The accuracy of cancer diagnoses as reported in families with head and neck cancer: a case-control study.

AIMS

The assessment of the risk of developing cancer is largely based on family history. This study aimed to confirm as many cancer diagnoses as possible in families of patients with second primary tumours after index squamous cell carcinoma of the head, single squamous cell carcinoma and spouse controls.

MATERIALS AND METHODS

Eighty patients with second primary tumours, 67 with single squamous cell carcinoma and 47 spouse controls were recruited into the study. A detailed interview was undertaken on all study participants, including information on all first-degree relatives for causes of death and possible diagnoses of cancer.

RESULTS

Details were available from 1340 first-degree relatives: second primary tumour group 570 relatives, single squamous cell carcinoma 429 relatives and the spouse controls 341 relatives. There were 174 cases of cancer (76, 56 and 42, respectively). For those relatives that had died, submissions were made to the Office of National Statistics to confirm the cause and date of death. One hundred and thirty-three cases were submitted for Office of National Statistics flagging (58, 42 and 33, respectively). Seventy-four of the first-degree relatives with cancer were successfully traced. Tracing confirmed a cancer diagnosis but no further details of the primary site in 31 of the relatives. Thirty-eight of the first-degree relatives were confirmed to have the cancer site of diagnosis reported by the study participants. Thirteen of the reported diagnoses in the first-degree relatives were inaccurate with regard to the cancer site of origin. Five of the first-degree relatives did not have a cancer diagnosis confirmed on the death certificate.

CONCLUSIONS

This study highlights the difficulties encountered in confirming cancer diagnoses among first-degree relatives within this cancer site. To improve such studies and our routine collection of data within the clinic setting, more robust systems are needed to interlink cancer registries, population data and hospital records.