Data and Sample
The National Ambulatory Medical Care Survey (NAMCS) is a nationally representative survey of physicians providing office-based ambulatory medical care in the United States. The survey has been conducted annually since 1973 by the National Center for Health Statistics (NCHS) of the U.S. Centers for Disease Control and Prevention. The target universe of the NAMCS includes ambulatory visits made to offices of nonfederally employed physicians, excluding visits made in hospital-based settings and care provided by anesthesiologists, radiologists and pathologists. Each physician is asked to complete a patient record for a random sample of visits occurring during a randomly assigned one-week period. Information is recorded by the physician regarding the patient’s demographic characteristics, symptoms, diagnosis and the medical services, tests, medications and procedures provided during the visit. In addition, supplementary information about the physician’s practice is collected by NCHS investigators. The NAMCS does not obtain information concerning the patient’s employment status, work history, medical care costs or the outcomes of care.
This study is based on NAMCS data for office visits made during 1997 and 1998. Two years of survey data were combined to provide more reliable estimates. NAMCS utilizes a multistage probability sampling process to identify eligible physicians that are representative of the national distribution of ambulatory care visits. Responses were obtained from 3,607 eligible physicians covering 48,054 patient visits. Physician response rates were 69.2% in 1997 and 67.9% in 1998. Item nonresponse rates were generally less than 5%. viagra jelly online

Study Variables
Patient race and ethnicity were determined by the physician on the basis of physical observation and information supplied by the patient. The physician selected from among five racial categories provided on the patient record form: white, black/African-American, Asian, native Hawaiian/other Pacific Islander and American Indian/Alaska native. For the purposes of this study, we used racial data for patients identified as white or black/African-American (henceforth referred to as “African-American”). Data for patients identified as Asian, native Hawaiian/other Pacific Islander and American Indian/ Alaska native were excluded from the racial analyses because of their low numbers. Available ethnicity categories were: “Hispanic or Latino” and “Not Hispanic or Latino.” Racial and ethnic categories were not mutually exclusive, so that, for example, a patient identified as African-American could also be classified as Hispanic. There was only a slight overlap between minority groups with 7.1% of African Americans identified also as Hispanic, and 4.1% of Hispanics identified also as African Americans.

Visits that were primarily for the care of work-related conditions were identified either by: 1) the physician specifying “workers’ compensation” as the primary expected source of payment for the visit, or 2) the physician answering “yes” to the question, “Was this visit related to injury or poisoning?” followed by an affirmative response to the question, “Was this injury work-related?” For the purposes of this study, a visit was considered to be a visit for a work-related condition if it satisfied either of these two criteria. Don’t suffer without medication. Buy Viagra Super Active online

Statistical Analysis
The basic unit of analysis for this study was the patient visit. Visit weights were calculated for each patient visit by the NCHS based on the sampling strata adjusted for response rates for the combined 1997-1998 period. An unadjusted subgroup analysis was performed, stratified by race, to describe the care for work-related conditions provided to African-American patients and to white patients. A similar unadjusted analysis described care for work-related conditions involving Hispanic patients and non-Hispanics. Bivariate statistics were calculated using weighted data to describe the percent distribution of medical care characteristics for patients within each of these racial/ethnic categories. T-tests for continuous variables and Chi-squared tests for discrete variables were performed to estimate the relative differences in care provided for work-related conditions between African-American and white patients, and between Hispanic and non-Hispanic patients. The level of statistical significance was set at p<0.05. The bivariate analyses were performed using SAS (version 8.2) analytical software. cialis super active online

Multiple logistic regression analyses were conducted to derive adjusted estimates of the influence of being African-American (compared to white) and Hispanic ethnicity (compared to non-Hispanic) on various dimensions of ambulatory care provided for patients with work-related conditions. The primary regression model included the following covariates: age of the patient, gender of the patient, geographic region of the physician practice and MSA status (urban/rural location) of the physician practice. The regression analysis was performed 10 times, each time using a different dimension of care as the dependent variable in the regression model Dimensions of care included in the regression analyses all had been found to vary significantly between African Americans and whites (or between Hispanics and non-Hispanics) in the initial unadjusted bivariate analyses. To investigate whether the results might vary by type of diagnosis, we repeated the regression analyses restricted to cases involving musculoskeletal disorders (ICD-9 codes 710-739) and acute injuries (ICD-9 codes 800-999), the two most common diagnostic categories for work-related conditions. zyprexa medication

Odds ratios and 95% confidence intervals were calculated for each of the regression analyses. Because the estimates derived from our analyses arebased on a sample rather than on the entire target universe of office visits, the results were subject to sampling error. To account for sampling error, the 95% confidence intervals around the reported odds ratios were calculated by Taylor approximation, using SUDAAN analytical software.