False Negative Answers Research Articles

Identifying and Minimizing Errors in the Measurement of Early Childhood Development: Lessons Learned from the Cognitive Testing of the ECDI2030.

Challenges in measuring early childhood development (ECD) at scale have been documented, yet little is known about the specific difficulties related to questionnaire design and question interpretation. The purpose of this paper is to discuss the challenges of measuring ECD at scale in the context of household surveys and to show how to overcome them. The paper uses examples from the cognitive interviewing exercises that were conducted as part of the methodological work to develop a measure of ECD outcomes, the ECDI2030. It describes the methodological work carried out to inform the selection and improvement of question items and survey implementation tools as a fundamental step to reduce and mitigate systematic measurement error and improve data quality. The project consisted of a total of five rounds of testing, comprising 191 one-on-one, in-depth cognitive interviews across six countries (Bulgaria, India, Jamaica, Mexico, Uganda, and the USA). Qualitative data analysis methods were used to determine matches and mismatches between intention of items and false positives or false negative answers among subgroups of respondents. Key themes emerged that could potentially lead to systematic measurement error in population-based surveys on ECD: (1) willingness of child to perform task versus ability of child to perform task; (2) performing task versus performing task correctly; (3) identifying letters or numbers versus recognizing letters or numbers; (4) consistently performing task versus correctly performing task; (5) applicability of skills being asked versus observability of skills being asked; and (6) language production versus language comprehension. Through an iterative process of testing and subsequent revision, improvements were made to item wording, response options, and interviewer training instructions. Given the difficulties inherent in population-level data collection in the context of global monitoring, this study’s findings confirm the importance of cognitive testing as a crucial step in careful, culturally relevant, and sensitive questionnaire design and as a means to reduce response bias in cross-cultural contexts.

The Assessment of Risk Factors for Development of Disability in Children with Congenital Hypothyroidism in Uzbekistan within a Neonatal Screening

The aim of this study was to detect the most significant risk factors leading to disability in children with congenital hypothyroidism (CH) in the autonomous Republic of Karakalpakstan (RK) during neonatal screening (NS). Methods and Results: We used data of patients with CH registered within NS in the RK in 1998-2019 by the Center for Screening of Mother and Child. To predict and calculate the most significant risk factors for disability in children with CH, we used the method of normalizing intensive indicators by E. Shigan, based on the Bayes theorem. The study recruited 111 patients with CH aged from 2 months to 20 years. Among the patients, there were 79(71.2%) girls and 32(28.8%) boys. Additionally, 34(30.6%) children with CH had been disabled since childhood. The lack of compensation after the start of treatment had the highest and most significant degree of disability risk (RR=6.39, 95% CI: 7.4-1.2). Among patients diagnosed outside of screening, disability developed 4.1 times more often than with the results of NS (RR=4.0, 95% CI: 1.1-10.6). In CH patients diagnosed outside of screening, “absence of reagents” was a significant factor increasing the risk of disability by 6.1 times (RR=6.1, 95% CI: 1.8-11.2). Such risk factors as “home delivery” and “parental refusal of the primary test” increased the risk of disability by 3.4 times (RR=3.4, 95% CI: 2.5-8.4) and 1.6 times (RR=2.4, 95% CI: 2.93-7.12), respectively. The possible errors or false-negative answers in the “normal” secondary test and the “normal” primary test increased the risk of disability by 3.3 times (RR=4.0, 95% CI: 3.2-10.7) and 2.4 times (RR=2.42, 95% CI: 2.93-7.12), respectively. Factors such as the “late response to retesting” (RR=0.82 95% CI: 0.65-0.54), “late awareness on the part of the medical staff” (RR=0.29, 95% CI: 0.27- 0.08), and "parental refusal of treatment" (RR=1.03, 95% CI: 0.81-0.84) showed less significance in patients' disability. The “starting treatment after 1 month” factor was 4.2 times more likely to result in disability than “starting treatment before 1 month” (RR=4.2, 95% CI: 4.5 -1.1). Cancellation of levothyroxine by parents for children up to 3 years of age and cancellation of treatment by parents after 3 years more likely resulted in disability by 1.4 times (RR=1.43, 95% CI: 1.4 -2.01) and 3.3 times (RR=3.33, 95% CI: 3.3-10.9), respectively. Conclusion: the most significant risk factors for the development of disability in children with CH in the RK were (in descending order): no compensation after starting treatment, no reagents for screening, starting treatment after 1 year, diagnostics outside of screening, cancellation of L-T4 by parents before and after age 3 years, false-negative secondary TSH test, false-negative primary test, parents refusing the primary test, and childbirth at home.

Investigation of various factors influencing Raman spectra interpretation with the use of likelihood ratio approach

The main aim of this study was to verify whether selected analytical parameters may affect solving the comparison problem of Raman spectra with the use of the likelihood ratio (LR) approach. Firstly the LR methodologies developed for Raman spectra of blue automotive paints obtained with the use of 785nm laser source (results published by the authors previously) were implemented for good quality spectra recorded for these paints with the use of 514.5nm laser source. For LR models construction two types of variables were used i.e. areas under selected pigments bands and coefficients derived from discrete wavelet transform procedure (DWT).Few experiments were designed for 785nm and 514.5nm Raman spectra databases after constructing well performing LR models (low rates of false positive and false negative answers and acceptable results of empirical cross entropy approach).In order to verify whether objective magnification described by its numerical aperture affects spectra interpretation, three objective magnifications —20×(N.A.=0.4.), 50×(N.A.=0.75) and 100×(N.A.=0.85) within each of the applied laser sources (514.5nm and 785nm) were tested for a group of blue solid and metallic automotive paints having the same sets of pigments depending on the applied laser source. The findings obtained by two types of LR models indicate the importance of this parameter for solving the comparison problem of both solid and metallic automotive paints regardless of the laser source used for measuring Raman signal. Hence, the same objective magnification, preferably 50× (established based on the analysis of within- and between-samples variability and F-factor value), should be used when focusing the laser on samples during Raman measurements.Then the influence of parameters (laser power and time of irradiation) of one of the recommended fluorescence suppression techniques, namely photobleaching, was under investigation. Analysis performed on a group of solid automotive paint samples showed that time of irradiation upon established laser power does not affect solving the comparison problem with the use of LR test. Likewise upon established time of irradiation 5% or 10% laser power could be used interchangeably without changing conclusions within this problem. However, upon the established time of irradiation changes in laser power between control and recovered sample from 5% or 10% to 50% may cause erroneous conclusions. Additionally it was also proved that prolonged irradiation of paint does not quantitatively affect pigments bands areas revealed after such a pre-treatment.

Evidential value of polymeric materials-chemometric tactics for spectral data compression combined with likelihood ratio approach.

Polymers have become a ubiquitous element of our culture. Therefore, these materials may play an important role in forensic investigations, serving as mute witnesses of occurrences such as car accidents. In this study, the possibilities provided by the likelihood ratio (LR) approach to estimate the evidential value of observed similarities and differences, and to discriminate among NIR spectral data originating from polypropylene automotive parts and household items, were investigated. Since the construction of LR models requires the introduction of only a few variables, the main objective was to reduce the dimensionality of registered spectra, which are characterised by over a thousand variables. The applied strategy was based on compression of NIR signals using discrete wavelet transform (DWT) followed by use of the SELECT algorithm for the selection and decorrelation of the most informative DWT coefficients. Selected features eventually served as an input for LR models. The performance of the developed models was assessed by measuring the rates of false positive and false negative answers as well as by applying an empirical cross entropy approach. Despite relatively small databases of polymeric objects, both univariate and multivariate LR models showed acceptable performances. The latter, however, gave the most satisfactory results, as it enabled successful discrimination of compared samples and delivered the lowest error rates. In addition, in order to verify the potential of NIR spectroscopy, the obtained results were compared with those obtained after application of the proposed tactics to the FTIR data, which is a well-established method in the forensic sphere.

Interpretation of FTIR spectra of polymers and Raman spectra of car paints by means of likelihood ratio approach supported by wavelet transform for reducing data dimensionality.

The problem of interpretation of common provenance of the samples within the infrared spectra database of polypropylene samples from car body parts and plastic containers as well as Raman spectra databases of blue solid and metallic automotive paints was under investigation. The research involved statistical tools such as likelihood ratio (LR) approach for expressing the evidential value of observed similarities and differences in the recorded spectra. Since the LR models can be easily proposed for databases described by a few variables, research focused on the problem of spectra dimensionality reduction characterised by more than a thousand variables. The objective of the studies was to combine the chemometric tools easily dealing with multidimensionality with an LR approach. The final variables used for LR models' construction were derived from the discrete wavelet transform (DWT) as a data dimensionality reduction technique supported by methods for variance analysis and corresponded with chemical information, i.e. typical absorption bands for polypropylene and peaks associated with pigments present in the car paints. Univariate and multivariate LR models were proposed, aiming at obtaining more information about the chemical structure of the samples. Their performance was controlled by estimating the levels of false positive and false negative answers and using the empirical cross entropy approach. The results for most of the LR models were satisfactory and enabled solving the stated comparison problems. The results prove that the variables generated from DWT preserve signal characteristic, being a sparse representation of the original signal by keeping its shape and relevant chemical information.

PS-039 DokuPIK – Pharmaceutical interventions: Do ward pharmacists know how to document?

BackgroundThe internet-based categorical, hierarchical documentation system for routine pharmaceutical interventions (PIs) DokuPIK is widely used especially among German ward-based pharmacists.PurposeTo conduct a survey-based study on its clinical usability.Materials and methodsOut...

Factors That Influence Standard Automated Perimetry Test Results in Glaucoma: Test Reliability, Technician Experience, Time of Day, and Season

To determine the influence of several factors on standard automated perimetry test results in glaucoma. Longitudinal Humphrey field analyzer 30-2 Swedish interactive threshold algorithm data from 160 eyes of 160 glaucoma patients were used. The influence of technician experience, time of day, and season on the mean deviation (MD) was determined by performing linear regression analysis of MD against time on a series of visual fields and subsequently performing a multiple linear regression analysis with the MD residuals as dependent variable and the factors mentioned above as independent variables. Analyses were performed with and without adjustment for the test reliability (fixation losses and false-positive and false-negative answers) and with and without stratification according to disease stage (baseline MD). Mean follow-up was 9.4 years, with on average 10.8 tests per patient. Technician experience, time of day, and season were associated with the MD. Approximately 0.2 dB lower MD values were found for inexperienced technicians (P < 0.001), tests performed after lunch (P < 0.001), and tests performed in the summer or autumn (P < 0.001). The effects of time of day and season appeared to depend on disease stage. Independent of these effects, the percentage of false-positive answers strongly influenced the MD with a 1 dB increase in MD per 10% increase in false-positive answers. Technician experience, time of day, season, and the percentage of false-positive answers have a significant influence on the MD of standard automated perimetry.

Evidential value of physicochemical data—comparison of methods of glass database creation

AbstractGlass databases used in forensic laboratories are commonly generated by collection of samples from a very small area of many different objects (Method 1). These frequently yield univariate data in the form of refractive indices measured using the Glass Refractive Index Measurement (GRIM) technique, or multivariate data in the form of elemental compositions obtained by the SEM‐EDX technique. We investigate whether the within‐ and between‐object variances and covariances estimated from data collected using Method 1 are as reliable as those obtained from databases formed by samples collected from several areas of an object (Method 2). Reliability of the Method 1‐ and Method 2‐estimated parameters is evaluated in terms of the performance of likelihood ratio (LR) models that measure the evidential value of physicochemical data for forensic purposes. Two‐level random effect models assuming that the within‐object distribution is normal, and that the between‐object distribution is either normal or obtained by kernel density estimation (KDE), are applied to comparison of glass fragments of known origin, and the rates of false positive and false negative answers recorded. These rates are similar for Methods 1 and 2 for the refractive index (RI) data, and Method 1 performs better than Method 2 for elemental composition data, suggesting that the current method of database generation is appropriate for the estimation of these sources of variability for glass samples. Copyright © 2010 John Wiley &amp; Sons, Ltd.

A Two‐Level Model for Evidence Evaluation in the Presence of Zeros*

Likelihood ratios (LRs) provide a natural way of computing the value of evidence under competing propositions. We propose LR models for classification and comparison that extend the ideas of Aitken, Zadora, and Lucy and Aitken and Lucy to include consideration of zeros. Instead of substituting zeros by a small value, we view the presence of zeros as informative and model it using Bernoulli distributions. The proposed models are used for evaluation of forensic glass (comparison and classification problem) and paint data (comparison problem). Two hundred and sixty-four glass samples were analyzed by scanning electron microscopy, coupled with an energy dispersive X-ray spectrometer method and 36 acrylic topcoat paint samples by pyrolysis gas chromatography hyphened with mass spectrometer method. The proposed LR model gave very satisfactory results for the glass comparison problem and for most of the classification tasks for glass. Results of comparison of paints were also highly satisfactory, with only 3.0% false positive answers and 2.8% false negative answers.

The sensitivity and specificity of scanning laser polarimetry in the detection of glaucoma in a clinical setting

A study was conducted to determine the sensitivity and specificity of masked experts in classifying results obtained from normal eyes, glaucomatous eyes, and eyes suspicious for glaucoma using the Nerve Fiber Analyzer GDx (Laser Diagnostic Technologies, San Diego, CA), a scanning polarimeter. Retrospective consecutive case series. Consecutive subjects seen in a glaucoma clinic were imaged with the GDX: The masked printouts were distributed to nine experienced users of the machine, who were asked to classify the scan from each eye as normal, suspicious for glaucoma, or consistent with glaucoma. No specific guidelines for interpretation were used. The interpretations were compared with the known diagnoses obtained from review of the subjects' medical records. The number of correct diagnoses, false-positive answers (specificity), and false-negative answers (sensitivity) were noted for each observer and averaged for all nine. Interobserver agreement was calculated by use of the kappa statistic. False-positive and false-negative responses from reviewers' assessments compared with known diagnoses from chart review. A total of 104 eyes was included: 29 normal eyes, 33 eyes from glaucoma suspects, and 42 eyes with glaucoma. Interobserver agreement was moderate for normals (kappa = 0.42, P = 0.00) and glaucoma (kappa = 0.48, P = 0.00), but slight for glaucoma suspects (kappa = 0.09, P = 0.00). Overall, the average sensitivity was 74% and the specificity was 74%. If only considering whether or not the scan represented glaucoma (i.e., reviewer assessment of "suspect" meant "not glaucoma"), the average sensitivity was 86% and the specificity 83%. If only considering whether a scan was normal or not (i.e., reviewer assessment of "suspect" meant "abnormal"), the sensitivity was 83% and the specificity 86%. Eliminating glaucoma suspects (but allowing the assessment of "suspect") increased overall sensitivity to 80% and specificity to 79%. With suspects eliminated, sensitivity was 80% and specificity 91% for determination of glaucoma, and sensitivity was 93% and specificity 79% for determination of abnormality. Scanning laser polarimetry may have good sensitivity and specificity for separating normal from abnormal eyes, but it is not as good for classifying unknown subjects when glaucoma suspects are included. Evaluation of new parameters and continued clinical experience may help develop guidelines for interpretation and/or specific indicators for glaucomatous damage to increase the sensitivity and specificity.

Tracing Lineage of Array Data

Arrays are a common and important class of data in many applications. Arrays can model data such as digital images, digital video, scientific and experimental data, matrices, and finite element grids. Although array manipulations are diverse and domain-specific, they often exhibit structural regularities. This paper describes an algorithm called sub-pushdown to trace data lineage in such array computations. Lineage tracing is a type of data-flow analysis that relates parts of a result array to those parts of the argument (base) arrays that have bearings on the result array parts. Sub-pushdown can be used to trace data lineage in array-manipulating computations expressed in the Array Manipulation Language (AML) that was introduced previously. Sub-pushdown has several useful features. First, the lineage computation is expressed as an AML query. Second, it is not necessary to evaluate the AML lineage query to compute the array data lineage. Third, sub-pushdown never gives false-negative answers. Sub-pushdown has been implemented as part of the ArrayDB prototype array database system that we have built.

False-negative responses in glaucoma perimetry: indicators of patient performance or test reliability?

PURPOSE. To study whether false-negative answers in computerized glaucoma perimetry indicate the patient's ability to perform perimetry or test result reliability. METHODS. A retrospective evaluation was performed of visual field test results obtained with a perimetry program (Humphrey 30-2 Sita Standard; Humphrey Instruments, San Leandro, CA) in 70 consecutive patients with unilateral glaucomatous field loss. Frequencies of false-negative answers were compared between the two eyes of each patient and related to amount of visual field damage in the glaucomatous eyes using linear regression analysis. RESULTS. Frequencies of false-negative answers were higher in eyes with field loss. The intrapatient intereye difference was 6.6% on average (P < 0.0001). In seven subjects with false-negative frequency of 5% or more in both eyes, the mean difference was 12.7% between eyes. The differences in false-negative answers depended significantly on the amount of field loss in the glaucomatous eyes (P = 0.0003). Larger differences were seen in patients with advanced field loss in the affected eye. CONCLUSIONS. The increased frequencies of false-negative answers in eyes with field loss were strongly associated with field status. The higher false-negative frequencies in eyes with glaucomatous field loss compared with unaffected eyes may be explained by the increased variability in threshold values typically found in such eyes. False-negative answers in patients with glaucoma therefore represent eye rather than patient status.

A Single Question about Prior Stroke versus a Stroke Questionnaire to Assess Stroke Prevalence in Populations

Background: False-positive and false-negative answers to screening questions influence prevalence and incidence estimations for stroke in population studies. Despite frequent use in screening, only a few studies have examined causes and influence of incorrect self-reports. We compared the rates of false-positive and false-negative answers to a single question about prior stroke to those of the Stroke Symptom Questionnaire (SSQ), a newly developed instrument based on 6 symptom questions. Differences in stroke prevalence estimations and risk factors for incorrect reports are described. Methods: The MEMO study (Memory and Morbidity in Augsburg Elderly) examines cognitive function and neurodegenerative diseases in an elderly population (n = 384) in southern Germany. All participants filled in the symptom questionnaire, received a neurological examination and a neuropsychological test battery. Medical records were obtained for event validation of subjects positive on screening and those negative on screening with symptoms suggesting a cerebrovascular event during examination. Results: Prevalence of total stroke was 5.3% using a single screening question and 6.8% using the questionnaire. The false-negative rate was higher for the single-question approach (34.2 versus 10.5%). It was strongly influenced by gender and cognitive function. The questionnaire had a higher false-positive rate than the single question. Based on the results, we established question combinations that best served three different research scenarios (frequency estimation, risk factor analysis, control selection), relevant to stroke research in population studies. Conclusions: A single screening question for stroke in the past with event validation by medical records underestimates stroke frequency in population studies by about 30%. Use of a number of questions for key symptoms combined with a general stroke question, as in the SSQ, improves the completeness of event ascertainment and allows the detection of stroke and transient ischemic attack at the same time.

Semantics and verification of action diagrams with linear timing

Specifications containing linear timing constraints, such as found in action diagrams (timing diagrams) defining interface behaviors, are often used in practice. Although efficient O(n 3 ) shortest path algorithms exist for computing the minimum and maximum time distances between actions, subject to the timing constraints, there is so far no accurate method that can decide (a) whether a specification of this kind is realizable (i.e., can be simulated by a causal system), and (b) given the action diagrams of the interfaces of two or more communicating systems, whether the systems implementing such independent specifications will correctly interoperate (i.e., satisfy the respective protocols and timing assumptions). First we illustrate the weakness of existing action diagram verification techniques: the causality issue is not addressed, and the proposed methods to answer the compatibility (interoperability) question yield false negative answers in many practical situations. We then define the meaning of causality in an action diagram specification and state a set of sufficient conditions for causality to hold. This development then leads to an exact procedure for the verification of the interface compatibility of communicating action diagrams. the results are illustrated on a practical example.

Feasibility of automated visual field examination in children between 5 and 8 years of age.

To investigate how young children develop the ability to undergo a visual field evaluation using regular automated perimetry. The study included 42 normal girls aged 5, 6, 7, and 8 years. Twelve locations in the 15 degrees eccentricity were tested in one eye, using an Octopus 2000R perimeter with a two level strategy. False positive and false negative catch trials were presented. The examination was performed three times in succession. Before the examination procedure, a specially designed programme was conducted for progressive familiarisation. During the familiarisation procedure, it was found that all of the 5-year-old children, seven of the 6-year-old children, and three of the 7-year-old children were unable to perform immediately, and correctly, the instructions given during the familiarisation phase; these children took from 30 seconds to 3 minutes to comply with the examiner's requests. With the exception of one 5-year-old child, all tested subjects completed the planned procedure. The mean proportion of false negative answers in catch trials was 1.6%. The mean proportion of false positive answers was 12.2%. The quadratic dependency on age suggested by the averages was not significant (F(3,116) = 0.88; p = 0.45). Detection stimulus improved with age, as shown by the fact that probability of perceiving dim stimulus increases significantly (F(3,116) = 12.68; p < 0.0001). Children did remarkably well regarding both the duration of the examination and the reliability of the answers. A preliminary familiarisation phase with a specially designed adaptation programme was found to be mandatory with children aged 7 or under. To our knowledge, this is the first time that such an investigation has been performed.

Incidence of Minamata disease in communities along the Agano river, Niigata, Japan--patterns of the exposure and official diagnosis of patients

Data obtained from all residents, who lived in the Agano river villages where Minamata disease occurred as of June 1965, were used to analyze its incidence patterns and to evaluate validity of the official recognition of the disease. The residents totaled 18,253 (8,911 males and 9,342 females), including 262 (151 males and 111 females) recognized patients and 378 (186 males and 192 females) rejected applicants. Consumption of the contaminated river fish was associated with hair Hg measurements (ppm), but there were few residents who denied the consumption but showed elevated Hg values, probably due to false-negative answers in the census survey in 1965. Incidence rates of the recognized patients increased with the upgrading codes for fish ingestion, but analysis of the association of fish ingestion among the recognized applicants indicated that there were ca. 50% false-positive diagnosis. This analysis showed that The Government made best efforts to help very mild cases despite the fact that such a generous attitude inevitably causes overdiagnosis. Incidence rates of the rejected applicants also showed a dose-response. This is extremely important because a rejection meant that compensation was refused by the polluting company. Protest groups insisted the existence of Minamata disease showing only sensory symptoms, but this concept has not been accepted by the Government and the polluters. The present study disclosed a dose-response relationship in the rejected cases giving positive ground for the opinions proposed by protest groups. Proportions of such "Hg-associated sensory disorders" (cases with unexplained sensory disorders which are associated with river fish ingestion) were 15-30% among the exposed applicants, 6.6-15% among the all rejected applicants, and 3.9-10% among the unexplained sensory disorders in the regional population. Owing to the river pollution, there was 5-11% increase of the unexplained sensory disorders among the local residents, in addition to recognizable Minamata disease.