The art and science of interpreting survey data

Homing in on the homeless: assessing the physical health of homeless adults in Los Angeles County using an original method to obtain physical examination data in a survey

OBJECTIVE

Public policy that decreases the finding for social services may combine with the ascendancy of corporate managed care to increase the health care deficit. Assessing the health impact of these policy changes on various populations is a fundamental challenge for health services research. Disadvantaged populations, such as the homeless, are likely to be affected disproportionately. Research quality data on the physical health of such populations are difficult and expensive to obtain. In particular, physical examination data have not been available and self-reports are insufficient. Our objective: to develop and utilize a structured physical exam system enabling lay survey researchers to report reliably physical findings related to six tracer conditions in a disadvantaged population.

STUDY SETTING

A field survey of homeless adults in Los Angeles County, California. Respondents were 363 homeless adults representing a subsample of a probability sample of the county's homeless adult population.

STUDY DESIGN

We integrated existing measures with expert clinical opinion and original means of data collection into a structured physical exam enabling lay interviewers to identify the prevalence of vision problems, significant skin disorders, peripheral vascular disease of the lower extremities, selected podiatric disorders, hypertension, and tuberculosis in a sample of homeless adults.

PRINCIPAL MEASURES

We describe lay interviewer performance in terms of mastery of the necessary material based on written and practical exams and in terms of the number of respondents successfully followed. We base our description of the instrument on the time necessary to complete it, and on the proportion of each component successfully completed during the field survey, as well as on interrater reliability. We report the prevalence of the various clinical conditions according to self-report and according to the structured limited physical exam, as well as the marginal proportion of respondents who were identified by the physical exam and not by self-report.

PRINCIPAL FINDINGS

Interviewers performed the exam successfully under field conditions. Respondent acceptance of the instrument was high. Interrater agreement was 100 percent regarding the need for referral on the basis of blood pressure and vision. Kappa statistics for skin, foot, and edema findings were .67,.71, and .81, respectively. Adjusted for sampling weights, 60 percent of this population required referral for at least one of the specified conditions. For those portions of the survey for which both self-report and physical exam data were available, lay interviewers made significant percentages of referrals on the basis of physical findings alone.

CONCLUSIONS

High blood pressure, poor vision, peripheral vascular diseases of the feet and legs, and significant skin conditions are prevalent among the homeless in Los Angeles County. Without physical exam data, estimates of the prevalence of these conditions will be incorrect. Researchers can use laypersons to collect reliable and valid physical exam data on disadvantaged populations. This represents a new tool for assessing and monitoring the health of these populations.

Calcium intakes of Mexican Americans, Cubans, Puerto Ricans, non-Hispanic whites, and non-Hispanic blacks in the United States

OBJECTIVE

To compare dietary calcium intakes from food in Mexican Americans, Cubans, Puerto Ricans, non-Hispanic whites, and non-Hispanic blacks aged 11 through 74 years.

DESIGN

Population survey data from the Hispanic Health and Nutrition Examination Survey and the second National Health and Nutrition Examination Survey were used to calculate calcium intake from a single 24-hour recall. These data were compared by age and sex between the five population groups. Food sources of calcium in the three Hispanic groups were also examined using 24-hour recall data.

SUBJECTS

The sample consisted of 11,773 non-Hispanic whites, 1,728 non-Hispanic blacks, 4,739 Mexican Americans, 1,076 Cubans, and 1,835 Puerto Ricans.

MAIN OUTCOME MEASURES

Mean calcium intake, percentage intake of Recommended Dietary Allowance, and, for Hispanics, food sources of calcium.

STATISTICAL ANALYSES

Means were compared within age and sex groups between the five population groups using a t test.

RESULTS

Calcium intakes from food in the three Hispanic groups were similar to intakes of non-Hispanic whites and higher than intakes of non-Hispanic blacks. Although dairy foods were the main sources of calcium for Hispanics, corn tortillas were important calcium sources among Mexican Americans. Women consumed less calcium than the Recommended Dietary Allowance in all age and racial or ethnic groups.

APPLICATIONS

When assessing calcium intakes of the three Hispanic groups, ethnic differences in food sources of calcium need to be considered. Efforts to increase calcium intake in Hispanics also need to account for ethnic differences.

Equity in environmental health: data collection and interpretation issues

In order to assess the issue of inequity in exposure to environmental hazards, researchers must identify subgroups whose exposure is disproportionately greater than the average exposure experienced by the remainder of the population. The general population is a complex mixture of subgroups, each consisting of individuals who experience a wide range of exposures and whose ability to cope with the consequences of those exposures is equally varied. Therefore, large efforts are needed to collect data that will enable researchers to determine comprehensively which subgroups are highly exposed and which subgroups have disproportionately greater health effects as a result of exposures to environmental hazards. The interpretation of findings is more of an art than a science, especially when two population subgroups are being contrasted. Addressing environmental equity requires explicit comparisons between groups, and racial and ethnic contrasts will be prominent. It is often difficult to identify the underlying mechanisms that produce particular patterns of results. However, researchers and policy makers must understand the dynamics that may have produced a particular pattern of results so they can separate those factors that are amenable to change from those that are not.

Adjusting for nonresponse bias in a health examination survey

There is a potential for nonresponse bias in most population studies using health examinations. This is true of the Mexican American portion of the Hispanic Health and Nutrition Examination Survey (HHANES), conducted by the National Center for Health Statistics, in which unit nonresponse to the examination accounted for 24 percent of the sample. Even though the full effect of nonresponse can never be really known, ancillary information from the interview sample can be used in an attempt to adjust for bias in estimates. Two techniques for nonresponse bias adjustment are presented and illustrated using health status level and hypertension status from published studies based on the HHANES of 1982-84. The first approach uses conditional probabilities and the second approach uses direct standardization. The examples examine whether or not an adjustment for socioeconomic status, sex, and age--variables related to both response status and the conditions under study--changes the prevalence estimates of (a) Mexican Americans who report poor, fair, or good health status and (b) hypertension among Mexican Americans.