Results

The age adjusted mortality rate for CHD in the 1477 fathers was 11.0 per 1000 person years compared with 12.0 per 1000 person years in 4779 other married men who took part in the original MIDSPAN study (who either did not have offspring in the study or who took part in the original study without their spouse). For mothers and 4642 other married women, the rates were 4.6 and 5.7 per 1000 person years respectively. In 18.4% of families both parents had died of any cause by 1996, compared with 24.5% of other couples who had participated in the original study.

Twenty nine offspring from 20 families reported themselves to be step-children or adopted and were excluded from further analyses. The study population comprised 1026 sons and 1283 daughters from 1457 families, who were aged 7–38 years when their parents took part in the original MIDSPAN study and 30–59 years when surveyed in 1996.

PREVALENCE OF PARENTAL DEATHS FROM HEART DISEASE

Death certificates showed that 379 fathers (26.0%) died of “heart disease” including 286 (19.6%) with CHD as the underlying cause of death; for mothers the figures were 180 (12.4%) and 116 (8.0%), respectively.

CHD deaths occurred at a wide range of ages, from 51–86 years in fathers and from 52–83 years in mothers. At the time of bereavement sons were aged 12–58 years and daughters 9–55 years (table 1). Seventeen sons and 23 daughters had two parents who had died from CHD.

The prevalence of a parental CHD death was 26% in both sons and daughters (table 2). Prevalence was higher in offspring in manual compared with non-manual occupations (OR=1.37, 95% CI 1.12, 1.68, p=0.0021) and in older offspring (for a 10 year increase in age, OR =1.18, 1.01, 1.37, p=0.034).

REPORTING OF PARENTAL DEATHS

Of the 647 sons with at least one deceased parent, according to death certificate records, 643 (99%) correctly identified the vital status of both parents; of 823 daughters with at least one deceased parent, 812 (99%) confirmed this information. Of the 379 sons whose parents were alive in 1996 according to the GRO, 367 (97%) reported them both to be alive; of 460 daughters, 448 (97%) reported that both parents were alive.

REPORTING OF A FAMILY HISTORY OF HEART DISEASE

Altogether 89.1% of parental deaths from CHD were reported correctly by offspring. Older offspring reported this cause of death less sensitively, but the association did not reach statistical significance: for a 10 year increase in age, OR =0.68 (0.45, 1.02), p=0.065. Daughters' reporting was more sensitive, again without reaching significance: OR=1.59, (0.95, 2.65), p=0.078. The effect of offspring social class was not significant: for manual relative to non-manual occupations, OR=0.81 (0.48, 1.36), p=0.42. The specificity of offspring reports of parental CHD deaths was 86.3%.

PERCEPTION OF A FAMILY WEAKNESS ATTRIBUTABLE TO HEART DISEASE

Heart disease was the most commonly perceived type of condition, weakness or illness running in a respondent's family. In the whole population 12.8% of sons and 18.7% of daughters considered that they had a family weakness attributable to heart disease. In both sons and daughters 47% of such cases were associated with at least one parental death from CHD (table 3). Some 22.9% of sons (202 of 881) and 20.7% of daughters (212 of 1025) who did not perceive a family weakness attributable to heart disease had at least one parent who had died of CHD.

Of those with at least one parental CHD death 23.2% of sons and 34.4% of daughters considered themselves to have a family weakness attributable to heart disease (table 3), compared with 9.1% of sons and 13.2% of daughters in families where neither parent had died from CHD. There was no difference between the proportions of maternal CHD deaths (24.3%, 43 of 177) and paternal CHD deaths (28.8%, 123 of 427) that were associated with a perceived family weakness attributable to heart disease.

The perception of a family weakness was lower in older offspring (for a 10 year increase in age, OR =0.82 (0.68, 0.98), p=0.034), higher in daughters (OR=1.56 (1.22, 1.98), p=0.0004) and slightly lower in offspring from manual occupations (OR = 0.80 (0.61, 1.04), p=0.089) (table 4). Relative to offspring whose parents had not died from CHD, those with a single parental CHD death had an odds ratio of 3.15 (2.47, 4.02, p<0.0001); for those whose parents had both died from CHD the prevalence of a perceived family weakness was again increased (OR=7.91 (4.12, 15.170, p<0.0001). In the statistical model confined to sons, perceptions of a family weakness attributable to heart disease were significantly less common in men from manual compared with non-manual occupations (OR=0.66 (0.44, 0.98), p=0.038) (table 4).

Discussion

Despite increasing interest in the recording of family histories of heart disease in general practice, there have been few studies linking the prevalence, reporting and perceptions of such histories in the general population. This is the first such study based on a large general population cohort in which parental mortality has been monitored over a 20 year period. Although it was more difficult to renew contact with families where both parents had died, the study population includes a substantial number of such families. The study excludes offspring who had left the area.

Over the 20 year period 26% of fathers and 12% of mothers had died of “heart disease”. Seventy five per cent of these paternal deaths and 64% of the maternal deaths had CHD as the underlying cause of death on their death certificate. The fact of death was reported correctly in 99% of cases and the cause of death was reported correctly in 89%, of deaths attributable to CHD.

Epidemiological definitions of a family history of heart disease are based on the fact of death and the associated disease risks.5-7 Lay perceptions of a family history, however, depend in addition on the meaning and implications of deaths and clinical events as interpreted by individuals.10 11

Our study of lay perceptions was based on offspring reporting “a family weakness due to heart disease”. This definition has some face validity, deriving from its strong association with parental deaths from CHD, the greater effect of two parental deaths and the greater effect of parental deaths from CHD on younger offspring.

A strength of the study is the high level of agreement between respondents' and the GRO's reports of parental deaths from heart disease. A limitation is that we have no data on non-fatal coronary events in family members12 and on fatal events in family members other than parents and siblings (although only 10 sons and 8 daughters from 13 families reported a sibling, in all cases a brother, as having died from heart disease).

Qualitative research is underway to investigate the lay reasoning that underpins people's understanding of family histories of heart disease.13 This suggests that while parental deaths are very important in the assessment of familial risk, people also take account of death and illness in other family members. Another quantitative study has demonstrated a strong relation between reporting of a family weakness attributable to heart disease and numbers of both close relatives (parents and siblings) and more distant relatives (grandparents, aunts, uncles and first cousins) with the disease.14 These observations help to explain why 9% of sons and 13% of daughters in the study whose parents had not died of CHD nevertheless reported a family weakness attributable to heart disease. The larger size of this category of offspring results in more than half of all offspring reporting a family weakness attributable to heart disease belonging to this group (table 3).

These data, and also the wide range of ages at which offspring experience parental deaths from CHD illustrate the heterogeneity of family histories of heart disease in the general population. This heterogeneity may be one reason why only a minority of offspring, comprising 23% of sons and 34% of daughters with at least one parent having died from CHD, reported a family weakness attributable to heart disease. There was no interaction between offspring and parental gender in how CHD deaths in parents were perceived.

While daughters and sons were equally likely to have experienced the death of a parent from heart disease, daughters were more likely to perceive such a death in terms of a family weakness. Our qualitative research suggests that while daughters who perceive a family weakness tended to have one parent who had died from heart disease, sons with this perception tended to have either two parents with heart disease or heart problems affecting their father and other relatives on his “side of the family” (unpublished data).

In the data reported here, it is of interest that men in manual occupations, who are at greatest risk of developing CHD, are least likely to perceive a family weakness attributable to heart disease (table 4), with only 17.5% of such sons reporting this perception (table 3). However, this group also had the lowest prevalence of perceived family weaknesses from any cause. A possible explanation is that the greater exposure of men in manual groups to premature death in older relatives makes them less likely to identify specific family weaknesses. It is also of interest that 26% of men in manual groups who did not report a family weakness attributable to heart disease had nevertheless experienced at least one parent dying from CHD.

Our findings may not be directly relevant to current clinical practice concerning the ascertainment and management of people with a family history of premature heart disease. Current clinical guidelines for the definition of a family history of CHD are based on coronary events occurring in first degree relatives under the age of 55.15Only 1.4% of sons and 1.1% of daughters in this population had experienced a parent dying from CHD below this age. Higher proportions are likely to have parents below this age who have survived non-fatal coronary events. By its sampling criteria, the study also excludes parents who died of heart disease below the age of 45.

As understanding of genetic risk of CHD improves, however, it is probable that new definitions of family histories involving larger proportions of the general population will be introduced to clinical practice. Knowledge of a family history of heart disease is also important in relation to activities in cardiovascular health promotion, including attempts to influence individual behaviour. Our findings show that health care professionals need to be sensitive to variations in lay understanding and perception if they are to avoid dysfunctional consultations concerning CHD risk.