Relative risk of contacting GPs within 24 hours of exposure to a media message – previous control dataset.a
| Number of control windows | Symptom adjustment | Number of patients (%) with exposure during case window | Number of patients without exposure during case window | Number of patients (%) with exposure during control window | Number of patients without exposure during control windows | OR (95% CI)b |
|---|---|---|---|---|---|---|
| 1c | No | 36 (11) | 76 | 36 (11) | 76 | 1.0 (0.5 to 1.9) |
| 1c | Yes | 24 (7) | 52 | 21 (6) | 55 | 1.2 (0.5 to 2.6) |
| 5d | No | 36 (11) | 76 | 32 (10) | 80 | 1.5 (0.8 to 2.7) |
| 5d | Yes | 12 (4) | 20 | 7 (2) | 25 | 1.7 (0.7 to 4.1) |
↵a This was estimated using the exposure odds ratio; this is the odds ratio of having received a media message in the 24-hour case window before the contact with the GP compared with either: the 24-hour control windows immediately before the case window, or with all five possible 24-hour control windows in the week before the contact (the previous control dataset). Subjects were all 322 patients in the interview study. Symptom adjustment restricts data to patients with comparable symptom intensity in the case and control windows.
↵b These results are for the 112 exposed patients, out of 322 interviewed patients. The exposure odds ratios are Mantel–Haenszel statistics for sparse data.12,13
↵c 24-hour control windows immediately before the case window.
↵d 24-hour control windows in the week before the contact with the GP. OR = odds ratio.