Improved accuracy in predicting coronary heart disease (CHD) risk in persons with diabetes and kidney disease is needed. Addition of albuminuria to established methods of CHD risk calculation was reported in the Strong Heart Study (SHS) cohort. In this article, the addition of estimated glomerular filtration rate (eGFR) is evaluated with data from 4,549 American Indian SHS participants, ages 45–74 years. After adjustment for Framingham CHD risk factors, hazard ratios for eGFR as a predictor of CHD were 1.69 (95% confidence interval: 1.34, 2.13) in women and 1.41 (95% confidence interval: 0.94, 2.13) in men. Models including albuminuria, eGFR, or both scored higher in discriminatory power than models using conventional risk factors alone in women; in men the improvement was seen only for albuminuria and the combination of albuminuria and eGFR. Hosmer–Lemeshow assessments showed good calibration for the models using eGFR alone in both genders, followed by models including albuminuria alone in both genders. Adding eGFR improved the NRI in women (0.085, p=.0004) but not in men (0.010, p=0.1967). NRI and IDI were improved in both genders using both albuminuria and eGFR (NRI: 0.135, p<.0001; IDI: 0.027, p<.0001 in women; NRI: 0.035, p<0.0196; IDI: 0.008 p<0.0156 in men). Therefore, a risk calculator including albuminuria enhances CHD prediction compared with one using only standard risk factors in men and women. Including eGFR alone improves risk prediction in women, but for men it is preferable to have both eGFR and albuminuria. In conclusion, this enhanced calculator should be useful in estimating CHD risk in populations with high prevalence of diabetes and renal disease.

Delays from diagnosis to treatment may account in part for poor head and neck cancer outcomes among patients of low socioeconomic status, minority patients, and the uninsured. The purpose of this study is to evaluate factors associated with treatment delay in an urban county hospital system. METHOD: Retrospective review of consecutive head and neck cancer patients treated at a county hospital serving an indigent urban population and uninsured county residents. Multivariate analysis was performed using the Cox regression model. RESULTS: Ninety consecutive patients were evaluated. The average time from diagnosis to initiation of treatment was 49 days. On univariate analysis, age Ͻ55 years, oropharyngeal primary tumor site, and primary treatment with chemotherapy and radiation were associated with a statistically significant longer delay to treatment. On multivariate analysis, male gender (HR 2.5, 95% CI 1.3-4.9, pϭ0.008), nonsurgical primary treatment modality (HR 0.8, 95% CI 0.3-0.6, pϭ0.003), and positive nodal status (HRϭ0.4, 95% CI 0.2-0.9, pϭ0.018) were associated with a significantly longer delay in treatment, while an association with oropharyngeal primary tumor site approached statistical significance (HRϭ2.2, 95% CI 1.0-4.8, pϭ0.052). CONCLUSION: Alarming delays in head and neck cancer care exist among indigent patients and the uninsured, with young adult males, patients with positive nodal disease, and those receiving primary chemoradiation at highest risk. Possible reasons for the delay include difficulty navigating multidisciplinary cancer care, poor patient compliance, and a lack of resources. Improved patient outreach and more efficient allocation of treatment resources are needed.

Kidney and cardiovascular disease are widespread among populations with high prevalence of diabetes, such as American Indians participating in the Strong Heart Study (SHS). Studying these conditions simultaneously in longitudinal studies is challenging, because the morbidity and mortality associated with these diseases result in missing data, and these data are likely not missing at random. When such data are merely excluded, study findings may be compromised. In this article, a subset of 2264 participants with complete renal function data from Strong Heart Exams 1 (1989–1991), 2 (1993–1995), and 3 (1998–1999) was used to examine the performance of five methods used to impute missing data: listwise deletion, mean of serial measures, adjacent value, multiple imputation, and pattern-mixture. Three missing at random models and one non-missing at random model were used to compare the performance of the imputation techniques on randomly and non-randomly missing data. The pattern-mixture method was found to perform best for imputing renal function data that were not missing at random. Determining whether data are missing at random or not can help in choosing the imputation method that will provide the most accurate results.

691 Background: In the absence of mature overall survival (OS) endpoints, interim clinical trial data can be used to predict long-term survival in patients (pts) with advanced malignancies and inform trial continuation, treatment preference, and reimbursement decisions. Research shows that HRQoL assessments can be associated with OS, offering potential utility for validated HRQoL scales. We describe a predictive model used to determine the extent to which HRQoL data could predict CS (survival conditional on progression at either 6 or 12 months). Methods: Using pt-level data from a large phase III trial of nivolumab vs everolimus, a simulation approach with a survival random forest algorithm identified factors statistically important in predicting CS from a large number of covariates measured at baseline. Stepwise Cox proportional hazard survival models were fitted using covariates identified as important. Baseline scores and change over time were tested to determine the influence on the predictive power of the HRQoL data. Results: For both nivolumab and everolimus, baseline FKSI values were significant predictors of CS; median survival times roughly doubled for pts with baseline FKSI scores ≥30 vs pts with scores < 30 (nivolumab, 31.3-16.6; everolimus, 26.6-11). Baseline FKSI scores were the most important predictor vs the other baseline covariates from the survival random forest simulation, and a statistically significant covariate when fitting a stepwise Cox proportional hazard survival model. Change in scores over time influenced CS for pts with high baseline scores, and pts who demonstrated improvement in scores vs baseline had significantly higher CS vs pts with low baseline and no improvement in scores vs baseline. However, when examining change in HRQoL score over time, the statistical importance of the covariate begins to diminish due to high correlations with factors such as adverse events and weight change. Conclusions: We found that HRQoL data, specifically the FKSI, could be useful in predicting CS, especially at the onset of a trial. The importance of the FKSI score in predicting CS could be a powerful complement to existing clinical prognostic factors.

No abstract