Marital status and risk factors for coronary heart disease. The United States health examination survey of adults.

Death rates from coronary heart disease (CHD) for persons aged 25 years and over are higher for the divorced, widowed, or the never-married than for persons who are currently married (Table I). This is true regardless of colour, and though there is some variation of this excess risk with age (e.g., white never-married women aged 55-64 and 65-74 have almost the same rates as white married women) it is present to some extent at all ages above 25. Is this mortality difference real, or can it be explained by systematic errors and misclassification in the registration system? Such errors could occur in the following ways: (1) Selective census under-enumeration, compared with married persons, of the never-married, widowed, and divorced: To the extent that this occurs, the estimate of the non-married population will be falsely low, and thus the calculated mortality rate falsely high. Because they tend to be more transient, non-married persons may indeed be under-enumerated. The magnitude of this error, although not known, is probably not very large. The variation in estimated completeness of enumeration by colour and sex in the 1960 census was as follows: white males 97 2 %, white females 98' 4%, non-white males 89-1 %, non-white females 91'9% (Siegel, 1968). It is unlikely that the difference among marital status groups would be more than the largest colour-sex difference, and this would not be sufficient to account for more than a small part of the excess rate among the non-married. (2) Discrepancies in reports of marital status between the census and matching death certificate: Compared with the 1960 census record of persons dying between May and August 1960, the corresponding death certificate overstated the number of non-married persons by 2 1% (National Center for Health Statistics, 1969). This tendency will