Predicting hospital readmission in Medicaid patients with diabetes using administrative and claims data.

OBJECTIVES: Readmission is common and costly for hospitalized Medicaid patients with diabetes. We aimed to develop a model predicting risk of 30-day readmission in Medicaid patients with diabetes hospitalized for any cause. STUDY DESIGN: Using 2016-2019 Medicaid claims from 7 US states, we identified patients who (1) had a diagnosis of diabetes or were prescribed any diabetes drug, (2) were hospitalized for any cause, and (3) were discharged to home or to a nonhospice facility. For each encounter, we assessed whether the patient was readmitted within 30 days of discharge. METHODS: Applying least absolute shrinkage and selection operator variable selection, we included demographic data and claims history in a logistic regression model to predict 30-day readmission. We evaluated model fit graphically and measured predictive accuracy by the area under the receiver operating characteristic curve (AUROC). RESULTS: Among 69,640 eligible patients, there were 129,170 hospitalizations, of which 29,410 (22.8%) were 30-day readmissions. The final model included age, sex, age-sex interaction, past diagnoses, US state of admission, number of admissions in the preceding year, index admission type, index admission diagnosis, discharge status, length of stay, and length of stay-sex interaction. The observed vs predicted plot showed good fit. The estimated AUROC of 0.761 was robust in analyses that assessed sensitivity to a range of model assumptions. CONCLUSIONS: Our model has moderate power for identifying hospitalized Medicaid patients with diabetes who are at high risk of readmission. It is a template for identifying patients at risk of readmission and for adjusting comparisons of 30-day readmission rates among sites or over time.

Survival of cancer survivors with a new pancreatic cancer diagnosis.

BACKGROUND: Persons newly diagnosed with pancreas cancer and who have survived a previous cancer are often excluded from clinical trials, despite limited evidence about their prognosis. We examined the association between previous cancer and overall survival. METHODS: This US population-based cohort study included older adults (aged ≥66 years) diagnosed with pancreas cancer between 2005 and 2015 in the linked Surveillance, Epidemiology, and End Results-Medicare data. We used Cox proportional hazards models to estimate stage-specific effects of previous cancer on overall survival, adjusting for sociodemographic, treatment, and tumor characteristics. RESULTS: Of 32,783 patients, 18.7% were previously diagnosed with another cancer. The most common previous cancers included prostate (29.0%), breast (18.9%), or colorectal (9.7%) cancer. More than half of previous cancers (53.9%) were diagnosed 5 or more years prior to pancreas cancer diagnosis or at an in situ or localized stage (47.8%). The proportions of patients surviving 1, 3, and 5 years after pancreas cancer were nearly identical for those with and without previous cancer. Median survival in months was as follows for those with and without previous cancer respectively: 7 versus 8 (Stage 0/I), 10 versus 10 (Stage II), 7 versus 7 (Stage III), and 3 versus 2 (Stage IV). Cox models indicated that patients with previous cancer had very similar or statistically equivalent survival to those with no previous cancer. CONCLUSIONS: Given nearly equivalent survival compared to those without previous cancer, cancer survivors newly diagnosed with pancreas cancer should be considered for inclusion in pancreas cancer clinical trials.

Survival of women diagnosed with breast cancer and who have survived a previous cancer.

PURPOSE: Many women diagnosed with breast cancer have survived previous cancer; yet little is known about the impact of previous cancer on overall and cancer-specific survival. METHODS: This population-based cohort study using SEER-Medicare data included women (age ≥ 66 years) diagnosed with breast cancer between 2005 and 2015. Separately by breast cancer stage, we estimated effect of previous cancer on overall survival using Cox regression and on cause-specific survival using competing risk regression; all survival analyses adjusted for covariates. RESULTS: Of 138,576 women diagnosed with breast cancer, 8% had a previous cancer of another organ site, most commonly colorectal or uterine cancer or melanoma. Many of these women (46.3%) were diagnosed within 5 years of breast cancer. For all breast cancer stages except IV wherein there was no difference, women with vs. without previous cancer had worse overall survival. This survival disadvantage was driven by deaths due to the previous cancer and other causes. In contrast, women with previous cancer generally had favorable breast-cancer-specific survival, although this varied by stage. Overall survival varied by previous cancer type, timing, and stage; previous lung cancer, cancer diagnosed within 1 year of incident breast cancer, and previous cancer at a distant stage were associated with the worst survival. In contrast, women with a previous melanoma had equivalent overall survival to women without previous cancer. CONCLUSION: We observed variable impact of previous cancer on overall and breast-cancer-specific survival depending on breast cancer stage at diagnosis and the type, timing, and stage of previous cancer.

Estimating cost-effectiveness from claims and registry data with measured and unmeasured confounders.

The analysis of observational data to determine the cost-effectiveness of medical treatments is complicated by the need to account for skewness, censoring, and the effects of measured and unmeasured confounders. We quantify cost-effectiveness as the Net Monetary Benefit (NMB), a linear combination of the treatment effects on cost and effectiveness that denominates utility in monetary terms. We propose a parametric estimation approach that describes cost with a Gamma generalized linear model and survival time (the canonical effectiveness variable) with a Weibull accelerated failure time model. To account for correlation between cost and survival, we propose a bootstrap procedure to compute confidence intervals for NMB. To examine sensitivity to unmeasured confounders, we derive simple approximate relationships between naïve parameters, assuming only measured confounders, and the values those parameters would take if there was further adjustment for a single unmeasured confounder with a specified distribution. A simulation study shows that the method returns accurate estimates for treatment effects on cost, survival, and NMB under the assumed model. We apply our method to compare two treatments for Stage II/III bladder cancer, concluding that the NMB is sensitive to hypothesized unmeasured confounders that represent smoking status and personal income.

Estimating lead-time bias in lung cancer diagnosis of patients with previous cancers.

Surprisingly, survival from a diagnosis of lung cancer has been found to be longer for those who experienced a previous cancer than for those with no previous cancer. A possible explanation is lead-time bias, which, by advancing the time of diagnosis, apparently extends survival among those with a previous cancer even when they enjoy no real clinical advantage. We propose a discrete parametric model to jointly describe survival in a no-previous-cancer group (where, by definition, lead-time bias cannot exist) and in a previous-cancer group (where lead-time bias is possible). We model the lead time with a negative binomial distribution and the post-lead-time survival with a linear spline on the logit hazard scale, which allows for survival to differ between groups even in the absence of bias; we denote our model Logit-Spline/Negative Binomial. We fit Logit-Spline/Negative Binomial to a propensity-score matched subset of the Surveillance, Epidemiology, and End Results-Medicare linked data set, conducting sensitivity analyses to assess the effects of key assumptions. With lung cancer-specific death as the end point, the estimated mean lead time is roughly 11 months for stage I&II patients; with overall survival, it is roughly 3.4 months in stage I&II. For patients with higher-stage lung cancers, the mean lead time is 1 month or less for both outcomes. Accounting for lead-time bias reduces the survival advantage of the previous-cancer group when one exists, but it does not nullify it in all cases.

Adaptive parametric prediction of event times in clinical trials.

BACKGROUND: In event-based clinical trials, it is common to conduct interim analyses at planned landmark event counts. Accurate prediction of the timing of these events can support logistical planning and the efficient allocation of resources. As the trial progresses, one may wish to use the accumulating data to refine predictions. PURPOSE: Available methods to predict event times include parametric cure and non-cure models and a nonparametric approach involving Bayesian bootstrap simulation. The parametric methods work well when their underlying assumptions are met, and the nonparametric method gives calibrated but inefficient predictions across a range of true models. In the early stages of a trial, when predictions have high marginal value, it is difficult to infer the form of the underlying model. We seek to develop a method that will adaptively identify the best-fitting model and use it to create robust predictions. METHODS: At each prediction time, we repeat the following steps: (1) resample the data; (2) identify, from among a set of candidate models, the one with the highest posterior probability; and (3) sample from the predictive posterior of the data under the selected model. RESULTS: A Monte Carlo study demonstrates that the adaptive method produces prediction intervals whose coverage is robust within the family of selected models. The intervals are generally wider than those produced assuming the correct model, but narrower than nonparametric prediction intervals. We demonstrate our method with applications to two completed trials: The International Chronic Granulomatous Disease study and Radiation Therapy Oncology Group trial 0129. LIMITATIONS: Intervals produced under any method can be badly calibrated when the sample size is small and unhelpfully wide when predicting the remote future. Early predictions can be inaccurate if there are changes in enrollment practices or trends in survival. CONCLUSIONS: An adaptive event-time prediction method that selects the model given the available data can give improved robustness compared to methods based on less flexible parametric models.

CDK 4/6 inhibitor palbociclib (PD0332991) in Rb+ advanced breast cancer: phase II activity, safety, and predictive biomarker assessment.

PURPOSE: The G1-S checkpoint of the cell cycle is frequently dysregulated in breast cancer. Palbociclib (PD0332991) is an oral inhibitor of CDK4/6. Based upon preclinical/phase I activity, we performed a phase II, single-arm trial of palbociclib in advanced breast cancer. EXPERIMENTAL DESIGN: Eligible patients had histologically confirmed, metastatic breast cancer positive for retinoblastoma (Rb) protein and measureable disease. Palbociclib was given at 125 mg orally on days 1 to 21 of a 28-day cycle. Primary objectives were tumor response and tolerability. Secondary objectives included progression-free survival (PFS) and assessment of Rb expression/localization, KI-67, p16 loss, and CCND1 amplification. RESULTS: Thirty-seven patients were enrolled; 84% hormone-receptor (HR)(+)/Her2(-), 5% HR(+)/Her2(+), and 11% HR(-)/Her2(-), with a median of 2 prior cytotoxic regimens. Two patients had partial response (PR) and 5 had stable disease ≥ 6 months for a clinical benefit rate (CBR = PR + 6moSD) of 19% overall, 21% in HR(+), and 29% in HR(+)/Her2(-) who had progressed through ≥2 prior lines of hormonal therapy. Median PFS overall was 3.7 months [95% confidence interval (CI), 1.9-5.1], but significantly longer for those with HR(+) versus HR(-) disease (P = 0.03) and those who had previously progressed through endocrine therapy for advanced disease (P = 0.02). Grade 3/4 toxicities included neutropenia (51%), anemia (5%), and thrombocytopenia (22%). Twenty-four percent had treatment interruption and 51% had dose reduction, all for cytopenias. No biomarker identified a sensitive tumor population. CONCLUSIONS: Single-agent palbociclib is well tolerated and active in patients with endocrine-resistant, HR(+), Rb-positive breast cancer. Cytopenias were uncomplicated and easily managed with dose reduction.

EVALUATING COSTS WITH UNMEASURED CONFOUNDING: A SENSITIVITY ANALYSIS FOR THE TREATMENT EFFECT.

Estimates of the effects of treatment on cost from observational studies are subject to bias if there are unmeasured confounders. It is therefore advisable in practice to assess the potential magnitude of such biases. We derive a general adjustment formula for loglinear models of mean cost and explore special cases under plausible assumptions about the distribution of the unmeasured confounder. We assess the performance of the adjustment by simulation, in particular, examining robustness to a key assumption of conditional independence between the unmeasured and measured covariates given the treatment indicator. We apply our method to SEER-Medicare cost data for a stage II/III muscle-invasive bladder cancer cohort. We evaluate the costs for radical cystectomy vs. combined radiation/chemotherapy, and find that the significance of the treatment effect is sensitive to plausible unmeasured Bernoulli, Poisson and Gamma confounders.

MRI assessment of early tumor response in metastatic renal cell carcinoma patients treated with sorafenib.

OBJECTIVE: The purpose of this study was to examine early MRI changes in renal cell carcinoma (RCC) treated with the antiangiogenic agent sorafenib and to identify MRI biomarkers of RCC response to sorafenib. MATERIALS AND METHODS: Sixteen patients with RCC were evaluated by MRI before and 3-12 weeks after commencing treatment with sorafenib. Two experienced MR radiologists, blinded to treatment status, independently graded tumor appearance on T1-weighted, T2-weighted, and gadolinium-enhanced images. The proportional odds mixed model was used to compare qualitative appearance of tumors before and after therapy. Time-to-progression was correlated with Response Evaluation Criteria in Solid Tumors (RECIST) 1.0 and MR-modified Choi criteria, incorporating changes in both tumor enhancement and size. RESULTS: After sorafenib therapy, there was a significant increase in T1 signal intensity of tumors (p < 0.0001) and a significant decrease in degree of tumor enhancement (p < 0.0001). The sum of unidimensional tumor diameters decreased significantly after therapy (p = 0.005). However, the average decrease in size at early follow-up was 13%, and all patients except one had stable disease by RECIST 1.0. Early responders defined by MR-modified Choi criteria had increased time-to-progression compared with nonresponders, whereas early RECIST evaluation did not predict clinical outcome. CONCLUSION: Decreased enhancement and T1 shortening of tumors on MRI may be useful biomarkers of RCC response to angiogenesis inhibitors. Response criteria combining early changes in size and enhancement lead to better correlation with clinical outcome compared with size decrease alone.

Comparative effectiveness of screening and prevention strategies among BRCA1/2-affected mutation carriers.

UNLABELLED: Comparative effectiveness research has become an integral part of health care planning in most developed countries. In a simulated cohort of women, aged 30-65, who tested positive for BRCA1 or BRCA2 mutations, we compared outcomes of mammography with and without MRI, prophylactic oophorectomy with and without mastectomy, mastectomy alone, and chemoprevention. METHODS: Using Treeage 9.02 software, we developed Markov models with 25,000 Monte Carlo simulations and conducted probabilistic sensitivity analysis. We based mutation penetrance rates, breast and ovarian cancer incidence, and mortality rates, and costs in terms of 2009 dollars, on published studies and data from the Surveillance, Epidemiology, and End RESULTS (SEER) Program and the Centers for Medicare and Medicaid Services. We used preference ratings obtained from mutation carriers and controls to adjust survival for quality of life (QALYs). RESULTS: For BRCA1 mutation carriers, prophylactic oophorectomy at $1,741 per QALY, was more cost effective than both surgeries and dominated all other interventions. For BRCA2 carriers, prophylactic oophorectomy, at $4,587 per QALY, was more cost effective than both surgeries. Without quality adjustment, both mastectomy and BSO surgeries dominated all other interventions. In all simulations, preventive surgeries or chemoprevention dominated or were more cost effective than screening because screening modalities were costly. CONCLUSION: Our analysis suggested that among BRCA1/2 mutation carriers, prophylactic surgery would dominate or be cost effective compared to chemoprevention and screening. Annual screening with MRI and mammography was the most effective strategy because it was associated with the longest quality-adjusted survival, but it was also very expensive.

A method comparison study of timeline followback and ecological momentary assessment of daily cigarette consumption.

INTRODUCTION: Uncertainty exists about how best to measure daily cigarette consumption. Two common measures are timeline followback (TLFB), which involves structured, prompted recall, and ecological momentary assessment (EMA), which involves recording consumption, as it occurs, on a handheld electronic device. METHODS: We evaluated the agreement between TLFB and EMA measures collected for 14 days prior to the target quit date from 236 smokers in a smoking cessation program. We performed a Bland-Altman analysis to assess agreement of TLFB and EMA using a regression-based model that allows for a nonuniform difference between methods and limits of agreement that can vary with the number of cigarettes smoked. RESULTS: For pairs of measurements taken on the same smoker, TLFB counts were on average 3.2 cigarettes higher than EMA counts; this difference increased for larger numbers of cigarettes. Using a model that allows for variable limits of agreement, the width of the 95% interval ranged from 8.7 to 61.8 cigarettes, with an average of 26.4 cigarettes. Variation between the methods increased substantially for larger cigarette counts, leading to wider limits and poorer agreement for heavy smokers. DISCUSSION: Throughout the measurement range, the estimated limits of agreement were far wider than the limits of clinical significance, defined a priori to be 20% of the number of cigarettes smoked. We conclude that TLFB and EMA cannot be considered equivalent for the assessment of daily cigarette consumption, especially for heavy smokers.

Weibull prediction of event times in clinical trials.

In clinical trials with interim analyses planned at pre-specified event counts, one may wish to predict the times of these landmark events as a tool for logistical planning. Currently available methods use either a parametric approach based on an exponential model for survival (Bagiella and Heitjan, Statistics in Medicine 2001; 20:2055) or a non-parametric approach based on the Kaplan-Meier estimate (Ying et al., Clinical Trials 2004; 1:352). Ying et al. (2004) demonstrated the trade-off between bias and variance in these models; the exponential method is highly efficient when its assumptions hold but potentially biased when they do not, whereas the non-parametric method has minimal bias and is well calibrated under a range of survival models but typically gives wider prediction intervals and may fail to produce useful predictions early in the trial. As a potential compromise, we propose here to make predictions under a Weibull survival model. Computations are somewhat more difficult than with the simpler exponential model, but Monte Carlo studies show that predictions are robust under a broader range of assumptions. We demonstrate the method using data from a trial of immunotherapy for chronic granulomatous disease.

Cost-effectiveness of preventive strategies for women with a BRCA1 or a BRCA2 mutation.

BACKGROUND: For BRCA1 or BRCA2 mutation carriers, decision analysis indicates that prophylactic surgery or chemoprevention leads to better survival than surveillance alone. OBJECTIVE: To evaluate the cost-effectiveness of the preventive strategies that are available to unaffected women carrying a single BRCA1 or BRCA2 mutation with high cancer penetrance. DESIGN: Markov modeling with Monte Carlo simulations and probabilistic sensitivity analyses. DATA SOURCES: Breast and ovarian cancer incidence and mortality rates, preference ratings, and costs derived from the literature; the Surveillance, Epidemiology, and End Results (SEER) Program; and the Health Care Financing Administration (now the Centers for Medicare & Medicaid Services). TARGET POPULATION: Unaffected carriers of a single BRCA1 or BRCA2 mutation 35 to 50 years of age. TIME HORIZON: Lifetime. PERSPECTIVE: Health policy, societal. INTERVENTIONS: Tamoxifen, oral contraceptives, bilateral salpingo-oophorectomy, mastectomy, both surgeries, or surveillance. OUTCOME MEASURES: Cost-effectiveness. RESULTS OF BASE-CASE ANALYSIS: For mutation carriers 35 years of age, both surgeries (prophylactic bilateral mastectomy and oophorectomy) had an incremental cost-effectiveness ratio over oophorectomy alone of 2352 dollars per life-year for BRCA1 and 100 dollars per life-year for BRCA2. With quality adjustment, oophorectomy dominated all other strategies for BRCA1 and had an incremental cost-effectiveness ratio of 2281 dollars per life-year for BRCA2. RESULTS OF SENSITIVITY ANALYSIS: Older age at intervention increased the cost-effectiveness of prophylactic mastectomy for BRCA1 mutation carriers to 73,755 dollars per life-year. Varying the penetrance, mortality rates, costs, discount rates, and preferences had minimal effects on outcomes. LIMITATIONS: Results are dependent on the accuracy of model assumptions. CONCLUSION: On the basis of this model, the most cost-effective strategies for BRCA mutation carriers, with and without quality adjustment, were oophorectomy alone and oophorectomy and mastectomy, respectively.

An index of local sensitivity to nonignorable drop-out in longitudinal modelling.

In longitudinal studies with potentially nonignorable drop-out, one can assess the likely effect of the nonignorability in a sensitivity analysis. Troxel et al. proposed a general index of sensitivity to nonignorability, or ISNI, to measure sensitivity of key inferences in a neighbourhood of the ignorable, missing at random (MAR) model. They derived detailed formulas for ISNI in the special case of the generalized linear model with a potentially missing univariate outcome. In this paper, we extend the method to longitudinal modelling. We use a multivariate normal model for the outcomes and a regression model for the drop-out process, allowing missingness probabilities to depend on an unobserved response. The computation is straightforward, and merely involves estimating a mixed-effects model and a selection model for the drop-out, together with some simple arithmetic calculations. We illustrate the method with three examples.

Bayesian estimation of cost-effectiveness from censored data.

We describe a Bayesian methodology for estimating the cost-effectiveness of a new treatment compared to a standard in a clinical trial, when censoring of survival, the effectiveness variable, induces censoring of total cost. The statistical model assumes that survival follows a Weibull distribution and that total health care cost follows a gamma distribution whose mean has a linear regression on survival time. We summarize the posterior distributions of key parameters by importance sampling. We illustrate the method with an analysis of data from a randomized clinical trial of a treatment for cardiovascular disease.

Bayesian estimation of cost-effectiveness: an importance-sampling approach.

We describe a method for estimating the cost-effectiveness of a new treatment compared to a standard, using data from a comparative clinical trial. We quantify the clinical effectiveness as a binary variable indicating success or failure. The underlying statistical model assumes that costs are uncensored and follow separate gamma distributions in each of the groups defined by the four possible combinations of treatment arm and effectiveness outcome. The method is subjectivist, in that it represents prior uncertainty about model parameters with a probability distribution, which we update via Bayes's theorem to produce a posterior distribution. We approximate the posterior by importance sampling, a straightforward simulation method. We illustrate the method with an analysis of cost (derived from resource usage data) and effectiveness (measured by one-year survival) in a clinical trial in heart disease. The example demonstrates that the method is practical and provides for a flexible data analysis.

Effect of prevention strategies on survival and quality-adjusted survival of women with BRCA1/2 mutations: an updated decision analysis.

PURPOSE: This study updates findings regarding the effects of prophylactic surgery, chemoprevention, and surveillance on the survival and quality-adjusted survival of women who test positive for BRCA1/2 mutations. MATERIALS AND METHODS: Markov modeling of outcomes was performed in a simulated cohort of 30-year-old women who tested positive for BRCA1/2 mutations. The model incorporated breast and ovarian cancer incidence rates from the literature and mortality rates from the Surveillance, Epidemiology, and End Results Program. Quality adjustment of survival estimates were obtained from a survey of women aged 33 to 50 years. Sensitivity analyses were performed of varied assumptions regarding timing and effects of preventive measures on cancer incidence and adverse effects. RESULTS: A 30-year-old woman could prolong her survival beyond that associated with surveillance alone by use of preventive measures: 1.8 years with tamoxifen, 2.6 years with prophylactic oophorectomy, 4.6 years with both tamoxifen and prophylactic oophorectomy, 3.5 years with prophylactic mastectomy, and 4.9 years with both surgeries. She could prolong her quality-adjusted survival by 2.8 years with tamoxifen, 4.4 years with prophylactic oophorectomy, 6.3 years with tamoxifen and oophorectomy, and 2.6 years with mastectomy, or with both surgeries. The benefits of all of these strategies would decrease if they were initiated at later ages. CONCLUSION: Women who test positive for BRCA1/2 mutations may derive greater survival and quality adjusted survival benefits than previously reported from chemoprevention, prophylactic surgery, or a combination. Observational studies and clinical trials are needed to verify the results of this analysis of the long-term benefits of preventive strategies among BRCA1/2-positive women.

Survival associated with 5-fluorouracil-based adjuvant chemotherapy among elderly patients with node-positive colon cancer.

BACKGROUND: Randomized clinical trials have demonstrated the efficacy of adjuvant 5-fluorouracil (5-FU)-based chemotherapy after surgical resection of node-positive colon cancer. Although this treatment became the standard in 1990 following a National Institutes of Health Consensus Conference, among those at least 65 years of age it is less likely to be offered to older or nonwhite patients. OBJECTIVE: To determine the association between 5-fu-based chemotherapy and survival in older patients. DESIGN: Retrospective cohort study. SETTING: Combined database of the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) program and Medicare. PATIENTS: 4768 patients 65 years of age or older who received a diagnosis of node-positive colon cancer from 1992 to 1996, were covered by Medicare Parts A and B, and resided in the population covered by the SEER program. MEASUREMENTS: Propensity scores to control for known predictors of receiving treatment, Cox proportional hazards models to assess the association of 5-FU therapy with survival, and sensitivity analyses to estimate the possible effects of unknown confounders. RESULTS: Fifty-two percent of patients received 5-FU therapy. For this sample, the hazard ratio for death associated with 5-FU therapy was 0.66 (95% CI, 0.60 to 0.73). Confounding could have accounted for this association only if an unmeasured confounder were extremely unequally distributed between the treated and untreated groups or increased mortality by at least 50%. CONCLUSIONS: 5-Fluorouracil adjuvant therapy is significantly associated with reduced mortality in older patients, similar to the association found in randomized, controlled trials among younger patients. More frequent use of 5-FU therapy in older patients would probably reduce death from colon cancer.

Outcomes of tamoxifen chemoprevention for breast cancer in very high-risk women: a cost-effectiveness analysis.

PURPOSE: To estimate the effects on survival, quality-adjusted survival, and health care costs of using tamoxifen for primary prevention in subgroups of women at very high risk for breast cancer. PATIENTS AND METHODS: A decision analysis was performed using a hypothetical cohort of women that included subgroups with atypical hyperplasia, Gail risk greater than 5, lobular carcinoma-in-situ, or two or more first-degree relatives with breast cancer. Data sources were the Breast Cancer Prevention Trial, the Surveillance, Epidemiology, and End-Results program, time trade-off preference ratings, the Group Health Cooperative of Puget Sound, and the United States Health Care Financing Administration. RESULTS: Our model predicted that tamoxifen would prolong the average survival of cohort members initiating use at ages 35, 50, and 60 years by 70, 42, and 27 days, respectively. It would prolong survival even more for those in the higher-risk groups, especially those with atypical hyperplasia (202, 89, and 45 days). Tamoxifen use was also projected to extend quality-adjusted survival by 158, 80, and 50 days in the atypical hyperplasia group. For younger women in the highest risk groups, chemoprevention with tamoxifen was estimated to have cost savings or be cost-effective, both with and without quality adjustments. CONCLUSION: Chemoprevention with tamoxifen may be particularly beneficial to women with atypical hyperplasia, 5-year Gail model risk greater than 5%, lobular carcinoma-in-situ, or two or more first-degree relatives with breast cancer. The benefits may be greater if tamoxifen is initiated before age 50 years rather than after and if the breast cancer risk reduction conferred by tamoxifen lasts longer than 5 years. For women with a very high risk of invasive breast cancer, chemoprevention with tamoxifen seems to be cost-effective.