We wish to draw the attention of the Committee to problems Cancer Registries are currently experiencing with the issue of confidentiality and consent. These problems appear to have been especially highlighted since the recent implementation (1 April 2000) of the 1998 Data Protection Act.

  Cancer registries are the sole sources of population-based cancer information in this country. The information collected and analysed by cancer registries is essential in the planning of patient services for cancer, in monitoring the effective delivery of such services and in epidemiological research. Without the data provided by cancer registries it would, for example, be impossible to monitor changes in cancer survival at the population level and very difficult to investigate issues such as whether social deprivation affects the incidence, treatment or outcomes from cancer; whether national screening programmes are improving cancer outcomes; or whether sources of environmental pollution are increasing cancer rates.

  All cancers (except some skin cancers) are registered, the information coming from a variety of NHS sources and from death certificates. Information from NHS records has been used as a means of registration on the basis of implied consent. The assumption has been that registration forms part of the auditing of cancer care and if a person does not want data from their records to be registered they will let it be down. Up to now the Data Protection Registrar has accepted the registries method of obtaining data, albeit with increasing reluctance. Registries have made patient data available to bonafide research organisations but only after the managing consultant and, usually, the patients' GP have given their consent to the data release and advice has been sought from the appropriate research ethics committee. Patients are asked for their informed consent prior to entering research studies. The cancer registries have an unblemished record in preserving patient confidentiality over many years.

  It has been suggested that explicit consent, where the patient is asked for permission to record and use the data for purposes other than treatment, should now be introduced for cancer registration. Cancer registries are aware that this method of working would result in incomplete registration, which would make the cancer registration system useless. Non registration would occasionally occur because the patient refused consent, but far more commonly because the doctor (or his/her staff) who would have to ask the patient, would forget or be reluctant to ask for a number of reasons. Recent German experience indicates that underreporting would be on a large scale. But even if the number of missing registrations were quite small, the figures would be difficult to interpret because it would not be known how many cases were missing and because the missing cases would be systematically different from the registered ones thereby introducing bias.

  We believe that explicit consent is not a feasible option if the population based coverage of cancer registries is to be maintained. The hypothetical additional safeguards introduced by explicit consent are likely to be negligible in comparison with the potential loss to the whole community and to future cancer patients if the population basis to registration becomes compromised.

  Given that the present system of implied consent is becoming unacceptable in some quarters, statutory notification of cancer (as practised, for example in Scandinavia) may be the best, or even the only, remaining option. This requires detailed consideration, from both practical and legal perspectives, as a priority before the long-term continuity of cancer registry information is threatened. We are aware that any legislation would need to be carefully drafted so as not to be too restrictive in relation to using the data for research purposes.

June 2000