Family history of breast cancer and cost of life assurance: a test case comparison of current UK industry practice

Introduction

Under the current recently extended moratorium, applicants for life assurance need not disclose the results of predictive genetic tests.1 The exception is for policies exceeding certain values, when insurers may seek results of tests approved by the government's Genetics and Insurance Committee. The committee expects to receive applications for the use of adverse results from tests for the BRCA1 and BRCA2 genes. Currently, insurers may, and often do, seek family histories. Substantial epidemiological data describe the relative risks of developing breast cancer depending on family history and age.2 Preventive action for women at risk can include early enrolment on surveillance programmes and prophylactic surgery. Evidence is emerging for the benefits of the former,3 and evidence is strong for risk reduction by the latter.4 5

Participants, methods, and results

We surveyed 21 companies representing 100% of the reinsurance market and 68% of the life and pensions market in the United Kingdom. We asked the companies to assess a fictional proposal for a 20 year policy (paying benefit only on death) by applying an excess mortality rating, defined as the percentage increase over the assumed rate of mortality. The applicant (scenario 1) was a 35 year old woman with unremarkable personal and family histories, except for breast cancer in the mother diagnosed at age 35. In line with standard insurance application forms, only first degree family history was given. In scenario 2, the same applicant had enrolled on a mammographic surveillance programme with no adverse results reported. In scenario 3, the same applicant had undergone prophylactic double mastectomy and oophorectomy (figure).

Percentage excess mortality ratings applied by life insurance and reinsurance companies in 2004-5. All companies gave one rating for the full product term, except for one that did not specify the rating or duration and is not shown

• Download figure
• Open in new tab
• Download powerpoint

Sixteen companies responded. The responding and non-responding groups included a similar variety of company profiles (for example, global operations and size). Nine would not increase premiums under any of the scenarios. Six would increase premiums under scenario 1 (rating +50% or +75%), and of these, four would not load under scenario 3, and two would not load under either scenario 2 or 3. One small UK company would not raise premiums under scenario 1 or 2, but would apply an unspecified increase under scenario 3.

Comment

Most life insurance companies that responded to the survey would offer a standard premium in scenario 1 (unremarkable except breast cancer in the mother diagnosed at age 35). The lifetime relative risk of developing breast cancer for the applicant in scenario 1 is 5.7 (2.7 to 11.8).2 Using life tables developed at the Cambridge Genetics Knowledge Park (A Butterworth, personal communication, 2005) this translates into a 20 year breast cancer mortality risk (from age 35) of 1.95%, compared with 0.6% for the general population–that is, three times higher.

It is reassuring that not only did all companies that rated under scenario 1 remove the rating after prophylactic surgery, but that two companies also gave credit for enrolment on a mammographic surveillance programme. Another company reported that this will be part of a review of policy, and in future it may also credit mammographic surveillance. The variation in the industry's responses show that women with a family history who are offered a high rating by one company may receive a lower rating if they approach a second company.

No respondent said that they would request further family history to see if second degree relatives were affected. Disclosure of more cases of cancer in second degree relatives could considerably increase the assessed risk. The artificial nature of the survey might have implied, however, that no further family history existed to be disclosed, so we cannot comment on whether second degree relative information would be requested in reality. How access to adverse test results for BRCA1 or BRCA2 would influence the response to a similar survey in the future remains to be seen.