia | Case note review |
| Blak, 201124 | UK | Primary care | Validation | 2006‐2007 | THIN | READ | Patients included in THIN | All | M/F | Dementia | External rate comparison |
| Van Staa, 199425 | UK | Primary care | Validation | 30 April 1990‐31 July 1992 | VAMP | ICD‐9 | All persons who received during study period at least one prescription for glibenclamide, gliclazide, chlorpropamide, glipizide, or tolbutamide. | >20 | M/F | Dementia | GP confirmation and review of discharge letters |
| Sommerlad, 201826 | England | Hospitalisation | Validation | 1 January 2008‐1 March 2016 | HES | ICD‐10 | Patients with a diagnosis of AD during study period | >65 | M/F | Dementia | External database comparison (CRIS) |
| Brown, 201627 | England | Hospitalisation | Validation | 1 April 1997‐.31 April 2011 | HES | ICD‐10 | Women recruited from England into the million women study | 55‐79 | F | Dementia | External database comparison (CPRD) + GP confirmation and review of discharge letters |
| Soo, 201428 | Scotland | Hospitalisation | Validation | 1 January 2003‐1 June 2003 | SMR01 | ICD‐10 | Patients included in the GLOMMS‐I cohort (ie, identified as having moderate/severe chronic kidney disease during study period). | >15 | M/F | Dementia | Case note review |
| Ryan, 199429 | Scotland | Hospitalisation | Validation | 1968‐1987 | SMR01/SMR04 | ICD‐9c | Admissions to Scottish hospitals in the Lothian region during study period | N/A | M/F | Dementia | Case note review |
Abbreviations: AD, Alzheimer's disease; CPRD, Clinical Practice Research Datalink; GLOMMS‐I, Grampian Laboratory Outcomes Mortality and Morbidity Study‐1; GP, general practitioner; GPASS, General Practice Administration System for Scotland; GPRD, General Practice Research Database; HES, Hospital Episode Statistics; ICD‐9, International Classification of Disease (9th Edition); ICD‐10, International Classification of Disease (10th Edition); QOF, Quality Outcome Framework; SMR01, Scottish Morbidity Records (General); SMR04, Scottish Morbidity Records (Psychiatric); THIN, The Health Improvement Network; VaD, Vascular Dementia; VAMP, Value Added Medical Products.
Separate studies were performed by same author in same year. Study (a) examined whether lithium therapy could be used to prevent the onset of dementia, while study (b) examined the association of dementia with infection. While many aspects of the two studies appear quite similar, the discrepancy between the number of patients validated in each (Table 3) indicates that they should be treated as separate studies.
For these studies, the method of GP confirmation is poorly described. Each study reported sending GPs a questionnaire to confirm the diagnosis of dementia (or its subtypes) but provided no details on the content of the questionnaire and did not mention requesting supporting materials such as hospital discharge letters.
Harmonisation between ICD‐8 and ICD‐9 occurred for the period 1968 to 1979.