Research note: proxy responses in health surveys: a methodological issue

Area of Science:

• Health Services Research
• Epidemiology
• Sociology of Health

Background:

• Sex differences in morbidity, with women reporting more illness, are well-documented.
• The use of proxy respondents in health surveys has been questioned as a potential source of bias, particularly regarding sex differences.
• It is hypothesized that proxy respondents may under-report illnesses of absent individuals, potentially skewing observed sex differences.

Purpose of the Study:

• To investigate the impact of proxy respondent effects on reported sex differences in morbidity.
• To determine if proxy reporting biases contribute to the observed excess female morbidity in health surveys.

Main Methods:

• A study involving 50 married couples was conducted.
• Participants were asked to report symptoms experienced by themselves and their spouses.
• Agreement between self-reports and proxy reports of symptoms was analyzed.

Main Results:

• A high level of agreement was observed between spouses regarding the number and severity of experienced symptoms.
• Discrepancies, when present, showed a tendency for proxy respondents to over-report symptoms, regardless of sex.
• This suggests that proxy effects, if any, would likely minimize, not create, observed sex differences in illness.

Conclusions:

• The observed excess female morbidity in health surveys is unlikely to be an artifact of proxy respondent bias.
• Proxy reporting does not appear to explain the higher rates of illness reported by women compared to men.
• Differences in symptom experience or reporting between sexes are more probable explanations for observed sex differences in morbidity.