True Population Values Research Articles

Overlap of genetic risk for Lewy body and Alzheimer’s disease pathology

AbstractBackgroundLewy body (LB) pathology is relatively understudied compared to Alzheimer’s disease (AD) pathology, and the genetic risk overlap between the two remains unclear. Chia et al. (2021) identified five genomic loci significantly associated with LB pathology in a genome‐wide association study (GWAS) of LB dementia cases and controls (N = 6,618), where 69% of cases and 15% of controls were pathologically confirmed. Here we gathered a largely distinct set of individuals (5.1% overlap) with genetic data and postmortem assessment for both AD and LB pathology (N = 6,002) to estimate the odds ratio (OR) for risk of sole AD pathology (AD+LB−), sole LB pathology (AD−LB+), and dual AD‐LB pathology (AD+LB+) for the five variants reported by Chia et al. We compared our results for APOE to ORs reported by Tsuang et al. (2013).MethodOur initial study population comprised 4,579 subjects from the National Alzheimer’s Coordinating Center and 1,423 from the Rush University Medical Center assessed for AD and LB pathology. We separately used criteria from Tsuang et al. and from Kaivola et al. (2021) to classify subjects as AD+LB+, AD+LB−, AD−LB+, or AD−LB−. We performed a GWAS for three phenotypic contrasts: each of AD+LB−, AD−LB+, and AD+LB+ against AD−LB−. We estimated ORs for the variants of interest.ResultWe found that rs769449 (APOE locus) was associated with AD+LB+ and AD+LB− at the genome‐wide significance level and nominally associated with AD−LB+, though with OR estimates lower than those reported by Tsuang et al. We obtained qualitatively similar results for rs6733839 (BIN1). The lead variants reported at the GBA, TMEM175, and SNCA‐AS1 loci (or proxies in high linkage disequilibrium) were not associated with a pathological phenotype, although an independent intronic variant of GBA was nominally associated with AD+LB+ and AD−LB+.ConclusionOur results suggest that APOE and BIN1 are associated with both AD and LB pathology, while GBA is associated with LB but not AD pathology. The ORs reported by Tsuang et al. for APOE‐associated risk for single and dual AD‐LB pathology may overestimate the true population values. Overall, AD and LB pathology share APOE and BIN1 as risk factors.

Bayesian analysis of diagnostic sensitivity and specificity for detecting infectious salmon anaemia virus (ISAV) using IFAT and real-time RT-PCR testing from laboratories in Atlantic Canada

Infectious Salmon Anaemia virus (ISAV) is the causative agent of Infectious Salmon Anaemia (ISA), an economically important disease in farmed Atlantic salmon (Salmo salar L.). Immunofluorescence antibody test (IFAT) and real-time reverse transcription polymerase chain reaction (qRT-PCR) are commonly used diagnostic tests for ISAV active surveillance in Atlantic Canada; however, their diagnostic accuracy is not well reported. In the absence of a reference test, we used Bayesian latent class analysis methods to estimate diagnostic sensitivity (DSe) and specificity (DSp) of IFAT and qRT-PCR. We compared diagnostic performance of qRT-PCR, performed at two laboratories, and IFAT (performed at another laboratory) using field samples obtained from two separate salmon populations (total n = 130) in province of Newfoundland and Labrador, Canada. Conditional independence models for two tests (IFAT and qRT-PCR) and two populations (low and high prevalence) were run separately on datasets from each laboratory. Informative priors were used for DSe and DSp of qRT-PCR and IFAT based on prior knowledge, while flat priors (beta 1,1) were used for prevalence. qRT-PCRs had similarly high DSe (>91%). IFAT was the least sensitive method (DSe range, 56.9%–60.4%). Both IFAT (DSp > 98.8%) and qRT-PCRs (DSp > 97.9%) showed high DSp. Real-time RT-PCR provided the greatest DSe for routine surveillance of Atlantic salmon farms while not sacrificing DSp. Sample size for the study was not likely adequate to estimate all parameters with very high confidence, which could have influenced the posterior probability interval calculated for DSe, resulting in less diagnostic confidence in the results (posterior probability for Sensitivities for the tests within 10% of the true population value instead of 5%). Although IFAT provided a high DSp, the qRT-PCR performed similarly in this regard. The findings indicate that the qRT-PCR is fit for ISAV surveillance programs.

Bayesian Multilevel Mediation: Evaluation of Inaccurate Priors in Latent 1-1-1 Designs

When latent constructs are measured by observed indicators from individuals nested within groups, multilevel structural equation modeling (MSEM) for 1-1-1 mediation designs allows researchers to simultaneously test indirect effects at each level of the data structure. However, with small samples (i.e., few clusters and/or small cluster sizes), such complex mediation models often run into estimation problems like nonconvergence, biased estimates, and insufficient power. Although Bayesian estimation with accurate informative priors can help alleviate these problems, it is unrealistic in practice to assume priors are correctly specified at the true population value. This study evaluates the performance of inaccurate (informative) priors in 1-1-1 MSEM mediation under varying sample sizes, ICCs, and effect sizes. Results indicate that while within-level indirect effect estimates are somewhat robust to inaccurate priors, between-level estimates are severely impacted, especially at small sample sizes. Implications and recommendations for conducting 1-1-1 MSEM mediation with Bayesian methods are discussed.

Impact of limited sample size and follow-up on single event survival extrapolation for health technology assessment: a simulation study

IntroductionExtrapolation of time-to-event data from clinical trials is commonly used in decision models for health technology assessment (HTA). The objective of this study was to assess performance of standard parametric survival analysis techniques for extrapolation of time-to-event data for a single event from clinical trials with limited data due to small samples or short follow-up.MethodsSimulated populations with 50,000 individuals were generated with an exponential hazard rate for the event of interest. A scenario consisted of 5000 repetitions with six sample size groups (30–500 patients) artificially censored after every 10% of events observed. Goodness-of-fit statistics (AIC, BIC) were used to determine the best-fitting among standard parametric distributions (exponential, Weibull, log-normal, log-logistic, generalized gamma, Gompertz). Median survival, one-year survival probability, time horizon (1% survival time, or 99th percentile of survival distribution) and restricted mean survival time (RMST) were compared to population values to assess coverage and error (e.g., mean absolute percentage error).ResultsThe true exponential distribution was correctly identified using goodness-of-fit according to BIC more frequently compared to AIC (average 92% vs 68%). Under-coverage and large errors were observed for all outcomes when distributions were specified by AIC and for time horizon and RMST with BIC. Error in point estimates were found to be strongly associated with sample size and completeness of follow-up. Small samples produced larger average error, even with complete follow-up, than large samples with short follow-up. Correctly specifying the event distribution reduced magnitude of error in larger samples but not in smaller samples.ConclusionsLimited clinical data from small samples, or short follow-up of large samples, produce large error in estimates relevant to HTA regardless of whether the correct distribution is specified. The associated uncertainty in estimated parameters may not capture the true population values. Decision models that base lifetime time horizon on the model’s extrapolated output are not likely to reliably estimate mean survival or its uncertainty. For data with an exponential event distribution, BIC more reliably identified the true distribution than AIC. These findings have important implications for health decision modelling and HTA of novel therapies seeking approval with limited evidence.

Personalising cancer medicine with prognostic markers.

Personalising cancer medicine with prognostic markers.

Estimation Under Mode Effects and Proxy Surveys, Accounting for Non-ignorable Nonresponse

We propose a new, model-based methodology to address two major problems in survey sampling: The first problem is known as mode effects, under which responses of sampled units possibly depend on the mode of response, whether by internet, telephone, personal interview, etc. The second problem is of proxy surveys, whereby sampled units respond not only about themselves but also for other sampled. For example, in many familiar household surveys, one member of the household provides information for all other members, possibly with measurement errors. Ignoring the existence of mode effects and/or possible measurement errors in proxy surveys could result in possible bias in point estimators and subsequent inference. Our approach accounts also for nonignorable nonresponse. We illustrate the proposed methodology by use of simulation experiments and real sample data, with known true population values.

Estimation of Forest Disturbance from Retrospective Observations in a Broad-Scale Inventory

Understanding the extent and timing of forest disturbances and their impacts is critical to formulating effective management and policy responses. Broad-scale inventory programs provide key estimates of forest parameters that indicate the extent and severity of disturbance impacts. Here, we review the use of a post-stratified estimator in a panelized design, in the context of disturbance observations that are collected retrospectively. We further develop a sample weight adjustment that is requisite for proper estimation of the extent and timing of disturbances. Using populations from areas of Arkansas, California, and Maine in the US, the weight adjustment technique was tested in a Monte Carlo simulation. We found that the estimated area of disturbance using the weight adjustment technique had satisfactory agreement with the true population values and performed considerably better than the conventional post-stratified estimation approach. The proliferation of panelized forest inventory designs globally suggests that accurate estimates of areal extent and timing of disturbances will often require that weighting adjustment techniques be employed in the estimation process.

How to carry out conceptual properties norming studies as parameter estimation studies: Lessons from ecology.

Conceptual properties norming studies (CPNs) ask participants to produce properties that describe concepts. From that data, different metrics may be computed (e.g., semantic richness, similarity measures), which are then used in studying concepts and as a source of carefully controlled stimuli for experimentation. Notwithstanding those metrics' demonstrated usefulness, researchers have customarily overlooked that they are only point estimates of the true unknown population values, and therefore, only rough approximations. Thus, though research based on CPN data may produce reliable results, those results are likely to be general and coarse-grained. In contrast, we suggest viewing CPNs as parameter estimation procedures, where researchers obtain only estimates of the unknown population parameters. Thus, more specific and fine-grained analyses must consider those parameters' variability. To this end, we introduce a probabilistic model from the field of ecology. Its related statistical expressions can be applied to compute estimates of CPNs' parameters and their corresponding variances. Furthermore, those expressions can be used to guide the sampling process. The traditional practice in CPN studies is to use the same number of participants across concepts, intuitively believing that practice will render the computed metrics comparable across concepts and CPNs. In contrast, the current work shows why an equal number of participants per concept is generally not desirable. Using CPN data, we show how to use the equations and discuss how they may allow more reasonable analyses and comparisons of parameter values among different concepts in a CPN, and across different CPNs.

Interpretation of clinical trial results: a committee opinion

This document provides guidance, background, and tips on how to recognize quality trials and focuses on evaluating the validity, importance, and relevance of clinical trial results. This document replaces the document of the same name, last published in 2008 (Fertil Steril® 2008;90:S114-20).

A meta-analysis of procedures to change implicit measures.

Using a novel technique known as network meta-analysis, we synthesized evidence from 492 studies (87,418 participants) to investigate the effectiveness of procedures in changing implicit measures, which we define as response biases on implicit tasks. We also evaluated these procedures' effects on explicit and behavioral measures. We found that implicit measures can be changed, but effects are often relatively weak (|ds| < .30). Most studies focused on producing short-term changes with brief, single-session manipulations. Procedures that associate sets of concepts, invoke goals or motivations, or tax mental resources changed implicit measures the most, whereas procedures that induced threat, affirmation, or specific moods/emotions changed implicit measures the least. Bias tests suggested that implicit effects could be inflated relative to their true population values. Procedures changed explicit measures less consistently and to a smaller degree than implicit measures and generally produced trivial changes in behavior. Finally, changes in implicit measures did not mediate changes in explicit measures or behavior. Our findings suggest that changes in implicit measures are possible, but those changes do not necessarily translate into changes in explicit measures or behavior. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

In search of an entity resolution OASIS

Entity resolution (ER) presents unique challenges for evaluation methodology. While crowdsourcing platforms acquire ground truth, sound approaches to sampling must drive labelling efforts. In ER, extreme class imbalance between matching and non-matching records can lead to enormous labelling requirements when seeking statistically consistent estimates for rigorous evaluation. This paper addresses this important challenge with the OASIS algorithm: a sampler and F-measure estimator for ER evaluation. OASIS draws samples from a (biased) instrumental distribution, chosen to ensure estimators with optimal asymptotic variance. As new labels are collected OASIS updates this instrumental distribution via a Bayesian latent variable model of the annotator oracle, to quickly focus on unlabelled items providing more information. We prove that resulting estimates of F-measure, precision, recall converge to the true population values. Thorough comparisons of sampling methods on a variety of ER datasets demonstrate significant labelling reductions of up to 83% without loss to estimate accuracy.

Stratification and sample allocation for reference burned area data

Statistical estimation protocols are one of the key means to ensure that independent and objective information on product accuracy is communicated to end-users. Methods for validating burned area products have been developed based on a probability sample of a space by time partitioning of the population. We extend this basic methodology to improve stratification and sample allocation, key elements of a sampling design used to collect burned area reference data. We developed and evaluated an approach to partition each year and biome into low and high burned area (BA) strata. Because the threshold used to separate the sampling units into low and high BA can vary by year and biome, this approach offers a more targeted stratification than used in previous studies for which a common threshold was applied to all biomes. A hypothetical population of validation data was then used to quantitatively compare the precision of accuracy estimates derived from different stratification and sample size allocation options. We evaluated two options that had been previously examined in the BA validation literature, and extended previous studies by adding two new options specifically developed for ratio estimates. Stratification based on mapped BA reduced standard errors of the global burned area accuracy estimates from one-half to one-eighth relative to standard errors of simple random sampling. Stratifying by mapped BA was also found to reduce standard errors of accuracy estimates for most year by biome strata indicating that this advantage of stratification and sample allocation applies generally to a range of conditions (i.e., biomes and years). The most precise estimates were obtained using a sample size per stratum allocation nh∝NhBA−h where Nh is the number of units in stratum h and BA−h is the mean mapped BA for stratum h. The best sampling design from our analyses was then used to select a set of 1,000 samples from a hypothetical population of validation data and confidence intervals were computed for each sample. Close to 95% of these confidence intervals contained the true population value thus confirming the validity of confidence intervals produced from the estimates and standard errors.

Biostatistics Series Module 1: Basics of Biostatistics.

Although application of statistical methods to biomedical research began only some 150 years ago, statistics is now an integral part of medical research. A knowledge of statistics is also becoming mandatory to understand most medical literature. Data constitute the raw material for statistical work. They are records of measurement or observations or simply counts. A variable refers to a particular character on which a set of data are recorded. Data are thus the values of a variable. It is important to understand the different types of data and their mutual interconversion. Biostatistics begins with descriptive statistics that implies summarizing a collection of data from a sample or population. Categorical data are described in terms of percentages or proportions. With numerical data, individual observations within a sample or population tend to cluster about a central location, with more extreme observations being less frequent. The extent to which observations cluster is summarized by measures of central tendency while the spread can be described by measures of dispersion. The confidence interval (CI) is an increasingly important measure of precision. When we observe samples, there is no way of assessing true population parameters. We can, however, obtain a standard error and use it to define a range in which the true population value is likely to lie with a certain acceptable level of uncertainty. This range is the CI while its two terminal values are the confidence limits. Conventionally, the 95% CI is used. Patterns in data sets or data distributions are important, albeit not so obvious, component of descriptive statistics. The most common distribution is the normal distribution which is depicted as the well-known symmetrical bell-shaped Gaussian curve. Familiarity with other distributions such as the binomial and Poisson distributions is also helpful. Various graphs and plots have been devised to summarize data and trends visually. Some plots, such as the box-and-whiskers plot and the stem-and-leaf plot are used less often but provide useful summaries in select situations.

Consumer Use of "Dr Google": A Survey on Health Information-Seeking Behaviors and Navigational Needs.

BackgroundThe Internet provides a platform to access health information and support self-management by consumers with chronic health conditions. Despite recognized barriers to accessing Web-based health information, there is a lack of research quantitatively exploring whether consumers report difficulty finding desired health information on the Internet and whether these consumers would like assistance (ie, navigational needs). Understanding navigational needs can provide a basis for interventions guiding consumers to quality Web-based health resources.ObjectiveWe aimed to (1) estimate the proportion of consumers with navigational needs among seekers of Web-based health information with chronic health conditions, (2) describe Web-based health information-seeking behaviors, level of patient activation, and level of eHealth literacy among consumers with navigational needs, and (3) explore variables predicting navigational needs.MethodsA questionnaire was developed based on findings from a qualitative study on Web-based health information-seeking behaviors and navigational needs. This questionnaire also incorporated the eHealth Literacy Scale (eHEALS; a measure of self-perceived eHealth literacy) and PAM-13 (a measure of patient activation). The target population was consumers of Web-based health information with chronic health conditions. We surveyed a sample of 400 Australian adults, with recruitment coordinated by Qualtrics. This sample size was required to estimate the proportion of consumers identified with navigational needs with a precision of 4.9% either side of the true population value, with 95% confidence. A subsample was invited to retake the survey after 2 weeks to assess the test-retest reliability of the eHEALS and PAM-13.ResultsOf 514 individuals who met our eligibility criteria, 400 (77.8%) completed the questionnaire and 43 participants completed the retest. Approximately half (51.3%; 95% CI 46.4-56.2) of the population was identified with navigational needs. Participants with navigational needs appeared to look for more types of health information on the Internet and from a greater variety of information sources compared to participants without navigational needs. However, participants with navigational needs were significantly less likely to have high levels of eHealth literacy (adjusted odds ratio=0.83, 95% CI 0.78-0.89, P<.001). Age was also a significant predictor (P=.02).ConclusionsApproximately half of the population of consumers of Web-based health information with chronic health conditions would benefit from support in finding health information on the Internet. Despite the popularity of the Internet as a source of health information, further work is recommended to maximize its potential as a tool to assist self-management in consumers with chronic health conditions.

Reporting Confidence Interval Instead of a Point Estimate a Review

In medical studies, investigators try to find out whether the difference of a measured outcome between groups is statistically significant by reporting a p value to reject or retain a null hypothesis. Such reporting of p value from comparisons of mean of the two groups has its limitations. Hypothesis testing does not take into account the variability of an observed sample statistic or its precision. Hence, it is important to quantify the uncertainty in this estimate by means of a confidence interval around the mean difference rather than reporting the mean difference simply as a point estimate. The confidence interval provides a range within which the true population value of the mean difference would lie with a certain degree of probability. Thus, a 95% confidence interval provides a range of mean differences that would contain the true population mean difference at least 95 times out of 100 repeated studies. The width of this range is determined by the sampling error or standard error. Greater the standard error, wider the range of the confidence interval. Conversely, a narrow range of confidence interval reflects a more precise study.

On accurate and precise generation of generalized Poisson variates

This article is concerned with comparison of a few random variate generation techniques for the generalized Poisson distribution. An evaluation is conducted on the degree of proximity between the estimates for its two distributional parameters and first four moments, and the specified or computed true population values via commonly accepted accuracy and precision measures in a simulated setting.

Accuracy of two techniques used to estimate pronghorn (Antilocapra americana) numbers in Chihuahuan Desert rangelands

Two techniques frequently used for monitoring pronghorn populations were compared in vegetation communities of the Chihuahuan Desert in New Mexico, USA. The evaluated techniques included the walking line transect and the driving line transect under two vegetation types and relative to season. The true population value (control) used to evaluate the techniques was a complete census using flights conducted over the entire study area (8000 ha). On grassland vegetation, the walking transect overestimates pronghorn density (n/km2), compared with the control (0.77 vs. 0.18 n/km2), as well as in the mesquite-grassland vegetation where the walking transect overestimates density (0.26 n/km2) compared with the control (0.11 n/km2). An increase in overestimation was found in the walking technique as the number of pronghorns in an area increased. In contrast, the driving technique showed no significant differences (P>0.05) compared with the control, in both grassland and mesquite-grassland pastures. It can be concluded that driving transect sampling is an accurate method to study pronghorn densities in the Chihuahuan Desert ecosystem, while walking transect sampling estimations need to be adjusted to obtain accurate population sizes.

Estimation of Population Percentiles

Percentiles play an important part in descriptive statistics of continuous data, and their use is recommended for reference interval estimation.1 We have selected various methods for the calculation of percentiles based on recommendations in the literature or use in popular software, and evaluated the accuracy of the percentile calculated in the sample as an estimate of the true population percentile2 using Monte-Carlo techniques. All selected methods calculate a rank or an index that points to a number in the sorted array of sample data, and linear interpolation is applied when the index does not correspond to an integer value. One method (method A1,3) calculates a rank or an index p(n+1) with p representing the centile (which is the percentile divided by 100) and n the sample size. Method B4 calculates an index 0.5+pn. Method C5 (commonly used in spreadsheets) uses p(n-1)+1 and method D6 uses p(n+1/3)+1/3. Details of the use of these 4 methods are given in the eAppendix (http://links.lww.com/EDE/A488). Experimental population data were obtained using a normal distribution pseudo-random number generator, programmed to generate a data set of 106 numbers with mean 0 and standard deviation 1. From our population data, 6 sets of 100 000 random samples each were drawn using a pseudo-random number generator with uniform distribution. Each of these sets consisted of 100 000 random samples with sample size 20, 120, 500 and 1000. The average of the 5th and 95th percentiles obtained with the 4 methods in these sample sets were calculated and compared with the population values. For each sample, the relative difference with the population values was expressed as a percentage, and the mean and standard deviation of these percentages were calculated. Next, the population data were transformed exponentially (base 10) to obtain a log-normal distribution and the experiments as described above were repeated. The results for 95th percentile in the normally distributed data are represented in table 1 (more comprehensive tables with figures are available in the eAppendix, http://links.lww.com/EDE/A488). The results for the 5th percentile were symmetrical to the results for the 95th percentile and are not shown. Method B presents the highest accuracy, followed by method D, A and C. Table 1 Accuracy of percentiles calculated in samples with various sizes from normally and log-normally distributed population data. Results are presented as average percentile, the average of the relative differences (%) and their standard deviation (SD). The results for 5th and 95th percentile in the log-normally distributed data are represented in the Table. For the 5th percentile, method A has a higher accuracy than the methods D, B and C, especially in small sample sizes, whereas for the 95th percentile method C presents the highest accuracy, followed by method B, D and A. We find that, for the calculation of percentiles, it may still be advantageous to transform log-normally distributed data. For example, the 95th percentile in the log-normal data should be about 44.16 (= 101.65). With method B and n = 20 we find an average of 114.69. But if we first log-transform the data we find, on average, 1.64, which back-transforms to 101.64 or 43.65, which is much closer to the true population value of 44.16. The effect of the log-transformation may be explained by the fact that linear interpolation is applied in the calculations of percentiles, and the transformation changes the distribution model within the interpolated interval. We conclude that method B is the preferred method in general for continuous data, taking into account the recommendation to transform the data to a normal distribution if necessary. Finally the large standard deviations of the observed differences illustrate the large statistical uncertainty associated with the estimated percentiles in small sample sizes. Therefore we stress the importance of reporting percentiles with their 95% confidence interval.

Statistics, probability, significance, likelihood: words mean what we define them to mean

Statisticians use words deliberately and specifically, but not necessarily in the way they are used colloquially. For example, in general parlance ‘statistics’ can mean numerical information, usually data. In contrast, one large statistics textbook defines the term ‘statistic’ to denote ‘a characteristic of a sample, such as the average score’, and the word ‘parameter’ to denote ‘a characteristic of a population’. However, for statisticians, statistics means more than just numerical information or a characteristic of a sample. Statistics is also a discipline. It is relevant to all areas of scientific enquiry, and spans study design, data collection, developing methods of analysis and analysing data, interpreting results, and making predictions. Other words that statisticians use carefully, and very particularly, form the cornerstones of statistical reasoning, words such as probability, significance, likelihood. Why does this matter? Well, it may well explain why non-statisticians struggle with the ideas and concepts used by statisticians. The reader is confused by the terminology, not the least because the statistician uses familiar words in different ways, a jargon similar to other professional vocabularies.

P - value and confidence intervals - facts and farces.

P - value and confidence intervals - facts and farces.