General practitioner notes as a source of information for case-control studies in young women. UK National Case-Control Study Group.

Abstract

GP notes in case-control studies of young women the diagnosis/pseudodiagnosis date were ignored, even if they referred to events before that date. The notes were abstracted onto a structured form and interviewers were instructed to use all information (including correspondence) in the notes. Where an item in the notes was undated, the abstractor used her judgment as to whether it happened before or after diagnosis/pseudodiagnosis date by using the chronology of the notes. Other sources of information kept by the GP (for example the notes of a woman's husband or children) were not abstracted.We wanted complete GP data on all eligible cases and their first selected controls, regardless of whether or not the interviews had been done. Where we were unable to interview a case, for whatever reason, permission was sought from the consultant and GP to abstract the information from her GP notes. Permission was sought from the GP, and if required by the GP, from the control also, to abstract the same information from her GP notes. If a first-selected control could not be interviewed for any reason, permission was also sought to abstract information from her GP notes even though a second (or subsequent) control had been successfully interviewed. Information was also sought from any family planning clinics that a woman recalled attending. The data from all sources were used to contruct the lifelong contraceptive calendar used for the main analyses of the study data.1 2 STATISTICAL METHODSAgreement between the interview and the GP data was measured by means of the kappa statistic.25 The analyses of breast cancer risk using interview data alone and GP data alone were carried out using multivariate logistic regression methods for individually matched case-control studies.6 Relative risks were estimated by odds ratios. Significance levels quoted are two-sided. Of the 755 first-selected controls, 675 (89%) were interviewed and the GP notes of 674 of these were abstracted. The 80 first controls not interviewed were replaced by second or subsequent choices and GP notes were abstracted for these also. Fifty three (66%) of the 80 firstselected controls who were not interviewed also had their GP notes abstracted. Reasons for not abstracting the remaining 27 sets of notes were GP refusal (15), control refusal (11), and failure to trace (1). We also abstracted 219 sets (99%) of notes for firstselected controls matched to the 222 noninterviewed cases whose notes had been successfully abstracted. (Subsequent controls were substituted for the three remaining control women who refused permission for note abstraction.)We compared information obtained from interview with that obtained from GP notes for the 754 pairs where both sets were available; one GP refused permission for abstraction. GP notes were abstracted for a total of 946 (90%)

Tubal ligation and risk of ovarian cancer subtypes: a pooled analysis of case-control studies

“…A second limitation was the potential for recall bias. However, a woman's self-reported tubal ligation history has been shown to be highly accurate compared with medical records, 32 and recall bias likely had minimal impact on the study findings. Finally, we cannot exclude the possibility that unmeasured confounders that act differently among disease subtypes could contribute to the observed subtypespecific differences.…”

Hodgkin's lymphoma and infection: findings from a UK case–control study

“…Although these data, which are principally collected with the aim of documenting and monitoring patient care, have been used in a limited way in a number of aetiological studies (McKinney et al, 1991;Mann et al, 1993;Chilvers et al, 1994;Ansell et al, 2005), their potential with respect to describing symptom profiles is yet to be fully realised. A critical advantage for aetiological and other studies -where the sequence and timing of events is important -is that information held in GP records is collected prior to the diagnosis of malignancy and so have the advantage of being unaffected by recall and reporting bias, having been recorded by the GP contemporaneously.…”

Tubal ligation and ovarian cancer risk in African American women

“…As with all retrospective case–control studies, there is the potential for recall bias. Overall, the literature shows good agreement between self-reported history of tubal ligation and medical record data [55], but we cannot rule out selective reporting in our study. There is also the potential for selection bias due to both the nature of random-digit dial surveys and the high fatality rate of EOC.…”