Robert Hiatt

It is becoming increasingly evident that early-life events and exposures have important consequences for cancer development later in life. However, epidemiological studies of early-life factors and cancer development later in life have had significant methodological challenges such as the long latency period, the distinctiveness of each cancer, and large number of subjects that must be studied, all likely to increase costs. These traditional hurdles might be mitigated by leveraging several existing large-scale prospective studies in the United States (US) and globally, as well as birth databases and birth cohorts, in order to launch both association and mechanistic studies of early-life exposures and cancer development later in life. Dedicated research funding will be needed to advance this paradigm shift in cancer research, and it seems justified by its potential to produce transformative understanding of how cancer develops over the life-course. This in turn has the potential to t...

Recent studies have suggested a role for dietary fat in the etiology of breast cancer. The relation of serum cholesterol and other serum lipid measures to breast cancer incidence was investigated in a cohort of 95, 179 women who had multiphasic health checkups from 1964 through 1972. Through 1977, 1,035 new breast cancer cases occurred in over 752,000 person-years of follow-up. Age-adjusted incidence rates were 1.45, 1.37, 1.31, and 1.40/1,000 person-years from the lowest to the highest quartile of serum cholesterol level, respectively. Similarly, no statistically significant relation was detected between beta-lipoprotein or total lipids and breast cancer. The sample size was sufficiently large to have detected a relative risk of 1.4 or larger with a probability of 99.9% at the 0.05 level of significance. The expected relation od breast cancer to established risk factors was confirmed by univariate analysis, and serum cholesterol and breast cancer were not associated after simultane...

The authors prospectively studied the incidence of cancers of the colon and rectum in 106,203 men and women, both white and black, who supplied data at northern California Kaiser Permanente facilities about use of alcoholic beverages in 1978-1984. Analysis controlling for age, sex, race, body mass index, coffee use, total serum cholesterol, and education showed a positive association of alcohol use to both types of cancer, which was stronger for rectal cancer (trend test, p = 0.03) than for colon cancer (trend test, p = 0.11). When persons with a daily intake of three or more drinks were compared with abstainers, relative risk for rectal cancer was 3.17 (95% confidence interval (CI): 1.05-9.57) and relative risk for colon cancer was 1.71 (95% CI: 0.92-3.19). Women with a daily intake of three or more drinks had a relative risk for colon cancer of 2.56 (95% CI: 1.03-6.40) compared with 1.16 (95% CI: 0.46-2.90) for men. Among drinkers, preference for wine, beer, or hard liquor had no ...

We examined the association between alcohol and caffeine consumption and premenstrual syndrome (PMS). One hundred two women with PMS matched by age and race to an equal number of women without PMS were compared for differences in intake of alcoholic and caffeinated beverages, with differences determined using Student's t-tests and conditional logistic regression for matched pairs. Information on alcohol and caffeinated beverage consumption was obtained from three 24-hour dietary recall interviews conducted during the postmenstrual period and from three conducted during the premenstrual period. Results showed that no significant difference was observed in total caffeine intake or in the individual caffeinated beverages consumed during either the post-menstrual or the premenstrual period. For alcohol consumption, however, women with PMS drank 1.41 (95% confidence interval; range, 0.34-2.47) more servings per week during the postmenstrual period. Based on post-menstrual consumption...

The rate of occurrence of stones of the urinary tract was assessed in a large population served by the Northern California Kaiser Foundation Health Plan. The study involved three separate groups. First, data were obtained by questionnaire from approximately 175,000 adults who took a multiphasic health checkup in the period 1964-1972; of these generally well adult members, 26.2/1000 persons (32.0/1000 men and 21.0/1000 women) reported having ever been told by a physician that they had a urinary tract stone. Second, data were obtained from 139,000 persons served by the San Francisco outpatient facility in 1970-1972; 1.22/1000 per year (1.81/1000 men and 0.59/1000 women) had an initial diagnosis of a "new or recurrent" stone of the upper urinary tract. The third set of data was procured from the entire Northern California region in 1971-1975; 0.36/1000 (0.52/1000 men and 0.19/1000 women) were discharged from a hospital each year with a diagnosis of upper urinary tract stone. ...

In a case-control study of urinalysis screening in the prevention of death from bladder cancer, hematuria was present in a higher proportion of cases than controls as long as five or six years before the diagnostic evaluation that led to the diagnosis of bladder cancer. In a separate cohort study data base that permitted the follow-up of 1046 persons with a physician's diagnosis of hematuria, 11 cases of bladder cancer were diagnosed more than two (mean 7.4) years after the hematuria diagnosis (4.3 cases expected; age-sex standardized morbidity ratio, 2.5; 95% confidence interval, 1.3-4.5). Bladder cancer was ruled out initially by cystoscopy in 8 of the 11 cases. Although we cannot be certain that preexisting bladder cancer or bladder cancer risk factors did not cause the bleeding, we hypothesize that hematuria can be a predictor as well as a manifestation of bladder cancer, based on a tendency for bladder mucosa with premalignant changes to bleed. The implications for screenin...

Barbiturate exposure during childhood was assessed from medical records of 237 children with intracranial and spinal cord tumors and 474 matched controls in a prepaid health plan. In utero exposure was also examined in a subset of 86 "cases" and 172 controls whose mothers were health plan members during pregnancy. No association of in utero exposure to barbiturates was found [odds ratio (O.R.) = 0.96, 95% confidence interval (C.I.) = 0.47, 1.94]. An association was noted for barbiturate use during childhood (O.R. = 1.80, 95% C.I. = 1.18, 2.74) but was reduced (O.R. = 1.41, 95% C.I. = 0.89, 2.21) when history of epilepsy was taken into account and was no longer significant. An apparent dose-response effect disappeared after adjustment for a history of epilepsy. Although barbiturate use for epilepsy due to preexisting brain tumors clearly explains some of the observed association, the small, residual risk prevents us from ruling out a possible carcinogenic effect of barbitur...

To evaluate screening urinalysis for its efficacy in preventing fatal bladder cancer. A case-control study of 290 case subjects with fatal bladder cancer upsilon 290 age and sex matched control subjects, with ascertainment of urinalyses by medical record review. In the five years before the symptoms that led to the detection of bladder cancer, case subjects had significantly fewer routine (screening) urinalyses than controls (matched odds ratio for one or more urinalyses upsilon none = 0.60 [95% confidence interval (CI) 0.41 to 0.87], adjusted for cigarette smoking and occupational bladder cancer risk). However urinalyses for urinary symptoms and problems before those that led to bladder cancer diagnosis were more frequent in the cases; hence there was only a small statistically non-significant deficit in cases of urinalyses for all reasons combined (corresponding adjusted odds ratio 0.94, 95% CI 0.61 to 1.46). Since urinalyses for symptoms and problems could have led to the detecti...

To investigate associations between maternal pregnancy hyperglycemia, gestational diabetes mellitus (GDM), and offspring adiposity. We evaluated these associations in a longitudinal study of 421 mother-daughter pairs at Kaiser Permanente Northern California. Maternal pregnancy glucose values were obtained from maternal medical records. Outcomes included three measures of girls' adiposity, measured annually: (1) ≥85th age-specific percentile for BMI; (2) percent body fat (%BF); and (3) waist-to-height ratio (WHR). Adjusting for maternal age at delivery, race/ethnicity, pregravid BMI, girl's age, and girl's age at onset of puberty, having a mother with GDM increased a girl's risk of having a BMI ≥85th percentile or having %BF or WHR in the highest quartile (Q4), compared with those in the lowest quintile of blood glucose (odds ratio [OR] 3.56 [95% CI 1.28-9.92]; OR 3.13 [95% CI 1.08-9.09]; and OR 2.80 [95% CI 1.00-7.84], respectively). There was a significant interacti...

It has long been accepted that modern reproductive patterns are likely contributors to breast cancer susceptibility because of their influence on hormones such as estrogen and the importance of these hormones in breast cancer. We conducted a meta-analysis to assess whether this 'evolutionary mismatch hypothesis' can explain susceptibility to both estrogen receptor positive (ER-positive) and estrogen receptor negative (ER-negative) cancer. Our meta-analysis includes a total of 33 studies and examines parity, age of first birth and age of menarche broken down by estrogen receptor status. We found that modern reproductive patterns are more closely linked to ER-positive than ER-negative breast cancer. Thus, the evolutionary mismatch hypothesis for breast cancer can account for ER-positive breast cancer susceptibility but not ER-negative breast cancer.

The objective of this study was to use successful quality improvement initiatives in large multispecialty medical groups to identify the organizational factors that were the most important to improvement. The study analyzed the most successful quality improvement initiatives from those submitted by the 24 members of the Council of Accountable Physician Practices. Twelve initiatives from 8 groups were selected that met the study criteria for large improvement for large numbers of patients. An independent group used these initiatives to identify potentially important factors and then asked key local leaders to rate the importance of these factors on a scale of 1 to 4, importance rating (1-4 scale) for each of 18 identified factors. Eighteen factors were identified and 5 stood out as ranked a 4 (Very Important) for at least 80% of the initiatives: Communication, Use of Evidence-Based Medicine, Leadership, Measurement, and Reporting. Another 7 of the 18 factors were ranked a 4 for more ...

In this paper, we describe the attributes of a comprehensive approach to breast cancer screening possible in an integrated health care system. We define an integrated health care system as one in which comprehensive preventive and medical care is provided to a defined population, by a defined panel of providers, and in which this care can be tracked using automated electronic data systems. Guided by the Pathways Conceptual Framework, it is possible to identify and systematically address (through research and interventions) the multiple predisposing, enabling, and reinforcing factors at the individual and organizational level associated with each step along the screening process. This framework is helpful as both a planning and an evaluation tool, in identifying places in the screening and follow-up process that could benefit from concerted quality improvement efforts and in guiding an evaluation of those efforts. We describe examples from research and organizational programmatic eff...

In preparation for jointly publishing official government cancer statistics, the Centers for Disease Control and Prevention (CDC) and the National Cancer Institute (NCI) compared incidence rates from NCI's Surveillance Epidemiology and End Results (SEER) Program and CDC's National Program of Cancer Registries (NPCR). Data for 1999 covering 78% of the US population were obtained from SEER and selected NPCR registries that met high quality data criteria. Incidence rates (per 100,000 population) were age-adjusted to the 2000 US standard population, and 95% gamma confidence intervals were estimated. NPCR rates for all sites combined were higher than SEER rates (males: NPCR 553.6, SEER 538.7; females: NPCR 420.8, SEER 412.5), but rates for specific cancer sites varied by registry program. Rates for colon cancer (males: NPCR 47.0, SEER 42.7; females: NPCR 36.5, SEER 33.8) and tobacco-related cancers were higher in NPCR than SEER. In contrast, NPCR rates were lower than SEER rates ...

Current knowledge is scarce on Latino dietary practices. This study compared the dietary practices, alcohol consumption, and smoking behavior of Latinos and non-Latino whites in two randomly selected samples. Telephone surveys of adults 35-74 years of age from the Kaiser Permanente Medical Care Program (Latinos = 844; non-Latino whites = 510) and from census tract-based areas (Latinos = 806; non-Latino whites = 436) were conducted in the San Francisco Bay Area. Latino ethnicity was a significant predictor of dietary and alcohol consumption practices in multivariate logistic regression models after adjustment for sex, education, age, employment, health insurance, martial status, county of residence, and self-perceived health status. Compared with non-Latino whites, Latinos were significantly less likely to report eating vegetables and more likely to eat rice, beans, and fried foods and to drink whole milk. Less acculturated Latinos were more likely to eat fruits, rice, beans, meat, a...

Difficulties in identification of Latinos from secondary datasets have often been addressed by the use of a Spanish surname list. To evaluate the sensitivity and specificity of the 1980 U.S. Census Spanish surname list in correctly identifying Latino ethnicity, we conducted a cross-sectional survey of adult members of a prepaid health plan in the San Francisco Bay Area. We compared a randomly selected list of 1345 Spanish surnames and 717 non-Spanish surnames from health plan membership lists with their self-identified race and ethnicity obtained by telephone interviews. The sample was stratified according to surname group (Spanish or non-Spanish), sex, age, and county. Compared with self-identified ethnicity, a Spanish surname was 88.4% sensitive in identifying Latino men and 70.4% sensitive in identifying Latina women. The non-Spanish surname list was 94.1% sensitive and 95.0% specific in correctly identifying non-Latinos. Although the overall negative predictive value was 97.8%, ...

The relation of work-related psychologic stress to spontaneous abortion was examined in a prospective study of 3,953 pregnant, employed members of a prepaid health plan in California, recruited between 1990 and 1991. Information on occupation, psychologic stress-related factors at work, social support at work, physical exertion at work, life events, pregnancy worries, and potential confounders was obtained by a telephone interview. Psychologic job stress and social support at work were assessed using an abbreviated version of the instrument developed by Karasek and colleagues (Los Angeles: University of Southern California, 1986). Results from the multivariate model indicated that, overall, stressful work was not associated with an increased risk of spontaneous abortion. However, interactions were observed between stressful work and maternal age over 32 years (p = 0.04), cigarette smoking (p = 0.02), and primigravidity (p = 0.06). Relative to the odds ratio for stressful work in you...

A case-control study of the relationship of serum cholesterol to large bowel cancer incidence was performed for 245 members of the Kaiser Permanente Medical Care Program in northern California. Five controls were obtained for each case, matched for age, sex, race, and time of examination. The mean serum cholesterol levels of cases were not significantly different from those of controls for all colorectal cancers, right-sided colon cancers, and left-sided colon cancers. Analysis by quartiles of serum cholesterol level suggested neither a direct nor an inverse relationship between serum cholesterol and large bowel cancer, and there was no evidence of a threshold value below which the risk of cancer increased. The study did not support the findings from other studies of an inverse relationship between serum cholesterol and large bowel cancer.

We examined breast cancer incidence in a cohort of about 69,000 women who were members of a large prepaid health plan in Northern California and who answered detailed questions about alcohol consumption from 1979 through 1984 as part of a voluntary multiphasic health checkup. Among white, black, and Hispanic women with no prior cancer, breast cancer had developed in 303 by the end of 1984. In analysis controlling only for age there was a progressive increase in breast cancer incidence with each higher level of reported alcohol consumption. In multivariate analyses controlled for age, race, body mass, and smoking, the relative risk at 1-2 drinks per day was 1.5 [95% confidence interval (CI) 1.0-2.3], at 3-5 drinks per day it was 1.5 (95% CI 0.8-2.8), and at 6 or more drinks per day it was 3.3 (95% CI 1.2-9.3). Past drinkers tended to have been heavier drinkers than current drinkers and had a relative risk of 2.2 (95% CI 1.2-3.9). Study of wine, beer, and liquor use did not suggest th...

The Pathways project studied the relationship of race/ethnicity, socioeconomic status, and culture to breast and cervical cancer screening. A multidisciplinary, multicultural team of investigators assessed the knowledge, attitudes, and practices of women from five race/ethnic groups (N = 4,228) and their barriers to screening utilization. A planning framework guided the development of interventions and analyses of the similarities and differences

To examine the association of breastfeeding or its duration with timing of girls' pubertal onset, and the role of BMI as a mediator in these associations. A population of 1,237 socio-economically and ethnically diverse girls, ages 6-8 years, was recruited across three geographic locations (New York City, Cincinnati, and the San Francisco Bay Area) in a prospective study of predictors of pubertal maturation. Breastfeeding practices were assessed using self-administered questionnaire/interview with the primary caregiver. Girls were seen on at least annual basis to assess breast and pubic hair development. The association of breastfeeding with pubertal timing was estimated using parametric survival analysis while adjusting for body mass index, ethnicity, birth-weight, mother's education, mother's menarcheal age, and family income. Compared to formula fed girls, those who were mixed-fed or predominantly breastfed showed later onset of breast development [hazard ratios 0.90 (95 % CI 0.75, 1.09) and 0.74 (95 % CI 0.59, 0.94), respectively]. Duration of breastfeeding was also directly associated with age at onset of breast development (p trend = 0.008). Associations between breastfeeding and pubic hair onset were not significant. In stratified analysis, the association of breastfeeding and later breast onset was seen in Cincinnati girls only. The association between breast feeding and pubertal onset varied by study site. More research is needed about the environments within which breastfeeding takes place in order to better understand whether infant feeding practices are a potentially modifiable risk factor that may influence age at onset of breast development and subsequent risk for disease in adulthood.

Individuals are increasingly involved in decisions about their health care. Shared decision making (SDM), an intervention in the clinical setting in which patients and providers collaborate in decision making, is an important approach for informing patients and involving them in their health care. However, SDM cannot bear the entire burden for informing and involving individuals. Population-oriented interventions to promote in-

Alcohol and tobacco consumption are closely correlated and published results on their association with breast cancer have not always allowed adequately for confounding between these exposures. Over 80% of the relevant information worldwide on alcohol and tobacco consumption and breast cancer were collated, checked and analysed centrally. Analyses included 58,515 women with invasive breast cancer and 95,067 controls from 53 studies. Relative risks of breast cancer were estimated, after stratifying by study, age, parity and, where appropriate, women's age when their first child was born and consumption of alcohol and tobacco. The average consumption of alcohol reported by controls from developed countries was 6.0 g per day, i.e. about half a unit/drink of alcohol per day, and was greater in ever-smokers than never-smokers, (8.4 g per day and 5.0 g per day, respectively). Compared with women who reported drinking no alcohol, the relative risk of breast cancer was 1.32 (1.19-1.45, P...

Hereditary nonpolyposis colorectal cancer (HNPCC), also known as Lynch syndrome, is a common autosomal domi- nant syndrome characterized by early age at onset, neoplas- tic lesions, and microsatellite instability (MSI). Because can- cers with MSI account for approximately 15% of all colorectal cancers and because of the need for a better understanding of the clinical and histologic manifestations of HNPCC,

The National Cancer Institute established the Cancer Genetics Network (CGN) to support collaborative investigations into the genetic basis of cancer susceptibility, explore mechanisms to integrate this new knowledge into medical practice, and identify ways of addressing the associated psychosocial, ethical, legal, and public health issues. The CGN has developed the complex infrastructure required to support the projects, including the establishment of guidelines and policies, uniform methods, standard questionnaires to be used by all of the centers, and a standard format for submission of data to the Informatics Center. Cancer patients and their family members have been invited to enroll and be included in a pool of potential study participants. The Information Technology Group is responsible for support of the design, implementation, and maintenance of the multicenter Network-wide research protocols. As of January 2004, the CGN contained data on 23,995 probands (participants) and 425,798 family members. As a resource for cancer genetic studies, the CGN has a large number of probands and first-degree relatives with and without cancer and with multiple ethnicities. Different study designs can be used including case-control, case-case, and family studies. The unique resources of the CGN are available for studies on cancer genetic susceptibility, translational research, and behavioral research. The CGN is now at a point where approved collaborators may have access to enrolled patients and their families for special studies, as well as to the clinical, environmental and family cancer history data banked in the Informatics Center.

Although commonly thought to be an effective method of screening for prostatic cancer, digital rectal examinations have yet to be shown by controlled study to help prevent advanced forms of this disease. 139 men with metastatic (stage D) prostatic cancer (cases) were compared with an equal number of matched men free of this condition (controls), with respect to rectal examinations recorded in the medical records up to, on average, 23 years before the cases' initial diagnosis of prostatic cancer. Cases and controls were members of a large health maintenance organisation in northern California. In the 10 years before initial diagnosis (excluding the last three months) the average number of examinations for routine screening (2.45 vs 2.52) or to evaluate intestinal or rectal symptoms (0.44 in both) were similar in cases and controls, respectively. After adjustment for racial differences, the relative risk of metastatic prostatic cancer for men with one or more screening rectal examinations compared with men with none was 0.9 with a 95% confidence interval of 0.5-1.7. Screening by routine digital rectal examination appears to have little if any effect in preventing metastatic prostatic cancer. If there is a small benefit, it will be difficult to demonstrate by conventional epidemiological study.

Epidemiologic studies of the relationship of alcohol consumption and breast cancer are generally in support of a direct association, although they are not entirely consistent. Studies supporting this finding have come from many countries around the world and from many investigators working with different study designs and types of populations. Ten of 16 case-control studies, and five of six cohort studies have produced results supporting a positive association. When evaluated together through techniques such as meta-analysis, these studies suggest that there is a 50% increase in breast cancer risk for women who average between one and two drinks per day. All types of alcoholic beverages seem to increase the risk, but there is little agreement between studies on dose-response relationships. No biologic explanation for alcohol's possible role in breast carcinogenesis is yet known, but several hypotheses have been advanced. In the absence of any effective method of primary breast cancer prevention, and because current risk factors explain so little of the incidence of this common cancer, alcohol consumption is of potential importance from both public health and etiologic perspectives. The possible protective effect of moderate alcohol consumption on heart disease should be considered when assessing the risks of this level of consumption on breast cancer.

The frequencies of kidney and urinary tract diseases (KUTD) in a defined population are presented. Rates of first hospitalizations with KUTD in the 5-year period, 1971 through 1975, were determined among over 1 million members of the Kaiser Foundation Health Plan (KFHP) of Northern California. Cases were identified from computer-stored ICDA-coded final discharge diagnoses for hospitalizations. Kidney and other urinary tract diseases were final diagnoses in 35.7 first hospitalizations per 10,000 members per year in the period 1971 through 1975. Disease-specific rates were highest for urinary tract infections, benign prostatic hypertrophy, and renal and ureteral calculi. The frequency of KUTD in this population is compared to available rates from other sources.

Differences in the epidemiology of lung cancer between Asians and non-Hispanic whites have brought to light the relative influences of genetic and environmental factors on lung cancer risk. We set out to describe the epidemiology of non-small cell lung cancer (NSCLC) among Asians living in California, and to explore the effects of acculturation on lung cancer risk by comparing lung cancer rates between U.S.-born and foreign-born Asians. Age-adjusted incidence rates of NSCLC were calculated for Chinese, Filipino, Japanese, Korean, Vietnamese, and South Asians in California between 1988 and 2003 using data from the California Cancer Registry. Incidence rates were calculated and stratified by sex and nativity. We analyzed population-based tobacco smoking prevalence data to determine whether differences in rates were associated with prevalence of tobacco smoking. Asians have overall lower incidence rates of NSCLC compared with whites (29.8 and 57.7 per 100,000, respectively). South Asians have markedly low rates of NSCLC (12.0 per 100,000). Foreign-born Asian men and women have an approximately 35% higher rate of NSCLC than U.S.-born Asian men and women. The incidence pattern by nativity is consistent with the population prevalence of smoking among Asian men; however, among women, the prevalence of smoking is higher among U.S.-born, which is counter to their incidence patterns. Foreign-born Asians have a higher rate of NSCLC than U.S.-born Asians, which may be due to environmental tobacco smoke or nontobacco exposures among women. South Asians have a remarkably low rate of NSCLC that approaches white levels among the U.S.-born. More studies with individual-level survey data are needed to identify the specific environmental factors associated with differential lung cancer risk occurring with acculturation among Asians.

The National Health Interview Survey (NHIS) has provided data about health behaviors at the national level since 1957. The 1987 and 1992 Cancer Control Supplements to the NHIS, along with other supplemental surveys administered intermittently on self-reported cancer-related behaviors, have contributed to important research and public health purposes. In this article, we reviewed 73 papers published between 1980 and 2001 that used NHIS data, including the first report from the 1998 NHIS, to examine what has been learned from past surveys. Our goal was to facilitate future analyses of recently released data on cancer screening practices from the Cancer Control Supplement to the 2000 NHIS, which is now known as the Cancer Control Module. We categorized the papers according to which of the following three study approaches they used: trends in screening rates, correlates of these rates with factors that may influence screening, and linkages or comparisons of NHIS data with other surveys or sources of information. We summarize knowledge gained in cancer screening for each of these three categories and identify areas that could benefit from more research. We highlight some of the new information available for the first time on the Cancer Control Module of the 2000 NHIS as fresh opportunities for cancer control research. Finally, we describe how the Cancer Control Supplements to the NHIS are integrated with the objectives of and developments in national cancer surveillance research that have emerged from federal planning efforts and collaborations with national partners in cancer surveillance in recent years.

Five small case-control studies have examined the relationship between exposure to organochlorines and the risk of breast cancer and have found inconsistent results. In these studies, organochlorine levels in breast cancer patients were measured after (or at most 6 months before) diagnosis. We tested the hypothesis that organochlorines are a risk factor for breast cancer, using prospectively gathered data on serum levels of DDE [1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene] (the main metabolite of the pesticide DDT [2,2-bis(p-chlorophenyl)-1,1,1-trichloroethane]) and polychlorinated biphenyls (PCBs). Study subjects belonged to a cohort of 57,040 women (46,629 white, 8123 black, and 2288 Asian) from the San Francisco Bay Area who took a multiphasic health examination, independent of concern about risk of breast cancer, in the late 1960s. At that time, a sample of blood was obtained, then frozen and stored. Follow-up was through December 31, 1990. We conducted a nested case-control study of 150 case patients and 150 matched control subjects. A random sample of 50 women per racial/ethnic group who had been diagnosed with breast cancer more than 6 months after the multiphasic examination (mean follow-up = 14.2 years) was selected, and each case patient was matched to a cancer-free control subject. Matched analyses found no differences in the case patients' and control subjects' serum levels of DDE (mean difference = 0.2 parts per billion [ppb]; 95% confidence interval [CI] = -6.7, 7.2) or PCBs (mean difference = -0.4 ppb; 95% CI = -0.8, 0.1). DDE levels, however, tended to be higher among black case patients compared with black controls (mean difference = 5.7 ppb; 95% CI = -3.3, 14.8), and PCBs were lower among white case patients compared with white controls (mean difference = -0.6 ppb; 95% CI = -1.2, -0.1). Organochlorine levels were significantly higher among black and Asian women compared with white women. The mean difference for DDE was 11.0 ppb for black women (95% CI = 4.3, 17.6) and 12.6 ppb for Asian women (95% CI = 6.0, 19.2); for PCBs, the respective differences were 0.8 ppb for black women (95% CI = 0.2, 1.4) and 1.4 ppb for Asian women (95% CI = 0.8, 1.9). The results were not altered by adjusting for relevant confounders, and the lack of association between exposure to organochlorines and breast cancer was present regardless of length of follow-up, year of diagnosis, or the case patient's menopausal and estrogen-receptor status. The data do not support the hypothesis that exposure to DDE and PCBs increases risk of breast cancer. Future investigations must consider the biologic mechanisms involved and variations in exposure to chemical pollutants and of breast cancer incidence rates among diverse groups of women.

To collect information regarding knowledge about and attitudes toward cancer in a sample of adult health plan members, self-identified as Latino or Anglo. Cross-sectional survey. Prepaid health plan. A random sample of 844 Latinos (mean age, 50.5 years) and 510 Anglos (51.8 years) completed the interview. Latinos were significantly more likely than Anglos to think that sugar substitutes (58% vs 42%), bruises from being hit (53% vs 34%), microwave ovens (47% vs 23%), eating pork (31% vs 11%), eating spicy foods (15% vs 8%), breast-feeding (14% vs 6%), and antibiotics (32% vs 12%) could cause cancer (P < .001 for each). Compared with Anglos, Latinos more often misidentified constant dizziness (39% vs 25%) and arthralgias (35% vs 20%) as being symptoms of cancer. A higher proportion of Latinos believed that having cancer is like getting a death sentence (46% vs 26%), that cancer is God's punishment (7% vs 2%), that there is very little one can do to prevent getting cancer (26% vs 18%), that it is uncomfortable to touch someone with cancer (13% vs 8%), and that they would rather not know if they had incurable cancer (35% vs 23%; P < .001 for each). Latino ethnicity was a significant predictor of these knowledge and attitude items in multivariate logistic regression models adjusted for sex, education, age, employment, marital status, county of residence, and self-perceived health status. We conclude that misconceptions about cancer are more prevalent among Latinos than Anglos and that selected attitudes about cancer among Latinos fit a cultural theme of fatalismo. These data can enable development of culturally appropriate cancer control interventions for Latinos.