Test result-based sampling: an efficient design for estimating the accuracy of patient safety indicators.

OBJECTIVE: Accuracy studies of Patient Safety Indicators (PSIs) are critical but limited by the large samples required due to low occurrence of most events. We tested a sampling design based on test results (verification-biased sampling [VBS]) that minimizes the number of subjects to be verified. METHODS: We considered 3 real PSIs, whose rates were calculated using 3 years of discharge data from a university hospital and a hypothetical screen of very rare events. Sample size estimates, based on the expected sensitivity and precision, were compared across 4 study designs: random and VBS, with and without constraints on the size of the population to be screened. RESULTS: Over sensitivities ranging from 0.3 to 0.7 and PSI prevalence levels ranging from 0.02 to 0.2, the optimal VBS strategy makes it possible to reduce sample size by up to 60% in comparison with simple random sampling. For PSI prevalence levels below 1%, the minimal sample size required was still over 5000. CONCLUSIONS: Verification-biased sampling permits substantial savings in the required sample size for PSI validation studies. However, sample sizes still need to be very large for many of the rarer PSIs.

Iterated local search

Iterated Local Search has many of the desirable features of a metaheuristic: it is simple, easy to implement, robust, and highly effective. The essential idea of Iterated Local Search lies in focusing the search not on the full space of solutions but on a smaller subspace defined by the solutions that are locally optimal for a given optimization engine. The success of Iterated Local Search lies in the biased sampling of this set of local optima. How effective this approach turns out to be depends mainly on the choice of the local search, the perturbations, and the acceptance criterion. So far, in spite of its conceptual simplicity, it has lead to a number of state-of-the-art results without the use of too much problem-specific knowledge. But with further work so that the different modules are well adapted to the problem at hand, Iterated Local Search can often become a competitive or even state of the artalgorithm. The purpose of this review is both to give a detailed description of this metaheuristic and to show where it stands in terms of performance.

Proving Efficiency of NPMLE and Identities

A method is given for proving efficiency of NPMLE directly linked to empirical process theory. The conditions in general are appropriate consistency of the NPMLE, differentiability of the model, differentiability of the parameter of interest, local convexity of the parameter space, and a Donsker class condition for the class of efficient influence functions obtained by varying the parameters. For the case that the model is linear in the parameter and the parameter space is convex, as with most nonparametric missing data models, we show that the method leads to an identity for the NPMLE which almost says that the NPMLE is efficient and provides us straightforwardly with a consistency and efficiency proof. This identify is extended to an almost linear class of models which contain biased sampling models. To illustrate, the method is applied to the univariate censoring model, random truncation models, interval censoring case I model, the class of parametric models and to a class of semiparametric models.

Analyzing left-truncated right-censored data with uncertain onset time with parametric models

Prevalent sampling is an efficient and focused approach to the study of the natural history of disease. Right-censored time-to-event data observed from prospective prevalent cohort studies are often subject to left-truncated sampling. Left-truncated samples are not randomly selected from the population of interest and have a selection bias. Extensive studies have focused on estimating the unbiased distribution given left-truncated samples. However, in many applications, the exact date of disease onset was not observed. For example, in an HIV infection study, the exact HIV infection time is not observable. However, it is known that the HIV infection date occurred between two observable dates. Meeting these challenges motivated our study. We propose parametric models to estimate the unbiased distribution of left-truncated, right-censored time-to-event data with uncertain onset times. We first consider data from a length-biased sampling, a specific case in left-truncated samplings. Then we extend the proposed method to general left-truncated sampling. With a parametric model, we construct the full likelihood, given a biased sample with unobservable onset of disease. The parameters are estimated through the maximization of the constructed likelihood by adjusting the selection bias and unobservable exact onset. Simulations are conducted to evaluate the finite sample performance of the proposed methods. We apply the proposed method to an HIV infection study, estimating the unbiased survival function and covariance coefficients. ^

A distribution of tumor size at detection and its limiting form.

A distribution of tumor size at detection is derived within the framework of a mechanistic model of carcinogenesis with the object of estimating biologically meaningful parameters of tumor latency. Its limiting form appears to be a generalization of the distribution that arises in the length-biased sampling from stationary point processes. The model renders the associated estimation problems tractable. The usefulness of the proposed approach is illustrated with an application to clinical data on premenopausal breast cancer.

Biased representation of disturbance rates in the roadside sampling frame in boreal forests: implications for monitoring design

The North American Breeding Bird Survey (BBS) is the principal source of data to inform researchers about the status of and trend for boreal forest birds. Unfortunately, little BBS coverage is available in the boreal forest, where increasing concern over the status of species breeding there has increased interest in northward expansion of the BBS. However, high disturbance rates in the boreal forest may complicate roadside monitoring. If the roadside sampling frame does not capture variation in disturbance rates because of either road placement or the use of roads for resource extraction, biased trend estimates might result. In this study, we examined roadside bias in the proportional representation of habitat disturbance via spatial data on forest “loss,” forest fires, and anthropogenic disturbance. In each of 455 BBS routes, the area disturbed within multiple buffers away from the road was calculated and compared against the area disturbed in degree blocks and BBS strata. We found a nonlinear relationship between bias and distance from the road, suggesting forest loss and forest fires were underrepresented below 75 and 100 m, respectively. In contrast, anthropogenic disturbance was overrepresented at distances below 500 m and underrepresented thereafter. After accounting for distance from road, BBS routes were reasonably representative of the degree blocks they were within, with only a few strata showing biased representation. In general, anthropogenic disturbance is overrepresented in southern strata, and forest fires are underrepresented in almost all strata. Similar biases exist when comparing the entire road network and the subset sampled by BBS routes against the amount of disturbance within BBS strata; however, the magnitude of biases differed. Based on our results, we recommend that spatial stratification and rotating panel designs be used to spread limited BBS and off-road sampling effort in an unbiased fashion and that new BBS routes be established where sufficient road coverage exists.

Tests for sex-biased dispersal using bi-parentally inherited genetic markers.

Understanding why dispersal is sex-biased in many taxa is still a major concern in evolutionary ecology. Dispersal tends to be male-biased in mammals and female-biased in birds, but counter-examples exist and little is known about sex bias in other taxa. Obtaining accurate measures of dispersal in the field remains a problem. Here we describe and compare several methods for detecting sex-biased dispersal using bi-parentally inherited, codominant genetic markers. If gene flow is restricted among populations, then the genotype of an individual tells something about its origin. Provided that dispersal occurs at the juvenile stage and that sampling is carried out on adults, genotypes sampled from the dispersing sex should on average be less likely (compared to genotypes from the philopatric sex) in the population in which they were sampled. The dispersing sex should be less genetically structured and should present a larger heterozygote deficit. In this study we use computer simulations and a permutation test on four statistics to investigate the conditions under which sex-biased dispersal can be detected. Two tests emerge as fairly powerful. We present results concerning the optimal sampling strategy (varying number of samples, individuals, loci per individual and level of polymorphism) under different amounts of dispersal for each sex. These tests for biases in dispersal are also appropriate for any attribute (e.g. size, colour, status) suspected to influence the probability of dispersal. A windows program carrying out these tests can be freely downloaded from http://www.unil.ch/izea/softwares/fstat.html

Policy Improvement for POMDPs Using Normalized Importance Sampling

We present a new method for estimating the expected return of a POMDP from experience. The estimator does not assume any knowle ge of the POMDP and allows the experience to be gathered with an arbitrary set of policies. The return is estimated for any new policy of the POMDP. We motivate the estimator from function-approximation and importance sampling points-of-view and derive its theoretical properties. Although the estimator is biased, it has low variance and the bias is often irrelevant when the estimator is used for pair-wise comparisons.We conclude by extending the estimator to policies with memory and compare its performance in a greedy search algorithm to the REINFORCE algorithm showing an order of magnitude reduction in the number of trials required.

Land Cover Sampling Biases Associated with Roadside Bird Surveys

Roadside surveys such as the Breeding Bird Survey (BBS) are widely used to assess the relative abundance of bird populations. The accuracy of roadside surveys depends on the extent to which surveys from roads represent the entire region under study. We quantified roadside land cover sampling bias in Tennessee, USA, by comparing land cover proportions near roads to proportions of the surrounding region. Roadside surveys gave a biased estimate of patterns across the region because some land cover types were over- or underrepresented near roads. These biases changed over time, introducing varying levels of distortion into the data. We constructed simulated population trends for five bird species of management interest based on these measured roadside sampling biases and on field data on bird abundance. These simulations indicated that roadside surveys may give overly negative assessments of the population trends of early successional birds and of synanthropic birds, but not of late-successional birds. Because roadside surveys are the primary source of avian population trend information in North America, we conclude that these surveys should be corrected for roadside land cover sampling bias. In addition, current recommendations about the need to create more early successional habitat for birds may need reassessment in the light of the undersampling of this habitat by roads.

Variance estimation for systematic sampling from deliberately ordered populations

The systematic sampling (SYS) design (Madow and Madow, 1944) is widely used by statistical offices due to its simplicity and efficiency (e.g., Iachan, 1982). But it suffers from a serious defect, namely, that it is impossible to unbiasedly estimate the sampling variance (Iachan, 1982) and usual variance estimators (Yates and Grundy, 1953) are inadequate and can overestimate the variance significantly (Särndal et al., 1992). We propose a novel variance estimator which is less biased and that can be implemented with any given population order. We will justify this estimator theoretically and with a Monte Carlo simulation study.

Bats from Fazenda Intervales, Southeastern Brazil: species account and comparison between different sampling methods

Assessing the composition of an area's bat fauna is typically accomplished by using captures or by monitoring echolocation calls with bat detectors. The two methods may not provide the same data regarding species composition. Mist nets and harp traps may be biased towards sampling low flying species, and bat detectors biased towards detecting high intensity echolocators. A comparison of the bat fauna of Fazenda Intervales, southeastern Brazil, as revealed by mist nets and harp trap captures, checking roosts and by monitoring echolocation calls of flying bats illustrates this point. A total of 17 species of bats was sampled. Fourteen bat species were captured and the echolocation calls of 12 species were recorded, three of them not revealed by mist nets or harp traps. The different sampling methods provided different pictures of the bat fauna. Phyllostomid bats dominated the catches in mist nets, but in the field their echolocation calls were never detected. No single sampling approach provided a complete assessment of the bat fauna in the study area. In general, bats producing low intensity echolocation calls, such as phyllostomids, are more easily assessed by netting, and bats producing high intensity echolocation calls are better surveyed by bat detectors. The results demonstrate that a combined and varied approach to sampling is required for a complete assessment of the bat fauna of an area.

Enhancing Sensitivity In Natural Product Nmr: Exploring Optimal Non-Uniform Sampling

Recent optimizations of NMR spectroscopy have focused their attention on innovations in new hardware, such as novel probes and higher field strengths. Only recently has the potential to enhance the sensitivity of NMR through data acquisition strategies been investigated. This thesis has focused on the practice of enhancing the signal-to-noise ratio (SNR) of NMR using non-uniform sampling (NUS). After first establishing the concept and exact theory of compounding sensitivity enhancements in multiple non-uniformly sampled indirect dimensions, a new result was derived that NUS enhances both SNR and resolution at any given signal evolution time. In contrast, uniform sampling alternately optimizes SNR (t < 1.26T2) or resolution (t~3T2), each at the expense of the other. Experiments were designed and conducted on a plant natural product to explore this behavior of NUS in which the SNR and resolution continue to improve as acquisition time increases. Possible absolute sensitivity improvements of 1.5 and 1.9 are possible in each indirect dimension for matched and 2x biased exponentially decaying sampling densities, respectively, at an acquisition time of ¿T2. Recommendations for breaking into the linear regime of maximum entropy (MaxEnt) are proposed. Furthermore, examination into a novel sinusoidal sampling density resulted in improved line shapes in MaxEnt reconstructions of NUS data and comparable enhancement to a matched exponential sampling density. The Absolute Sample Sensitivity derived and demonstrated here for NUS holds great promise in expanding the adoption of non-uniform sampling.

Effects of passage history and sampling bias on phylogenetic reconstruction of human influenza A evolution

In this paper we determine the extent to which host-mediated mutations and a known sampling bias affect evolutionary studies of human influenza A. Previous phylogenetic reconstruction of influenza A (H3N2) evolution using the hemagglutinin gene revealed an excess of nonsilent substitutions assigned to the terminal branches of the tree. We investigate two hypotheses to explain this observation. The first hypothesis is that the excess reflects mutations that were either not present or were at low frequency in the viral sample isolated from its human host, and that these mutations increased in frequency during passage of the virus in embryonated eggs. A set of 22 codons known to undergo such “host-mediated” mutations showed a significant excess of mutations assigned to branches attaching sequences from egg-cultured (as opposed to cell-cultured) isolates to the tree. Our second hypothesis is that the remaining excess results from sampling bias. Influenza surveillance is purposefully biased toward sequencing antigenically dissimilar strains in an effort to identify new variants that may signal the need to update the vaccine. This bias produces an excess of mutations assigned to terminal branches simply because an isolate with no close relatives is by definition attached to the tree by a relatively long branch. Simulations show that the magnitude of excess mutations we observed in the hemagglutinin tree is consistent with expectations based on our sampling protocol. Sampling bias does not affect inferences about evolution drawn from phylogenetic analyses. However, if possible, the excess caused by host-mediated mutations should be removed from studies of the evolution of influenza viruses as they replicate in their human hosts.

Effectiveness of three sampling methods to survey saproxylic beetle assemblages in Mediterranean woodland

The choice of sampling methods to survey saproxylic beetles is a key aspect to assessing conservation strategies for one of the most endangered assemblages in Europe. We evaluated the efficiency of three sampling methods: baited tube traps (TT), window traps in front of a hollow opening (WT), and emergence traps covering tree hollows (ET) to study richness and diversity of saproxylic beetle assemblages at species and family levels in Mediterranean woodlands. We also examined trap efficiency to report ecological diversity, and changes in the relative richness and abundance of species forming trophic guilds: xylophagous, saprophagous/saproxylophagous, xylomycetophagous, predators and commensals. WT and ET were similarly effective in reporting species richness and diversity at species and family levels, and provided an accurate profile of both the flying active and hollow-linked saproxylic beetle assemblages. WT and ET were the most complementary methods, together reporting more than 90 % of richness and diversity at both species and family levels. Diversity, richness and abundance of guilds were better characterized by ET, which indicates higher efficiency in outlining the ecological community of saproxylics that inhabit tree hollows. TT were the least effective method at both taxonomic levels, sampling a biased portion of the beetle assemblage attracted to trapping principles, however they could be used as a specific method for families such as Bostrichiidae, Biphyllidae, Melyridae, Mycetophagidae or Curculionidae Scolytinae species. Finally, ET and WT combination allows a better characterization of saproxylic assemblages in Mediterranean woodland, by recording species with different biology and linked to different microhabitat types.

Biased Random-key Genetic Algorithms For The Winner Determination Problem In Combinatorial Auctions.

Abstract In this paper, we address the problem of picking a subset of bids in a general combinatorial auction so as to maximize the overall profit using the first-price model. This winner determination problem assumes that a single bidding round is held to determine both the winners and prices to be paid. We introduce six variants of biased random-key genetic algorithms for this problem. Three of them use a novel initialization technique that makes use of solutions of intermediate linear programming relaxations of an exact mixed integer-linear programming model as initial chromosomes of the population. An experimental evaluation compares the effectiveness of the proposed algorithms with the standard mixed linear integer programming formulation, a specialized exact algorithm, and the best-performing heuristics proposed for this problem. The proposed algorithms are competitive and offer strong results, mainly for large-scale auctions.