Advertisement
Table 4 Errors in negative Interphone studies [4, 36–55, 65, 72], and reliability of positive Hardell studies [1–3, 7–10, 64, 71, 78].
| study, design, methods | negative studies | positive studies |
|---|---|---|
| Mobile phone use | inadequate: 2-5 min/day | significant: 16-32 min/day |
| Latency time | <5% cases with latency ≥10 y | >18% cases with latency ≥10 y |
| Cordless phone users | considered unexposed | considered exposed |
| Ipsilateral tumour latency | ≥10 y for only 2% cases | ≥10 y for >16% cases |
| Head tumours identified | only gliomas, meningiomas, neuromas, parotid tumours | also other head tumours types |
| Deceased cases | not included | included: proxy interviews |
| Interviews | not blind | always double blind |
| Type of interviews | face-to-face | mailed questionnaires |
| Time of interviews | cases: during hospitalisation | cases: after hospitalisation |
| controls: at home | controls: at home | |
| Exposure assessment | non blind interview | blind questionnaire |
| Data processing | not stated (not blind?) | Blind |
| Laterality attribution bias | present | Absent |
| Delayed interviews | for controls compared to cases | not delayed |
| Participation | reduced up to 40% | always near to 90% |
| Selection | exposed controls prevail | no selection bias |
| Documentation | positive data ignored | no documentation bias |
| Funding | co-funded by MP Companies | funded only by Public Bodies |
