Research Briefs

Ball, J.K., and Elixhauser, A. (1996, September). "Treatment differences between blacks and whites with colorectal cancer." Medical Care 34(9), pp. 970-984.

Blacks have a 59 percent to 98 percent higher odds of dying in the hospital from colorectal cancer than whites, possibly because of the less aggressive treatment that blacks receive for this form of cancer. According to a recent study conducted by AHCPR researchers, whites are more likely than blacks to be hospitalized while colorectal cancer is still localized. Blacks, however, are more often not hospitalized until the cancer is advanced, and they are less likely to receive major therapeutic procedures for this condition. These findings suggest that this type of cancer is unmanaged or poorly managed in black patients, according to the authors. They found that blacks hospitalized with primary tumor were 41 percent less likely than whites to receive a major colorectal therapeutic procedure when the cancer had not yet spread (no metastasis) and therefore was more treatable and 27 percent less likely to receive major therapy when the tumor had metastasized. Blacks were nearly twice as likely to die in the hospital with primary tumor and 1.7 times as likely to die in the hospital with cancer-related sequelae such as nutritional problems. Blacks were more likely than whites overall—458.6 vs. 380.2 discharges per 1,000—to be hospitalized with evidence of progressive cancer and, when evidence of metastasis was present, complications of radiation or chemotherapy, all conditions that capture the effects of unmanaged or poorly managed cancer. These racial differences existed even after allowances were made for differences in patient demographics, insurance status, other clinical factors, and provider characteristics. These findings are based on a study of over 20,000 black and white patients hospitalized with colorectal cancer, using discharge data from 500 hospitals participating in the Hospital Cost and Utilization Project (HCUP) from 1980-1987.

Reprints (AHCPR Publication No. 96-R131) are available from the AHCPR Publications Clearinghouse.

Hallstrom, A.P., Cobb, L.A., Yu, B.H. (1996, June). "Influence of comorbidity on the outcome of patients treated for out-of-hospital ventricular fibrillation." (AHCPR grant HS08197). Circulation 93(11), pp. 2019-2022.

Persons who suffer cardiac arrest with ventricular fibrillation (VF, an irregular heart beat) outside of the hospital have a better chance of surviving if they are younger, someone sees them collapse, and they quickly receive cardiopulmonary resuscitation (CPR) and early defibrillation. A person's coexisting (comorbid) medical conditions are an additional important but often overlooked predictor of out-of-hospital VF survival. The researchers constructed a comorbidity index of a history of 10 chronic conditions and recent (within 2 days) symptoms in 282 victims of out-of-hospital VF, which was strongly associated with outcome. However, the index, when combined with previously identified VF survival predictors, could account for only one-fourth of the variation in patient outcomes. Much of the difference in survival among these patients remains unexplained, according to the researchers. Chronic conditions included diabetes, hypertension, previous heart attack or angina, lung disease, gastrointestinal disorders, and cancer. Symptoms within 2 days of heart attack included chest pain, dizziness, indigestion, dyspnea, nausea, and fatigue. The comorbidity index was significantly lower in patients who survived compared with those who died (0.87 vs. 1.08) and was weakly but significantly correlated with age.

Handler, A., Rauge, K., Kelly, M.A., and Giachello, A. (1996, March). "Women's satisfaction with prenatal care settings: A focus group study." (AHCPR grant HS07376). Birth 23(1), pp. 31-37.

To assess the effects of ethnicity on satisfaction with prenatal care, the researchers convened eight small focus groups comprising 50 low-income Mexican-American, Puerto Rican, black, and white women in Chicago, IL. The majority of the women had several children, did not work, and did not have more than a high school education. Despite their diverse ethnic backgrounds, the women valued the same qualities in prenatal care. They wanted providers who communicated with and respected them and were technically competent. The women valued continuity of care with the same provider, a clean and friendly setting, short waiting times, and sufficient time with their providers. Only 16 percent of the women thought it was very important to have a caregiver from the same ethnic background as themselves. Nearly all of the women (94 percent) thought it was very important that their caregivers explain what they were doing and answer all of their questions. The women expressed some negativity toward foreign-trained or foreign caregivers, but this seemed to have more to do with concerns about training and communication skills than the caregivers' ethnicity.

Morise, A.P., Diamond, G.A., Detrano, R., and others (1996, April-June). "The effect of disease-prevalence adjustments on the accuracy of a logistic prediction model." (AHCPR grant HS06065). Medical Decision Making 16(2), pp. 133-142.

Logistic prediction models, which can estimate the probability of a patient outcome, can aid in clinical decisionmaking. Since not all practitioners or institutions develop their own models using local populations, published models need to be transportable to populations far removed from the population from which the model was derived. The accuracy of the model is degraded when it is transported to populations with outcome prevalences different from those of the derivation population, and the resulting errors from its use can have major clinical implications. The authors developed a logistic prediction model with respect to the noninvasive diagnosis of coronary disease based on 1,824 patients who underwent exercise testing and coronary angiography. They designed the model to adjust only for differences in prevalence. When they applied the adjusted algorithm to three geographically remote populations, calibration improved 87 percent, while discrimination fell by 1 percent. The researchers conclude that adjustment with a mathematical correction algorithm can reduce clinical errors in the accuracy of logistic prediction models used in nonderivative populations.

Stanton, B.F., Li, X., Ricardo, I., and others (1996, April). "A randomized, controlled effectiveness trial of an AIDS prevention program for low-income African-American youths." (AHCPR grant HS07392). Archives of Pediatric and Adolescent Medicine 150, pp. 363-372.

Targeting discussions about condom use to prevent AIDS and other sexually transmitted diseases (STDs) to small groups of urban black adolescent friends increases their use of condoms up to 6 months later. Use of friendship groups appears to help motivate this behavioral change. But to sustain this protective behavior beyond 6 months requires "booster" interventions, according to these authors. They recruited 383 black urban youths from nine community recreation centers in a large Eastern U.S. city and assessed the youths' use of condoms before and after participation in eight sessions designed to encourage condom use. The 206 intervention youths were organized into 76 already-formed friendship groups to reinforce messages about condom use. Prior to the sessions, 36 percent of the youths were sexually experienced, and by 1 year later, 49 percent were sexually active. The 177 control youths participated in discussion groups but not necessarily with friends, and personal decisionmaking and protection motivation were not emphasized. Six months later, 85 percent of intervention youths versus 61 percent of control youths reported condom use. However, after 12 months the differences in condom use were no longer apparent.

Wolinsky, F.D., and Stump, T.E. (1996). "A measurement model of the medical outcomes study 36-item short-form health survey in a clinical sample of disadvantaged, older, black and white men and women." (AHCPR grant HS07632). Medical Care 34(6), pp. 537-548.

The Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36) has become the functional health status measure of choice in outcomes studies. However, its use has been limited to young, healthy, and socioeconomically advantaged populations. This study examined the validity of its use in a clinical sample of disadvantaged, older adults with significant coexisting medical conditions such as hypertension and heart disease. The authors performed confirmatory factor analysis on data obtained from baseline face-to-face interviews with 1,051 patients who were at risk for acute deterioration of their clinical condition due either to their age alone (75 years or older), or to their age (50 to 74 years old) combined with major coexisting medical problems. They found that some changes in the SF-36 are needed if it is to be applied routinely to disadvantaged, older, sicker adults. For example, they recommend including health optimism as an additional factor to the current eight factors.

Current as of September 1996

Internet Citation:

Research Activities newsletter. September 1996, No. 196. Agency for Health Care Policy and Research, Rockville, MD. http://www.ahrq.gov/research/sep96/