The family history has long been recognized as a risk assessment tool in primary care.13
As new genetic tests emerge, it has particular importance as a means of identifying people with increased risk for genetic disorders or susceptibilities. People at increased risk because of their family history may be among the most appropriate candidates for genetic tests. A traditional screening test is defined by both positive predictive value—the chance of having the disease under consideration if the test result is positive—and negative predictive value—the chance of not having the disease if the test is negative. Predictive value is a function of the test characteristics (sensitivity and specificity) and most importantly of the patient's “pretest probability” of disease, that is, the patient's personal risk of having the illness being screened. As shown in , if patients being tested have a low risk of the disease (e.g., 1/1,000), they are likely to be healthy even if their screening test is positive. Thus, the positive predictive value (chance that the screened individual actually has the disease) is only 9/100 when using an unusually accurate test with 99% sensitivity and 99% specificity (far better diagnostic accuracy than typically applicable to most screening tests in use today). In choosing screening tests, the background or population risk reflects the proportion of individuals in the general population affected with a particular disorder,7
but the patient's personal risk may be higher given his or her family history or other circumstances. Therefore, family history can be crucial information increasing the individual patient's pretest probability, and therefore the predictive value of genetic or other screening tests.
Prior Probability and Predictive Value
The role of family history in the predictive value for screening can be illustrated with colorectal cancer. The overall population risk is approximately 1/1,000 for a 45-year-old. Eight percent of the population have a first-degree relative with a history of colorectal cancer that confers a personal risk of colorectal cancer 2-fold higher than the general population14
(), and individuals with a positive family history also present earlier with colorectal cancer.14
As a result, the estimated risk at age 40 is similar to that of an average risk person at age 50 (). Assuming that the sensitivity and specificity of the screening tests for colorectal cancer (e.g., fecal occult blood testing, colonoscopy) are the same for people in their 40s as for older patients, an argument can be made for initiating colorectal cancer screening at age 40 for patients with a positive family history.15,16
Thus, even before consideration of genetic testing, the family history has been important in determining the appropriate use of traditional screening tests.
FIGURE 1 Colorectal CA risk ± family history. From Fuchs et al.14
Predictive use of genetic testing may be defined as testing offered to asymptomatic individuals in order to determine their risk to develop the disorder themselves.7
Determining the predictive value of these genetic tests involves the same mathematical principles as traditional screening tests. Note, however, that the predictive value of a genetic test may be subject to 2 different interpretations. It could be considered the likelihood that an individual with a positive result actually has the gene mutation in question. Alternatively, predictive value can be interpreted as the chance that the individual with a positive genetic test result will develop the associated disease. Medical geneticists address this dual meaning through the concepts of “analytic validity” and “clinical validity.” Analytic validity refers to the accuracy with which the test identifies the mutation, whereas clinical validity refers to the accuracy with which the test result predicts clinical outcome (which is a function of gene penetrance).17
Given the incomplete penetrance of many mutations, the clinical validity of many genetic tests is lower than the analytic validity.17
These concepts apply to family history as well as to molecular genetic tests; thus, the analytic validity of family history as a risk predictor may be limited by the patient's knowledge about family members; and the clinical validity may depend on such factors as number of affected relatives and age of onset of disease. For many diseases, there are insufficient data to provide precise estimate of the analytic and clinical validity of a positive family history, although it is a well-established risk factor for most common diseases.18
Family history plays an important role in identifying patients who may benefit from predictive genetic testing. Familial adenomatous polyposis (FAP), in which innumerable colon polyps develop at an early age, offers an excellent example. The population or background risk of FAP is 1/8,000; individuals who inherit a mutation in the APC gene causing FAP are estimated to have a risk of colorectal cancer that approaches 100% (see ). Conversely, the background population's lifetime risk of colorectal cancer is 5/100. A positive family history of FAP dramatically influences the patient's personal lifetime risk of colorectal cancer. If a parent has FAP, the risk of the individual patient having an APC mutation associated with FAP is 50%. Penetrance is estimated to approach 100%, so lifetime risk is close to 50%. Current sensitivity of full gene sequencing analysis for predicting FAP is approximately 95%. Taken together, these facts suggest that predictive testing for FAP can identify individuals with this condition—that is, can provide a genetic diagnosis of FAP. Preventive treatment, in the form of a subtotal colectomy, can be offered to reduce the risk of developing colorectal cancer.
This predictive testing scenario must, however, take into account limitations in test specificity. Because of the precision of molecular testing, the analytic validity of genetic tests is estimated to be very high, although for most genetic conditions, the general population studies have not been performed to determine the actual rate of false positive test results (i.e., specificity). If we assume, for the purpose of this example, that the accumulated random or procedural errors related to genetic testing for FAP yield a potential specificity of 99.9% (i.e., a false positive rate of 1/1,000), genetic testing for FAP in the baseline population (where risk of FAP is 1/8,000) would result in a positive predictive value of 11% (). That is, despite the high sensitivity and specificity of FAP testing, and the high penetrance of FAP mutations, most of the positive test results occurring with FAP testing of the general population would represent false positives. In contrast, if the same genetic test were conducted on a patient with a positive family history of FAP, the predictive value of the positive test (i.e., the patient's risk of contracting FAP-related colorectal cancer) would be 99.9% (). Therefore, the individual patient's family history remains a critical factor in determining the value of predictive genetic testing.