Awareness and Use of the Prostate-Specific Antigen Test: Method

Posted by James

The NHIS 2000 is a nationally representative health survey conducted by the National Center for Health Statistics. NHIS data are collected through personal household interviews. The NHIS 2000 used a complex sample design involving stratification, clustering and multistage sampling. The NHIS 2000 survey collected information related to cancer prevention and control. African-American and Hispanic populations were oversampled to allow for more precise estimations. In the overall survey, men aged >40 were asked questions related to prostate cancer (Eulexin drug is used along with drugs such as Lupron to treat prostate cancer) screening awareness and test use. Sample weights were constructed to reflect the total population of the United States in 2000. More details of the overall study methods are available in the 2000 NHIS Survey Description.

Study Population

This sample was restricted to non-Hispanic African-American men aged >45. Those men who reported having prostate cancer were excluded. Interviews were conducted with approximately 736 African-American male adults, with an overall adult sample response rate of 72.1%.

Data Collection

Respondents were asked if they had heard of the PSA test, and if so, whether they had ever had a PSA test. We examined four sets of factors that may influence both having heard of the PSA test and ever having had a PSA test: sociodemographic, perceived risk, health status and family composition. Demographic variables included age, level of education, region of the country and metropolitan statistical area (MSA). Regions and MSAs are defined by the Census Bureau and Office of Management and Budget (OMB). The OMB defines an MSA as one or more counties that contain a city of 50,000 or more, or an urbanized area that has a total population of at least 100,000. Family income was expressed in relation to federal poverty levels published by the U.S. Census. For 2000, the poverty level for a family of four was set at $17,603. “Unknown” family income was included as a category due to the relatively large rate (26.2%) of nonre-sponse for the question on income.

For family composition variables, marital status was divided into four categories, family size into three, and family structure into four categories (Table 1). Health-related variables included health insurance (none, public or private), reported health status or a self-assessment of general health (four categories), family history of prostate cancer (Xeloda tablets is the only FDA-approved oral chemotherapy for both metastatic breast cancer and adjuvant and metastatic colorectal cancer), and whether respondents had received colorectal cancer screening.

Table 1. Characteristics of Non-Hispanic Black Men Aged 45 or Older, NHIS 2000 Survey

Sociodemographics	N		%a	Health Status	N	%a
Age (Years)	736			Health Insurance	730
45 to 49			29.5	None		13.4
50 to 64			45.2	Public		23.1
65+			25.3	Private/military		63.5
Education	723			Health Status	731
< High-school graduate			34.6	Excellent		18.7
High-school graduate			30.9	Very good		20.5
Some college/tech school			20.9	Good		32.1
College grad			13.6	Fair or poor		28.7
Poverty Threshold	736			Family History of Prostate Cancer	686
<200%			28.9	Yes		6.2
200-299%			13.7	No		93.8
300-399%			9.1
400-599%			7.5	Had Colorectal Screening10	505
500%+			14.6	Yes		30.8
Unknown			26.2	No		69.2
Region	736			Family Composition                 N		%a
Northeast			16.6	Marital Status                        733
Midwest			16.8	Married/living with partner		63.0
South			58.8	Widowed		7.3
West			7.8	Divorced or separated		18.0
				Never married		11.7
Metropolitan Statistical Area	736
No			13.9	Family Size (Persons in the Family) 736
Yes			86.1	1		27.6
				2		34.8
Perceived Risk		N	%a	3+		37.5
Perceived Risk of Getting Cancer		599
Low			69.6	Family Structure                     736
Medium			22.5	One adult, no children		27.6
High			8.0	Multiple adults, no children		47.9
				One adult, 1+ children		1.1
Perceived Amount of Cancer in Family 657				Multiple adults, 1+ children		23.3
Low			80.5
Medium			14.8	a Adjusted to the US 2000 Census; b Question asked of men >50
High			4.7	years of age.

For family history and perceived risk variables, first-degree family history of prostate cancer (Casodex medication is an oral non-steroidal anti-androgen for prostate cancer) was obtained from several questions about cancer diagnoses in respondents’ biological fathers, brothers and sons. For perceived risk of getting cancer (in general) and perceived amount of cancer (Generic Nolvadex is an anti-estrogen used to treat or prevent breast cancer) in one’s family, respondents were asked subjective assessment questions about their likelihood of getting any type of cancer and the amount (low, medium or high) of any type of cancer in their families.

Data Analysis

The statistical program SUDAAN was used in the analysis to take into account the complex multistage probability sampling design of the NHIS 2000. The 95% confidence interval was used to assess variability in the estimates of percentages. General linear contrast measures were calculated to assess statistical significance of differences among men within the categories of a variable—for example, differences in test use with men from the northeast and men from the south. P values for general linear contrasts compare a row percent to its reference level. The Wald Chi-squared test was used to assess the association of each variable with the outcome variable after adjusting for age. The final logistic model contained 12 main effects. Odds ratios were used to examine the associations of independent variables with having heard of the PSA test and having been tested while adjusting for other variables in the model. Family structure was dropped due to its strong association with marital status and family size. Fifty-five of the 736 men interviewed were excluded from the study because they did not answer the question, “Have you ever heard of a PSA, or prostate-specific antigen test?”