CHANCES : History of coronary heart disease, stroke, and diabetes

Prepared by: Mark O'Doherty and Maria Hughes, UKCRC Centre of Excellence for Public Health, Queens University Belfast, United Kingdom

1. Introduction

The cohorts were asked to report on their individual data items in respect of a past history of CHD, stroke and diabetes among study participants at baseline, so that the types and quality of data available for use within CHANCES could be compared. An overview of the responses to the questionnaire, and also information we have collected from published manuscripts, personal communication and cohort websites can be found in attached Excel table .

2. History of CHD and stroke

The relevant variables are broadly divided into two types depending on whether the past history is based on clinically documented or self-reported data.

The majority of cohorts have used self-reported history of CHD and stroke from the baseline questionnaire/examination to identify prevalent cases within their cohorts. However, documented history has been reported by EPIC Elderly Umeå (MONICA register), RES (self-report subsequently validated by research physicians based on medical records), and Tromsø (linkage to the discharge diagnosis register at the University Hospital of North Norway, the only local hospital serving the Tromsø population). ESTHER has self-reported history of CHD and stroke available, but has not used the self-report to identify prevalent CHD. Only the physician reported history of CHD from the physician's health check-up report was used (available for 97.4% of participants). A physician validation questionnaire was used for incident cases of stroke only, with self-report used for prevalent stroke identification at baseline. Information on self-reported heart attack, angina or coronary artery disease, atrial fibrillation, coronary artery bypass or angioplasty, TIA (Transient Ischemic Attack) and stroke are available at recontact 2 (2004-2005) for the NIH_AARP cohort, based on retrospective date-of-onset assignment. A literature search (see excel table) for RES also found that baseline self-reported MIs and strokes were validated by research physicians, and a baseline ECG was analysed off-line (automated Modular ECG Analysis System (MEANS)) to identify further prevalent MIs. Some of the individual cohorts within MORGAM had documented prevalent cases of CHD and stroke (e.g. MONICA coronary event register or other population-based coronary event register, person's medical records, a hospital discharge register or other health information system). Even though it is generally accepted that a clinically documented history is more reliable than self-reported history, MORGAM included self-reported history because a documented history was available from only a few cohorts.

A history of angina pectoris is only available for a small number of cohorts e.g. EPIC Elderly Greece, HAPIEE, MORGAM (some individual cohorts), and NHS, but when based on self-report may be considered to be unreliable. A history of revascularization may also be a strong indicator of prevalent CHD, though only a few cohorts have reported on this also e.g. MORGAM (some individual cohorts) and NIH_AARP (recontact 2).

Self-reported and clinically documented data both provide estimates of the prevalence of specific diagnoses such as CHD and stroke. However, as indicated above, many of the cohorts within CHANCES relied upon self-reported history of these diseases. Obviously, the quality of information obtained through patient self-reports depends substantially on both the questions asked and the characteristics of the person completing the questionnaire. Nevertheless, some previous studies have found good agreement between self-reported and documented diagnoses of both MI and stroke (Haapanen et al, 1997; Okura et al, 2004). The implication for CHANCES is that both self-reported and documented data sources need to be considered, with the choice depending on the actual research question. Therefore, our proposed variables (see below) cover both self-reported and documented cases.

3. History of diabetes

All cohorts have a self-reported history of diabetes at baseline. These include ESTHER, EPIC Elderly Umeå, RES, HAPIEE, EPIC Elderly Greece, EPIC Elderly Denmark, EPIC Elderly Bilthoven, Tromsø, MORGAM, Zutphen, NHS, SENECA, SHARE and NIH_AARP. Comparison of questionnaire items finds that most cohorts ask whether a participant has a doctor diagnosis of diabetes (except SENECA and Tromsø). All cohorts except NIH_AARP have recorded current treatment for diabetes which is used to support diagnosis. RES has linked current treatment to a drugs registry (99% coverage). EPIC Elderly Umeå subjects' records are linked to a diabetes registry recording treatment details. ESTHER has a doctor validated treatment record (98% coverage), while other cohorts have self-reported treatment details. SENECA has a question on use of medicines, vitamins/minerals (tablets or injections) or tonics, with space for the interviewer to record the name/dose/timing, but this question was not directly associated with a question on diabetes. However, SENECA retrospectively analysed fasting glucose and insulin measures at baseline (13 years after collection) which doubled the number of prevalent diabetes cases at baseline (based on the American Diabetes Association (ADA) criteria) (Teuscher et al, 2001). All cohorts can distinguish between those subjects managed through lifestyle change from those on oral diabetic medication or insulin. Some MORGAM cohorts have included doctors’ "suspicions of diabetes and glucose intolerance" with some inconsistencies. For those that have reported levels, missing data can be high in some of the MORGAM cohorts e.g.>15% in Denmark Glostrup, FINRISK, Italy Rome and Italy Friuli. EPIC Elderly Umeå, which is linked geographically to diabetes and population registries, estimates that 10% of data are missing for their cohort members. ESTHER estimates 3% missing data.

ESTHER, EPIC Elderly Umeå, RES and SENECA have a documented history of diabetes that can be compared to their self-reported history of diabetes from questionnaire items. SENECA found that self-report under-estimated prevalent diabetes. On the other hand, ESTHER estimated that self-reports may over-estimate prevalent cases, with 50% of their self-reported cases failing to accord with the clinical definition of diabetes used by their doctor validation method. The literature in this area is mixed, with some studies reporting poor agreement (Newell et al, 1999), while others indicating good concordance (Haapanen et al, 1997; Okura et al, 2004; Margolis et al, 2008; Leikauf et al, 2009; Barber et al, 2010).

HAPIEE, ESTHER, RES and EPIC Elderly Umeå have measured glucose, oral glucose tolerance test (OGTT) and HbA1c at baseline. RES and EPIC Elderly Umeå are currently using these measurements to diagnose prevalent diabetes. RES use non-fasting plasma glucose levels based on ADA criteria to define prevalent diabetes. EPIC Elderly Umeå has defined prevalent diabetes at baseline using a combination of linkage to a diabetes registry, self-reported diabetes and medication and an OGTT. Inclusion on the registry requires fasting glucose and OGTT tests according to ADA criteria, with medication use and medical record confirmation of diabetes. ESTHER has baseline fasting glucose (coverage: 81.3%) and HbA1c (coverage: 98.3%) for nearly all participants. Using ADA definitions, these biomarkers, in addition to physician diagnosis and current diabetes treatment, identified a further 97 cases of prevalent diabetes in addition to 1,427 identified on the basis of doctor validation report. HAPIEE measured HbA1c and fasting glucose on 3000 Polish participants and of these, 3.8% (n=118) had HbA1c ≥6.5%. Among those without a diagnosis of diabetes, 1% had HbA1c ≥6.5% (n=30), and among those with a diagnosis of diabetes, 32% had Hb1Ac ≥6.5% (n=88).

4. Proposed harmonized variables

The implication for CHANCES is that both self-reported and documented data sources need to be considered, with the choice depending on the actual research question. We have proposed variables that cover both self-reported and documented cases individually and also combined. We have also proposed a combined variable to include either documented or self-reported MI or stroke at baseline. These variables should be widely available and reasonably reliable indicators of MI and stroke at baseline for being used when excluding baseline cases from analysis of incident events during follow-up and also for identifying baseline cases when these are used as study end-point. Combined variables including self-reported history was included because documented history is available from few cohorts only. Similar variables have been proposed for diabetes. However, a variable for treatment of diabetes has also been proposed, which can be used in conjunction with either the documented and/or self-reported diabetes variables.

References

Barber J, Muller S, Whitehurst T, et al. Measuring morbidity: self-report or health care records? Fam Pract. 2010;27:25-30.

Haapanen N, Miilunpalo S, Pasanen M, et al. Agreement between questionnaire data and medical records of chronic diseases in middle-aged and elderly Finnish men and women. Am J Epidemiol. 1997;145:762-9.

Leikauf J, Federman AD. Comparisons of self-reported and chart-identified chronic diseases in inner-city seniors. J Am Geriatr Soc. 2009;57:1219-25.

Margolis KL, Lihong Qi, Brzyski R, et al. Validity of diabetes self-reports in the Women's Health Initiative: comparison with medication inventories and fasting glucose measurements. Clin Trials. 2008;5:240-7.

Newell SA, Girgis A, Sanson-Fisher RW, et al. The accuracy of self-reported health behaviors and risk factors relating to cancer and cardiovascular disease in the general population: a critical review. Am J Prev Med. 1999;17:211-29.

Okura Y, Urban LH, Mahoney DW, et al. Agreement between self-report questionnaires and medical record data was substantial for diabetes, hypertension, myocardial infarction and stroke but not for heart failure. J Clin Epidemiol. 2004;57:1096-103.

Teuscher AU, Reinli K, Teuscher A, et al. Glycaemia and insulinaemia in elderly European subjects (70-75 years). Diabet Med. 2001;18:150-3.

History CHD-Stroke-diabetes.xlsx (application/octet-stream)

History CHD-Stroke-diabetes.xlsx (application/vnd.openxmlformats-officedocument.spreadsheetml.sheet)

1. Introduction

2. History of CHD and stroke

3. History of diabetes

4. Proposed harmonized variables

References

Attachments: