Equation used to calculate estimate (E) | G1 (80 families)* | G2 (106 families)* | G3 (101 families)* | ||||||
---|---|---|---|---|---|---|---|---|---|
O | E | p | O | E | p | O | E | p | |
*Some families could not be included because of unreliable data on ages. | |||||||||
On average, there were six observed affected individuals in generation 1 (G1) from every 10 families, 15 from G2, and five from G3. Two classes of risk were estimated (low and high) based on age and probability of being a carrier (Pr). This was used to estimate the number of affected individuals in each generation (E) E was compared with the observed number of affected individuals per family (O). Two sided p values of less than 0.01 were taken to indicate that the estimated values were not consistent with observed values. A risk calculation giving expected cases equivalent with observed values in G2 would overestimate expected values in G1 and underestimate in G3. Clearly, a still greater risk could be assumed that would give a better estimate of G3 cancer incidence. This implies that age related risk increased with each generation. | |||||||||
Low risk | 0.65 | 0.374 | 1.52 | 0.82 | <10−6 | 0.5 | 0.09 | <10−6 | |
E = ∑Pr (1−((age×0.03)−1.2) | |||||||||
(for age = 55–90) + ∑Pr (for over 90) | |||||||||
High risk | 0.63 | 1.14 | 0.003 | 1.52 | 1.32 | 0.417 | 0.5 | 0.13 | 0.028 |
E = ∑Pr (1–eln (age)×5.35–23) | |||||||||
(for age = 40–75) + ∑Pr (for over 75) |