Pregnancy plasma glucose levels exceeding the American Diabetes Association thresholds, but below the National Diabetes Data Group thresholds for gestational diabetes mellitus, are related to the risk of neonatal macrosomia, hypoglycaemia and hyperbilirubinaemia.

AIMS/HYPOTHESIS: Gestational diabetes mellitus (GDM) is a risk factor for perinatal complications. In several countries, the criteria for the diagnosis of GDM have been in flux, the American Diabetes Association (ADA) thresholds recommended in 2000 being lower than those of the National Diabetes Data Group (NDDG) that have been in use since 1979. We sought to determine the extent to which infants of women meeting only the ADA criteria for GDM are at increased risk of neonatal complications. MATERIALS AND METHODS: In a multiethnic cohort of 45,245 women who did not meet the NDDG criteria and were not treated for GDM, we conducted nested case-control studies of three complications of GDM that occurred in their infants: macrosomia (birthweight >4,500 g, n = 494); hypoglycaemia (plasma glucose <2.2 mmo/l, n = 488); and hyperbilirubinaemia (serum bilirubin > or =342 micromol/l (20 mg/dl), n = 578). We compared prenatal glucose levels of the mothers of these infants and mothers of 884 control infants. RESULTS: Women with GDM by ADA criteria only (two or more glucose values exceeding the threshold) had an increased risk of having an infant with macrosomia (odds ratio OR = 3.40, 95% CI = 1.55-7.43), hypoglycaemia (OR = 2.61, 95% CI = 0.99-6.92) or hyperbilirubinaemia (OR = 2.22, 95% CI = 0.98-5.04). Glucose levels 1 h after the 100-g glucose challenge that exceeded the ADA threshold were particularly strongly associated with each complication. CONCLUSIONS/INTERPRETATION: These results lend support to the ADA recommendations and highlight the importance of the 1-h glucose measurement in a diagnostic test for GDM.

Developing a prediction rule from automated clinical databases to identify high-risk patients in a large population with diabetes.

OBJECTIVE: To develop and validate a prediction rule for identifying diabetic patients at high short-term risk of complications using automated data in a large managed care organization. RESEARCH DESIGN AND METHODS: Retrospective cohort analyses were performed in 57,722 diabetic members of Kaiser Permanente, Northern California, aged > or =19 years. Data from 1994 to 1995 were used to model risk for macro- and microvascular complications (n = 3,977), infectious complications (n = 1,580), and metabolic complications (n = 316) during 1996. Candidate predictors (n = 36) included prior inpatient and outpatient diagnoses, laboratory records, pharmacy records, utilization records, and survey data. Using split-sample validation, the risk scores derived from logistic regression models in half of the population were evaluated in the second half. Sensitivity, positive predictive value, and receiver operating characteristics curves were used to compare scores obtained from full models to those derived using simpler approaches. RESULTS: History of prior complications or related outpatient diagnoses were the strongest predictors in each complications set. For patients without previous events, treatment with insulin alone, serum creatinine > or =1.3 mg/dl, use of two or more antihypertensive medications, HbA(1c) >10%, and albuminuria/microalbuminuria were independent predictors of two or all three complications. Several risk scores derived from multivariate models were more efficient than simply targeting patients with elevated HbA(1c) levels for identifying high-risk patients. CONCLUSIONS: Simple prediction rules based on automated clinical data are useful in planning care management for populations with diabetes.

Self-monitoring of blood glucose levels and glycemic control: the Northern California Kaiser Permanente Diabetes registry.

PURPOSE: We sought to evaluate the effectiveness of self-monitoring blood glucose levels to improve glycemic control. SUBJECTS AND METHODS: A cohort design was used to assess the relation between self-monitoring frequency (1996 average daily glucometer strip utilization) and the first glycosylated hemoglobin (HbA1c) level measured in 1997. The study sample included 24,312 adult patients with diabetes who were members of a large, group model, managed care organization. We estimated the difference between HbA1c levels in patients who self-monitored at frequencies recommended by the American Diabetes Association compared with those who monitored less frequently or not at all. Models were adjusted for age, sex, race, education, occupation, income, duration of diabetes, medication refill adherence, clinic appointment "no show" rate, annual eye exam attendance, use of nonpharmacological (diet and exercise) diabetes therapy, smoking, alcohol consumption, hospitalization and emergency room visits, and the number of daily insulin injections. RESULTS: Self-monitoring among patients with type 1 diabetes (> or = 3 times daily) and pharmacologically treated type 2 diabetes (at least daily) was associated with lower HbA1c levels (1.0 percentage points lower in type 1 diabetes and 0.6 points lower in type 2 diabetes) than was less frequent monitoring (P < 0.0001). Although there are no specific recommendations for patients with nonpharmacologically treated type 2 diabetes, those who practiced self-monitoring (at any frequency) had a 0.4 point lower HbA1c level than those not practicing at all (P < 0.0001). CONCLUSION: More frequent self-monitoring of blood glucose levels was associated with clinically and statistically better glycemic control regardless of diabetes type or therapy. These findings support the clinical recommendations suggested by the American Diabetes Association.

Class inequalities in women's health: combined impact of childhood and adult social class--a study of 630 US women.

To assess contributions of childhood and adult social class to class gradients in women's health, the authors used gender-neutral household measures of class position in a retrospective cohort study of 630 women enrolled in Examination II of the Kaiser Permanente Women Twins Study (1989-1990, Oakland, CA). The age-adjusted odds of reporting fair or poor health was 2.3 times higher (95% confidence interval (CI)=1.2-4.1), using adult class measures, among women categorized as working class vs non-working class/professional. When stratified by childhood social class, however, the elevated risk of fair/poor health among adult working class compared to non-working class/professional women was evident only among those with a non-working class/professional childhood. Similarly, a working class tendency (based on adult class position) towards elevated levels of low density lipoprotein (LDL) cholesterol (odds ratio (OR)=1.5, 95% CI=0.9-2.7) and post-load glucose (OR=1.8, 95% CI=1.0-3.3) was apparent only among women who were non-working class in childhood. These results indicate that both childhood and adult class position influence class gradients in women's health in the United States.

Hormone replacement therapy is associated with better glycemic control in women with type 2 diabetes: The Northern California Kaiser Permanente Diabetes Registry.

OBJECTIVE: In women with diabetes, the changes that accompany menopause may further diminish glycemic control. Little is known about how hormone replacement therapy (HRT) affects glucose metabolism in diabetes. The aim of this study was to examine whether HbA(1c) levels varied by current HRT among women with type 2 diabetes. RESEARCH DESIGN AND METHODS: In a cohort of 15,435 women with type 2 diabetes who were members of a health maintenance organization, HbA(1c) and HRT were assessed by reviewing records in the health plan's computerized laboratory and pharmacy systems. Sociodemographic and clinical information were collected by survey. RESULTS: The mean age was 64.7 years (SD +/- 8.7). The study cohort comprised 55% non-Hispanic whites, 14% non-Hispanic blacks, 12% Hispanics, 11% Asians, 4% "other" ethnic groups, and 4% with missing ethnicity data. Current HRT was observed in 25% of women. HbA(1c) levels were significantly lower in women currently using HRT than in women not using HRT (age-adjusted mean +/- SE: 7.9 +/- 0.03 vs. 8.5 +/- 0.02, respectively, P = 0.0001). No differences in HbA(1c) level were observed between women using unopposed estrogens and women using opposed estrogens. In a Generalized Estimating Equation model, which took into account patient clustering within physician and adjusted for age, ethnicity, education, obesity, hypoglycemic therapy, diabetes duration, self-monitoring of blood glucose, and exercise, HRT remained significantly and independently associated with decreased HbA(1c) levels (P = 0.0001). CONCLUSIONS: HRT was independently associated with decreased HbA(1c) level. Clinical trials will be necessary to understand whether HRT may improve glycemic control in women with diabetes.

Glycemic control and heart failure among adult patients with diabetes.

BACKGROUND: Glycemic control is associated with microvascular events, but its effect on the risk of heart failure is not well understood. We examined the association between hemoglobin (Hb) A(Ic) and the risk of heart failure hospitalization and/or death in a population-based sample of adult patients with diabetes and assessed whether this association differed by patient sex, heart failure pathogenesis, and hypertension status. METHODS AND RESULTS: A cohort design was used with baseline between January 1, 1995, and June 30, 1996, and follow-up through December 31, 1997 (median 2.2 years). Participants were 25 958 men and 22 900 women with (predominantly type 2) diabetes, >/=19 years old, with no known history of heart failure. There were a total of 935 events (516 among men; 419 among women). After adjustment for age, sex, race/ethnicity, education level, cigarette smoking, alcohol consumption, hypertension, obesity, use of beta-blockers and ACE inhibitors, type and duration of diabetes, and incidence of interim myocardial infarction, each 1% increase in Hb A(Ic) was associated with an 8% increased risk of heart failure (95% CI 5% to 12%). An Hb A(Ic) >/=10, relative to Hb A(Ic) <7, was associated with 1.56-fold (95% CI 1.26 to 1.93) greater risk of heart failure. Although the association was stronger in men than in women, no differences existed by heart failure pathogenesis or hypertension status. CONCLUSIONS: These results confirm previous evidence that poor glycemic control may be associated with an increased risk of heart failure among adult patients with diabetes.

Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study.

CONTEXT: Atrial fibrillation is the most common arrhythmia in elderly persons and a potent risk factor for stroke. However, recent prevalence and projected future numbers of persons with atrial fibrillation are not well described. OBJECTIVE: To estimate prevalence of atrial fibrillation and US national projections of the numbers of persons with atrial fibrillation through the year 2050. DESIGN, SETTING, AND PATIENTS: Cross-sectional study of adults aged 20 years or older who were enrolled in a large health maintenance organization in California and who had atrial fibrillation diagnosed between July 1, 1996, and December 31, 1997. MAIN OUTCOME MEASURES: Prevalence of atrial fibrillation in the study population of 1.89 million; projected number of persons in the United States with atrial fibrillation between 1995-2050. RESULTS: A total of 17 974 adults with diagnosed atrial fibrillation were identified during the study period; 45% were aged 75 years or older. The prevalence of atrial fibrillation was 0.95% (95% confidence interval, 0.94%-0.96%). Atrial fibrillation was more common in men than in women (1.1% vs 0.8%; P<.001). Prevalence increased from 0.1% among adults younger than 55 years to 9.0% in persons aged 80 years or older. Among persons aged 50 years or older, prevalence of atrial fibrillation was higher in whites than in blacks (2.2% vs 1.5%; P<.001). We estimate approximately 2.3 million US adults currently have atrial fibrillation. We project that this will increase to more than 5.6 million (lower bound, 5.0; upper bound, 6.3) by the year 2050, with more than 50% of affected individuals aged 80 years or older. CONCLUSIONS: Our study confirms that atrial fibrillation is common among older adults and provides a contemporary basis for estimates of prevalence in the United States. The number of patients with atrial fibrillation is likely to increase 2.5-fold during the next 50 years, reflecting the growing proportion of elderly individuals. Coordinated efforts are needed to face the increasing challenge of optimal stroke prevention and rhythm management in patients with atrial fibrillation.

Beta agonists, inhaled steroids, and the risk of intensive care unit admission for asthma.

Although inhaled corticosteroid (ICS) use is associated with a decreased risk of hospitalization for asthma, the impact of ICS on the risk of life-threatening asthma exacerbation is less clear. The effect of ICS and inhaled beta agonist (IBA) dispensing on the risk of intensive care unit admission for asthma, a surrogate for life-threatening exacerbation, is evaluated. Using computerized International classification of diseases (ICD)-9 discharge diagnoses, a cohort of all 2,344 adult Northern California members of a health maintenance organization hospitalized for asthma over a 2-yr period were identified. Computerized pharmacy data was used to ascertain asthma medications dispensed during the 3-,6-, and 12-month intervals preceding index hospitalization for asthma. During the 3-months preceding hospitalization, a minority of subjects had no IBA units dispensed (34%), with 14% receiving low level (1 unit), 20% medium level (2-3 units), and 32% high level (> or = 4 units) therapy. A substantial proportion received no ICS units (55%), whereas 13% had low, 16% medium, and 15% high level therapy. In multiple logistic regression analysis, high level IBA use was associated with a greater risk of intensive care unit (ICU) admission for asthma after controlling for asthma severity. There was no relationship, however, between low or medium level IBA use and ICU admission. Conversely, medium level and high level ICS use were associated with a reduced risk of ICU admission. Analysing 6- and 12-month medication dispensing data, similar risk patterns were observed. Inhaled corticosteroid dispensing was associated with reduced risk of intensive care unit admission among adults hospitalized for asthma, whereas the opposite applied for high dose beta agonist usage. This suggests that ICS prescription to adults with moderate-to-severe asthma could reduce the risk of life-threatening exacerbation.

Effect of physician and patient gender concordance on patient satisfaction and preventive care practices.

OBJECTIVE: To explore the role of the gender of the patient and the gender of the physician in explaining differences in patient satisfaction and patient-reported primary care practice. DESIGN: Crosssectional mailed survey [response rate of 71%]. SETTING: A large group-model Health Maintenance Organization (HMO) in northern California. PATIENTS/PARTICIPANTS: Random sample of HMO members aged 35 to 85 years with a primary care physician. The respondents (N = 10,205) were divided into four dyads: female patients of female doctors; male patients of female doctors; female patients of male doctors; and male patients of male doctors. Patients were also stratified on the basis of whether they had chosen their physician or had been assigned. MEASUREMENTS AND MAIN RESULTS: Among patients who chose their physician, females who chose female doctors were the least satisfied of the four groups of patients for four of five measures of satisfaction. Male patients of female physicians were the most satisfied. Preventive care and health promotion practices were comparable for male and female physicians. Female patients were more likely to have chosen their physician than males, and were much more likely to have chosen female physicians. These differences were not seen among patients who had been assigned to their physicians and were not due to differences in any of the measured aspects of health values or beliefs. CONCLUSIONS: Our study revealed differences in patient satisfaction related to the gender of the patient and of the physician. While our study cannot determine the reasons for these differences, the results suggest that patients who choose their physician may have different expectations, and the difficulty of fulfi11ing these expectations may present particular challenges for female physicians.

The outcome and cost of alcohol and drug treatment in an HMO: day hospital versus traditional outpatient regimens.

OBJECTIVE: To compare outcome and cost-effectiveness of the two primary addiction treatment options, day hospitals (DH) and traditional outpatient programs (OP) in a managed care organization, in a population large enough to examine patient subgroups. DATA SOURCES: Interviews with new admissions to a large HMO's chemical dependency program in Sacramento, California between April 1994 and April 1996, with follow-up interviews eight months later. Computerized utilization and cost data were collected from 1993 to 1997. STUDY DESIGN: Design was a randomized control trial of adult patients entering the HMO's alcohol and drug treatment program (N = 668). To examine the generalizability of findings as well as self-selection factors, we also studied patients presenting during the same period who were unable or unwilling to be randomized (N = 405). Baseline interviews characterized type of substance use, addiction severity, psychiatric status, and motivation. Follow-up interviews were conducted at eight months following intake. Breathanalysis and urinalysis were conducted. Program costs were calculated. DATA COLLECTION: Interview data were merged with computerized utilization and cost data. PRINCIPAL FINDINGS: Among randomized subjects, both study arms showed significant improvement in all drug and alcohol measures. There were no differences overall in outcomes between DH and OP, but DH subjects with midlevel psychiatric severity had significantly better outcomes, particularly in regard to alcohol abstinence (OR = 2.4; 95% CI = 1.2, 4.9). The average treatment costs were $1,640 and $895 for DH and OP programs, respectively. In the midlevel psychiatric severity group, the cost of obtaining an additional person abstinent from alcohol in the DH cohort was approximately $5,464. Among the 405 self-selected subjects, DH was related to abstinence (OR = 2.1; 95% CI = 1.3, 3.5). CONCLUSIONS: Although significant benefits of the DH program were not found in the randomized study, DH treatment was associated with better outcomes in the self-selected group. However, for subjects with mid-level psychiatric severity in both the randomized and self-selected samples, the DH program produced higher rates of abstention and was more cost-effective. Self-selection in studies that randomize patients to services requiring very different levels of commitment may be important in interpreting findings for clinical practice.

Implications of stroke risk criteria on the anticoagulation decision in nonvalvular atrial fibrillation: the Anticoagulation and Risk Factors in Atrial Fibrillation (ATRIA) study.

BACKGROUND: Warfarin dramatically reduces the risk of stroke in patients with nonvalvular atrial fibrillation (NVAF) but increases the likelihood of bleeding. Accurately identifying patients who need anticoagulation is critical. We assessed the potential impact of prominent stroke risk classification schemes on this decision in a large sample of patients with NVAF. METHODS AND RESULTS: We used clinical and electrocardiographic databases to identify 13 559 ambulatory patients with NVAF from July 1996 through December 1997. We compared the proportion of patients classified as having a low enough stroke risk to receive aspirin using published criteria from the Atrial Fibrillation Investigators (AFI), American College of Chest Physicians (ACCP), and the Stroke Prevention in Atrial Fibrillation Investigators (SPAF). In this cohort, AFI criteria classified 11% as having a low stroke risk, compared with 23% for ACCP and 29% for SPAF (kappa range, 0.44 to 0.85). This 2- to-3-fold increase in low stroke risk patients by ACCP and SPAF criteria primarily resulted from the inclusion of many older subjects (65 to 75 years+/-men >75 years) with no additional clinical stroke risk factors. CONCLUSIONS: The age threshold for assigning an increased stroke risk has a dramatic impact on whether to recommend warfarin in populations of patients with NVAF. Large, prospective studies with many stroke events are needed to precisely determine the relationship of age to stroke risk in AF and to identify which AF subgroups are at a sufficiently low stroke risk to forego anticoagulation.

Self-monitoring of blood glucose: language and financial barriers in a managed care population with diabetes.

OBJECTIVE: Self-monitoring of blood glucose (SMBG) is a cornerstone of diabetes care, but little is known about barriers to this self-care practice. RESEARCH DESIGN AND METHODS: This cross-sectional study examines SMBG practice patterns and barriers in 44,181 adults with pharmacologically treated diabetes from the Kaiser Permanente Northern California Region who responded to a health survey (83% response rate). The primary outcome is self-reported frequency of SMBG. RESULTS: Although most patients reported some level of SMBG monitoring, 60% of those with type 1 diabetes and 67% of those with type 2 diabetes reported practicing SMBG less frequently than recommended by the American Diabetes Association (three to four times daily for type 1 diabetes, and once daily for type 2 diabetes treated pharmacologically). Significant independent predictors of nonadherent practice of SMBG included longer time since diagnosis, less intensive therapy, male sex, age, belonging to an ethnic minority, having a lower education and neighborhood income, difficulty communicating in English, higher out-of-pocket costs for glucometer strips (especially for subjects with lower incomes), smoking, and excessive alcohol consumption. CONCLUSIONS: Considerable gaps persist between actual and recommended SMBG practices in this large managed care organization. A somewhat reduced SMBG frequency in subjects with linguistic barriers, some ethnic minorities, and subjects with lower education levels suggests the potential for targeted, culturally sensitive, multilingual health education. The somewhat lower frequency of SMBG among subjects paying higher out-of-pocket expenditures for strips suggests that removal of financial barriers by providing more comprehensive coverage for these costs may enhance adherence to recommendations for SMBG.

The cost of health conditions in a health maintenance organization.

In this retrospective cohort analysis of all adults who were members of Kaiser Permanente, Northern California, between July 1995 and June 1996 (N = 2,076,303), the authors estimated the prevalence, average annual costs per person, and percentage of total direct medical expenditures attributable to each of 25 chronic and acute conditions. Ordinary least squares regression was used to adjust for age, gender, and comorbidities. The costs attributable to the 25 conditions accounted for 78 percent of the health maintenance organization's total direct medical expense for this age-group. Injury accounted for a higher proportion (11.5 percent) of expenditures than any other single condition. Three cardiovascular conditions--ischemic heart disease, hypertension, and congestive heart failure--together accounted for 17 percent of direct medical expense and separately accounted for 6.8 percent, 5.7 percent, and 4.0 percent, respectively. Renal failure ($22,636), colorectal cancer ($10,506), pneumonia ($9,499), and lung cancer ($8,612) were the most expensive conditions per person per year.

Racial and ethnic differences in a patient survey: patients' values, ratings, and reports regarding physician primary care performance in a large health maintenance organization.

BACKGROUND: Few studies have investigated the influence of race and/or ethnicity on patients' ratings of quality of care. None have incorporated patients' values and beliefs regarding medical care in assessing these possible differences. OBJECTIVES: We explored whether patients' values, ratings, and reports regarding physicians' primary care performance differed by race and/or ethnicity. RESEARCH DESIGN: This was a cross-sectional, mailed patient survey. SUBJECTS: The study subjects were adult primary care patients in a large health maintenance population (7,747 whites, 836 blacks, 710 Latinos, and 1,007 Asians). MEASURES AND METHODS: Ratings of the following dimensions of primary care were measured: technical competence, communication, accessibility, prevention and health promotion, and overall satisfaction. Patients' values regarding these dimensions and their confidence in medical care were measured. Multivariate analyses yielded associations of race/ethnicity with satisfaction and with reports of prevention services received. RESULTS: For 7 of the 10 dimensions of primary care measured, Asians rated physician performance significantly less favorably than did whites, including differences among Asian ethnic subgroups. Latinos rated physicians' accessibility less favorably than did whites. Blacks rated physicians' psychosocial and lifestyle health promotion practices higher than did whites. No differences were found in patient reports of prevention services received, except Pacific Islanders reported receiving significantly more prevention services than whites. CONCLUSIONS: In a large HMO population, significant differences were found by race and ethnicity, and among Asian ethnic subgroups, in levels of patient satisfaction with primary care. These findings may represent actual differences in quality of care or variations in patient perceptions, patient expectations, and/or questionnaire response styles. More research is needed to assess, in accurate and culturally appropriate ways, whether health plans are meeting the needs of all enrollees.

Evaluating hypertension control in a managed care setting.

BACKGROUND: We conducted a retrospective cohort study on a random sample of adult patients with hypertension in a large health maintenance organization to assess the feasibility of documenting blood pressure (BP) control and to compare different measures for defining BP control. METHODS: Three criteria for BP control were assessed: systolic BP less than 140 mm Hg; diastolic BP less than 90 mm Hg; and combined BP control, with systolic BP less than 140 mm Hg and diastolic BP less than 90 mm Hg. Four methods of assessing hypertension control by the above criteria were examined: proportion of patients with BP under control at 75% and 50% or more of their office visits; the mean of all pressures during the study period; and the BP from the last visit during the study period. RESULTS: The proportion of patients meeting each criterion for control was similar whether we used the mean BP for all visits, the last recorded BP, or control at 50% or more of visits. Control rates were substantially lower when the more stringent assessment, 75% of visits, was used. The proportion of patients with combined BP control at 75% or more of their visits was half that of the other methods. CONCLUSIONS: In this health maintenance organization population, results with the use of the simplest approach, the last BP measurement recorded, were similar to results with the mean BP. Our findings indicate that evaluation of BP control in a large health maintenance organization will find substantial room for improvement, and clinicians should be encouraged to be more aggressive in their management of hypertension, especially with regard to the systolic BP, which until recent years has been underemphasized.

Warfarin use among ambulatory patients with nonvalvular atrial fibrillation: the anticoagulation and risk factors in atrial fibrillation (ATRIA) study.

BACKGROUND: Warfarin dramatically reduces the risk for ischemic stroke in nonvalvular atrial fibrillation, but its use among ambulatory patients with atrial fibrillation has not been widely studied. OBJECTIVE: To assess the rates and predictors of warfarin use in ambulatory patients with nonvalvular atrial fibrillation. DESIGN: Cross-sectional study. SETTING: Large health maintenance organization. PATIENTS: 13428 patients with a confirmed ambulatory diagnosis of nonvalvular atrial fibrillation and known warfarin status between 1 July 1996 and 31 December 1997. MEASUREMENTS: Data from automated pharmacy, laboratory, and clinical-administrative databases were used to determine the prevalence and determinants of warfarin use in the 3 months before or after the identified diagnosis of atrial fibrillation. RESULTS: Of 11082 patients with nonvalvular atrial fibrillation and no known contraindications, 55% received warfarin. Warfarin use was substantially lower in patients who were younger than 55 years of age (44.3%) and those who were 85 years of age or older (35.4%). Only 59.3% of patients with one or more risk factors for stroke and no contraindications were receiving warfarin. Among a subset of "ideal" candidates to receive warfarin (persons 65 to 74 years of age who had no contraindications and had previous stroke, hypertension, or both), 62.1% had evidence of warfarin use. Among our entire cohort, the strongest predictors of receiving warfarin were previous stroke (adjusted odds ratio, 2.55 [95% CI, 2.23 to 2.92]), heart failure (odds ratio, 1.63 [CI, 1.51 to 1.77]), previous intracranial hemorrhage (odds ratio, 0.33 [CI, 0.21 to 0.52]), age 85 years or older (odds ratio, 0.35 [CI, 0.31 to 0.40]), and previous gastrointestinal hemorrhage (odds ratio, 0.47 [CI, 0.40 to 0.57]). CONCLUSIONS: In a large, contemporary cohort of ambulatory patients with atrial fibrillation who received care within a health maintenance organization, warfarin use was considerably higher than in other reported studies. Although the reasons why physicians did not prescribe warfarin could not be elucidated, many apparently eligible patients with atrial fibrillation and at least one additional risk factor for stroke, especially hypertension, did not receive anticoagulation. Interventions are needed to increase the use of warfarin for stroke prevention among appropriate candidates.