Estimating racial/ethnic disparity in mammography rates: it all depends on how you ask the question

BACKGROUND: Estimates of racial disparity in mammography appear to differ depending on the data source. This study examined the impact of different survey methodology on estimates of racial disparity in mammography. METHODS: Responses from 3,090 women > or =40 years to two different questions from the 1996 Medical Expenditure Panel Survey (MEPS) were compared when a mammogram was last obtained versus what medical services, including mammography, were obtained over a 4-month interval, aggregated across 1 year. RESULTS: There was no significant racial disparity in 1-year mammography prevalence based on the first question (white-black difference, 3.3%; 95% confidence interval [CI], -2.5, 9.2). In contrast, a significant disparity in 1-year mammography prevalence was found based on the medical services question (difference, 13.1%; 95% CI 8.6, 17.6). Disparity estimates by Hispanic ethnicity were similar for the two questions: white-Hispanic difference, 1.6%; 95% CI -4.3, 7.5, and white-Hispanic difference 5% (-0.2, 10.1). Adjustment for age, income, and insurance did not alter these findings. CONCLUSIONS: Estimates of racial, but not ethnic, disparities in mammography seem to depend on how the question is asked. These results caution against exclusive reliance on annual self-reports for monitoring disparities in preventive care.     

What they tell you to forget: from child sexual abuse to adolescent motherhood

This study explores the relationship between child sexual abuse and adolescent motherhood, using a life story interview method. The sample consists of 27 mothers participating in a home-visitation parenting program for mothers at risk of child maltreatment. The failure to articulate the violation of child sexual abuse and to appropriately construct blame resulted in a range of self-destructive behaviors, some of which placed mothers at greater risk of teen pregnancy. Repressed feelings associated with the trauma often resurfaced with motherhood as victims re-experienced their innocence and vulnerability as children.     

Medicaid and mental health: be careful what you ask for

Medicaid has had an enormous impact on the shape and impact of public mental health care. Medicaid mental health policy has expanded access, fostered consumerism, and created incentives for expansion of community-based providers. It also has dramatically changed the economic rules governing public mental health care, leading state governments to alter their behavior. The result has been a tilting of public mental health care toward Medicaid-covered people and services.     

Women's perceptions of breast cancer risk: how you ask matters.

BACKGROUND: Some studies suggest that women dramatically overestimate the risk of having breast cancer while others conclude that they underestimate it. To understand better how women perceive the chance of getting breast cancer, the authors asked women to estimate the risk in several ways. Each woman's answer was related to her actual risk. METHODS: Women were randomly selected from a registry of female veterans in New England. A mailed questionnaire asked each woman to estimate her ten-year risk of dying from breast cancer as a number out of 1,000 ("___ in 1,000" perceived risk) and whether this risk was higher than, the same as, or lower than that of an average woman her age (comparative perceived risk). The woman was also asked to compare her risk of dying from breast cancer with her risk of dying from heart disease. Risk-factor data were collected so that each woman's actual risk of breast cancer death could be estimated (actual risk). RESULTS: 201 women had complete data. The median age of the respondents was 62 years (range 27-80), and 98% were high school graduates. Most women (98%) overestimated the "___ in 1,000" risk of breast cancer death-half by eightfold or more (interquartile range, 4-36-fold overestimates). In contrast, only 10% of these women thought that they were at higher risk than an average woman their age. Most correctly thought that their risk of dying from breast cancer was lower than their risk of dying from heart disease. The women's "____in 1,000" perceived risks of breast cancer death were unrelated to their actual risks and had no significant agreement with an external bench-mark of importantly "high risk" (i.e., met risk criteria for the Tamoxifen primary prevention trial). In contrast, the women's comparative perceptions of being at low, average or high risk were related to actual risks and significantly agreed with the "high risk" benchmark. Most women not at importantly "high risk" correctly classified themselves; however, almost two thirds of "high risk" women misclassified themselves as "average or lower than average risk." CONCLUSIONS: The method used to elicit perceptions of risk matters. These women's responses to the comparative questions showed that they "knew more" about their actual risks than their open-ended numeric responses suggested.     

Socio-economic position and health: what you observe depends on how you measure it

BACKGROUND: A number of different socio-economic classifications have been used in relation to health in the United Kingdom. The aim of this study was to compare the predictive power of different socio-economic classifications in relation to a range of health measures. METHODS: A postal questionnaire was sent to a random sample of adults in the West of Scotland (sampling from 1997 electoral roll, response rate 50 percent achieved sample 2,867). RESULTS: Associations between social position and health vary by socio-economic classification, health measure and gender. Limiting long-standing illness is more socially patterned than recent illness; income, Registrar General Social Class, housing tenure and car access are more predictive of health than the new National Statistics Socio Economic Classification; and men show steeper socio-economic gradients than women. CONCLUSION: Although there is a consistent picture of poorer health among more disadvantaged groups, however measured, in seeking to explain and reduce social inequalities in health we need to take a more differentiated approach that does not assume equivalence among social classifications and health measures.     

Te contaria mi vida: I would tell you my life,if only you would ask

Universal screening for domestic violence is recommended in many health care settings. This qualitative study was undertaken to explore the thoughts and feelings of Mexican American women regarding being asked questions about domestic violence by a health care provider. We wanted to further explore what characteristics about a nurse, or other health care provider, would give a woman confianza, the trust necessary to discuss this issue. Seven women, who self identified as abused or formerly abused, were recruited from a pool of Spanish-speaking women receiving services from a rural domestic violence agency in the midwestern United States. The researchers found that, given certain characteristics and actions of the health care provider, women welcome the opportunity to discuss this issue. The implications for practice are these: be sincerely present for the client, ask about her life, listen to her response, and when necessary assist her to connect with appropriate domestic violence community services.     

When you can't ask their names: linking anonymous respondents with the Hogben number

Abstract: This article describes a method of linking anonymous subjects with a respondent-generated code using an algorithm based on personal details to produce unique identifiers. It was used to increase confidentiality and statistical power in a yearlong work-place health promotion evaluation. Subjects were employees of a large retail chain; 80 per cent were female, and the majority educated to high school level. Of the 385 possible, 81 per cent matched; 67 per cent of the codes were matched on all elements and another 14 per cent were accepted as ‘fuzzy’ matches. Linking respondents increased the statistical power of the study from an unacceptable 0.4 to an acceptable 0.8. Other research on linking records is briefly discussed, including sample bias and probabilistic matching. This technique is useful when anonymity is likely to raise response rates, but the ideal code could be further sought.