Search

Parameters of subpopulations can be more relevant than those of superpopulations. For example, a healthcare provider may be interested in the effect of a treatment plan for a specific subset of their patients; policymakers may be concerned with the impact of a policy in a particular state within a given population. In these cases, the focus is on a specific finite population, as opposed to an infinite superpopulation. Such a population can be characterized by fixing some attributes that are intrinsic to them, leaving unexplained variations like measurement error as random. Inference for a population with fixed attributes can then be modelled as inferring parameters of a conditional distribution. Accordingly, it is desirable that confidence intervals are conditionally valid for the realized population, instead of marginalizing over many possible draws of populations. We provide a statistical inference framework for parameters of finite populations with known attributes. Leveraging the attribute information, our estimators and confidence intervals closely target a specific finite population. When the data are from the population of interest, our confidence intervals attain asymptotic conditional validity, given the attributes, and are shorter than those for superpopulation inference. In addition, we develop procedures to infer parameters of new populations with differing covariate distributions; the confidence intervals are also conditionally valid for the new populations under mild conditions. Our methods extend to situations where the fixed information has a weaker structure or is only partially observed. We demonstrate the validity and applicability of our methods using simulated data and a real-word dataset for predicting car prices. [ABSTRACT FROM AUTHOR]

Background the relationship between low-density lipoprotein cholesterol (LDL-C) and adverse outcomes among the older people remains controversial. Objective to further clarify the association between admission LDL-C levels and cardiovascular mortality (CVM) among oldest old individuals (≥80 years) with acute myocardial infarction (AMI). Design a prospective cohort study. Setting two-centre. Subjects a consecutive sample of 1,224 oldest old individuals with AMI admitted to Beijing FuWai and Shenzhen FuWai hospitals. Methods all individuals were subdivided according to baseline LDL-C levels (<1.8, 1.8–2.6 and ≥ 2.6 mmol/l) and further stratified by high-sensitivity C-reactive protein (hsCRP) concentrations (<10 and ≥10 mg/l). The primary outcome was CVM. The time from admission to the occurrence of CVM or the last follow-up was analysed in Kaplan–Meier and Cox analyses. Results the median age of the overall population was 82 years. During an average of 24.5 months' follow-up, 299 cardiovascular deaths occurred. Kaplan–Meier analysis showed that LDL-C < 1.8 mmol/l group had the highest CVM among oldest old individuals with AMI. Multivariate Cox regression analysis further revealed that compared with those with LDL-C levels <1.8 mmol/l, subjects with LDL-C levels ≥2.6 mmol/l (hazard ratio: 0.67, 95% confidence interval: 0.46–0.98) had significantly lower risk of CVM, especially in those with high hsCRP levels. Moreover, when categorising according to LDL-C and hsCRP together, data showed that individuals with low LDL-C and high hsCRP levels had the highest CVM. Conclusions LDL-C < 1.8 mmol/l was associated with a high CVM after AMI in oldest old individuals, especially when combined with high hsCRP levels, which may need to be confirmed by randomised controlled trials. [ABSTRACT FROM AUTHOR]