Validity Metrics Research Articles

Evaluation of automated anthropometrics produced by smartphone-based machine learning: a comparison with traditional anthropometric assessments.

Automated visual anthropometrics produced by mobile applications are accessible and cost effective with the potential to assess clinically relevant anthropometrics without a trained technician present. Thus, the aim of this study was to evaluate the precision and agreement of smartphone-based automated anthropometrics against reference tape measurements. Waist and hip circumference (WC; HC), waist:hip ratio (WHR) and waist:height ratio (W:HT) were collected from 115 participants (69 F) using a tape measure and two smartphone applications (MeThreeSixty®, myBVI®) across multiple smartphone types. Precision metrics were used to assess test-retest precision of the automated measures. Agreement between the circumferences produced by each mobile application and the reference were assessed using equivalence testing and other validity metrics. All mobile applications across smartphone types produced reliable estimates for each variable with intraclass correlation coefficients ≥ 0·93 (all P < 0·001) and root mean square coefficient of variation between 0·5 and 2·5 %. Precision error for WC and HC was between 0·5 and 1·9 cm. WC, HC, and W:HT estimates produced by each mobile application demonstrated equivalence with the reference tape measurements using 5 % equivalence regions. Mean differences via paired t-tests were significant for all variables across each mobile application (all P < 0·050) showing slight underestimation for WC and slight overestimation for HC which resulted in a lack of equivalence for WHR compared with the reference tape measure. Overall, the results of our study support the use of WC and HC estimates produced from automated mobile applications, but also demonstrates the importance of accurate automation for WC and HC estimates given their influence on other anthropometric assessments and clinical health markers.

Patient-Centered Communication (PCC) scale: Psychometric analysis and validation of a health survey measure.

Patient-centered communication (PCC) is one important component of patient-centered care and seen as a goal for most clinical encounters. Previous research has shown that higher PCC is related to an increase in healthy behaviors and less morbidity, among other outcomes. Given its importance, the National Cancer Institute (NCI) commissioned a monograph in 2007 to synthesize the existing literature on PCC and determine measurement objectives and strategies for measuring this construct, with a particular focus on cancer survivors. Based on this effort, a seven-item PCC scale was included on the Health Information National Trends Survey (HINTS), a probability-based survey of the US adult population. This study used HINTS data collected in 2018 to evaluate the psychometric properties of the PCC scale for the general US adult population including measures of reliability and validity. Through an exploratory factor analysis, the seven-item PCC scale was shown to be unidimensional with good internal consistency (Cronbach's alpha = .92). A confirmatory factor analysis verified the factor structure. Other construct validity metrics included known groups and discriminant validity. Known group comparisons were conducted for several sociodemographic factors and health self-efficacy confirming a priori assumptions. Discriminant validity tests with measures of social support and anxiety/depression showed relatively weak associations. The psychometric properties of this scale demonstrate its scientific utility for both surveillance research and other smaller-scale studies. Given its association with many health outcomes, it can also be used to better understand the dynamics in a clinical encounter.

Development and application of novel performance validity metrics for computerized neurocognitive batteries.

Data from neurocognitive assessments may not be accurate in the context of factors impacting validity, such as disengagement, unmotivated responding, or intentional underperformance. Performance validity tests (PVTs) were developed to address these phenomena and assess underperformance on neurocognitive tests. However, PVTs can be burdensome, rely on cutoff scores that reduce information, do not examine potential variations in task engagement across a battery, and are typically not well-suited to acquisition of large cognitive datasets. Here we describe the development of novel performance validity measures that could address some of these limitations by leveraging psychometric concepts using data embedded within the PennComputerized Neurocognitive Battery (PennCNB). We first developed these validity measures using simulations of invalid response patterns with parameters drawn from real data. Next, we examined their application in two large, independent samples: 1) children and adolescents from the Philadelphia Neurodevelopmental Cohort (n = 9498); and 2) adult servicemembers from the Marine Resiliency Study-II (n = 1444). Our performance validity metrics detected patterns of invalid responding in simulated data, even at subtle levels. Furthermore, a combination of these metrics significantly predicted previously established validity rules for these tests in both developmental and adult datasets. Moreover, most clinical diagnostic groups did not show reduced validity estimates. These results provide proof-of-concept evidence for multivariate, data-driven performance validity metrics. These metrics offer a novel method for determining the performance validity for individual neurocognitive tests that is scalable, applicable across different tests, less burdensome, and dimensional. However, more research is needed into their application.

Optimal Parameters for Spatial Distribution Modeling of Global Horizontal Solar Radiation in Iraq

Solar radiation plays an essential role in all interior physical processes and the radiative budget of the earth-atmosphere system. However, information about the spatial distribution of global solar radiation or its components is limited and exclusive at some scientific centers that specialized in solar radiation physics. Based on daily data of global solar radiation from (1984) to (2004) available at (16) locations in Iraq retrieved from NASA e-archive, this study aims to determine the optimal parameters of the experimental model to estimate the amount of global solar radiation on a horizontal surface as a function of latitude only. The results showed that the suggested model which is mathematically expressed by the sum of sine of latitude is highly appropriate for estimation of global solar radiation in clear sky conditions after adjusting the parameters according to geographical location. Through comparison between modeled output and an independent measured, globally used, and satellite available data and based on overall results besides some validity metrics: MBE, RMSE, pRMSD, NSE, R2, and r; It can be concluded the proposed model is achievable with high accurate practically performance to calculate and estimate the horizontal global solar radiation (GH) in Iraq region.

Measurement of respectful maternity care in exit interviews following facility childbirth: a criterion validity assessment in Nigeria

Background Valid methods for assessing women’s experiences of maternity care are essential to the global efforts toward providing positive childbirth environments for all women in all health facilities. Methods This criterion validity study used observation of childbirth as the reference standard compared to exit interviews with women upon discharge with a live baby, usually within 24 hours of childbirth. We investigated eight positive and sixteen negative maternity care experience indicators. Data were collected from ten primary healthcare facilities in Gombe State, northeast Nigeria, in August 2018 and August 2019. Data analysis involved tabulation of demographic characteristics of women and childbirth context, computation of individual level validity metrics including the area under the receiver operating curve (AUC) and estimating population level validity using the inflation factor (IF). Results A total of 724 women were observed and interviewed at the time of discharge (exit) following facility childbirth, 15% of whom were adolescents, 99% were married, 43% had no formal education, and a skilled birth attendant had attended only 12%. The prevalence of positive maternity care experience indicators ranged between 25% and 96% in childbirth observations. For these positive indicators, the agreement between childbirth observations and exit interviews ranged from 55% to 97%. Six of the eight positive maternity care experience indicators had high overall validity, meaning AUC≥0.70 and 0.75&lt;IF&lt;1.25, with high sensitivity (89% to 99%) and moderate to high specificity (44% to 84%). The prevalence of the 16 negative maternity care experience indicators ranged between 0.1% and 18% in childbirth observations. For these negative indicators, agreement ranged from 87% to 99%. Just six of the 16 negative maternity care experience indicators met the criteria for validation analysis; and these showed low to moderate sensitivity (32% to 74%), high specificity (97% to 100%) and moderate (0.60&lt;AUC&lt;0.70) to high (AUC≥0.70) individual-level validity. Conclusions In this high mortality setting with relatively low coverage of skilled attendance at birth, exit interviews with women following facility-based childbirth care provided responses consistent with the observation of childbirth for eight positive and six negative maternity care experience indicators.

Visual body composition assessment methods: A 4-compartment model comparison of smartphone-based artificial intelligence for body composition estimation in healthy adults

Visual body composition (VBC) estimates produced from smartphone-based artificial intelligence represent a user-friendly and convenient way to automate body composition remotely and without the inherent geographical and monetary restrictions of other body composition methods. However, there are limited studies that have assessed the reliability and agreement of this method and thus, the aim of this study was to evaluate VBC estimates compared to a 4-compartment (4C) criterion model. A variety of body composition assessments were conducted across 184 healthy adult participants (114F, 70M) including dual-energy X-ray absorptiometry and bioimpedance spectroscopy for utilization in the 4C model and automated assessments produced from two smartphone applications (Amazon Halo®, HALO; and myBVI®) using either Apple® or Samsung® phones. Body composition components were compared to a 4C model using equivalence testing, root mean square error (RMSE), and Bland-Altman analysis. Separate analyses by sex and racial/ethnic groups were conducted. Precision metrics were conducted for 183 participants using intraclass correlation coefficients (ICC), root mean squared coefficients of variation (RMS-%CV) and precision error (PE). Only %fat produced from HALO devices demonstrated equivalence with the 4C model although mean differences for HALO were <±1.0kg for FM and FFM. RMSEs ranged from 3.9% to 6.2% for %fat and 3.1-5.2kg for FM and FFM. Proportional bias was apparent for %fat across all VBC applications but varied for FM and FFM. Validity metrics by sex and specific racial/ethnic groups varied across applications. All VBC applications were reliable for %fat, fat mass (FM), and fat-free mass (FFM) with ICCs ≥0.99, RMS-%CV between 0.7% and 4.3%, and PEs between 0.3% and 0.6% for %fat and 0.2-0.5kg for FM and FFM including assessments between smartphone types. Smartphone-based VBC estimates produce reliable body composition estimates but their equivalence with a 4C model varies by the body composition component being estimated and the VBC being employed. VBC estimates produced by HALO appear to have the lowest error, but proportional bias and estimates by sex and race vary across applications.

An Analytical Model of College Students’ Mental Health Education Based on the Clustering Algorithm

This paper proposes an improved k-means clustering algorithm to analyze the mental health education of college students. It offers an improved k-means clustering algorithm with optimized centroid selection to address the problems of randomly selected class cluster centroids that lead to inconsistent algorithm results and easily fall into local optimal solutions of the traditional k-means clustering algorithm. The algorithm determines the neighborhood parameter based on the Euclidean distance between the data object and its nearest neighbor in the data set. It counts the object density based on the neighborhood parameter Eps. In the initial class cluster centroid selection phase, the algorithm randomly selects the first-class cluster centroid, and subsequent class cluster centroids are chosen based on the data object density information and the distance information between the data object and the existing class cluster centroids. The proposed improved k-means clustering algorithm and clustering validity metrics are tested using several simulated and real datasets. In this paper, the characteristics and application areas of the improved k-means clustering algorithm are sorted out, the self-determination theory related to the enhanced k-means clustering algorithm is investigated, and the behavior of the improved k-means clustering algorithm in the enhanced k-means clustering algorithm system and the octagonal behavior analysis method is also sorted out through the improved k-means clustering algorithm mental health management cases. The path of intervention in mental health education is designed through the improved k-means clustering algorithm. The intervention points are explained, including motivation discovery, mechanism setting, and component matching of the enhanced k-means clustering algorithm.

A scalable and unbiased discordance metric with H.

A standard unsupervised analysis is to cluster observations into discrete groups using a dissimilarity measure, such as Euclidean distance. If there does not exist a ground-truth label for each observation necessary for external validity metrics, then internal validity metrics, such as the tightness or separation of the clusters, are often used. However, the interpretation of these internal metrics can be problematic when using different dissimilarity measures as they have different magnitudes and ranges of values that they span. To address this problem, previous work introduced the "scale-agnostic" $G_{+}$ discordance metric; however, this internal metric is slow to calculate for large data. Furthermore, in the setting of unsupervised clustering with $k$ groups, we show that $G_{+}$ varies as a function of the proportion of observations assigned to each of the groups (or clusters), referred to as the group balance, which is an undesirable property. To address this problem, we propose a modification of $G_{+}$, referred to as $H_{+}$, and demonstrate that $H_{+}$ does not vary as a function of group balance using a simulation study and with public single-cell RNA-sequencing data. Finally, we provide scalable approaches to estimate $H_{+}$, which are available in the $\mathtt{fasthplus}$ R package.

Development and Validation of a Short Questionnaire Assessing the Behavior of Local Food Procurement in Quebec, Canada

ABSTRACTBackgroundVery few validated instruments, particularly screening tools applicable to large-cohort studies, are available to assess the behavior of local food procurement.ObjectiveThe aim was to develop and validate a short questionnaire that measures local food procurement in a sample of French-speaking adults from Quebec, Canada, and to assess the association between local food-procurement behavior and diet quality.MethodsA comprehensive questionnaire developed previously to measure local food procurement [Locavore-Index (Locavore-I)] was simplified through a series of steps that included face-validity, exploratory factor analysis, and reliability testing (internal consistency). Construct validity of the resulting short Locavore-I Short Form (Locavore-I-SF) was examined in a sample of 299 adults (85% women) from the Quebec City metropolitan community.ResultsThe Locavore-I-SF comprises 12 questions that measure the frequency of short food supply chain use (self-production, farmers’ markets, and community-supported agriculture box scheme) for 3 locally produced foods (carrot, tomato, and lettuce) as well as the geographical origin of those 3 foods. The Locavore-I-SF, which is scored on a 12-point scale, had a high internal consistency (Cronbach ɑ: 0.74). The Locavore-I-SF scores were strongly correlated with the reference scores obtained from the Locavore-I from which it was developed (r = 0.84, P < 0.0001). Locavore-I-SF scores also correlated (r = 0.50, P < 0.0001) with the geographical origin of foods measured by pictures of food labels taken by participants. Higher Locavore-I-SF scores were associated with behaviors consistent with eating local foods, such as gardening (vs. not gardening; mean ± SEM difference: 2.3 ± 0.4 points; P < 0.0001) and not being preoccupied by the foods’ appearance standards (vs. being preoccupied; 1.4 ± 0.4 points; P = 0.0002). Finally, the Locavore-I-SF scores were weakly associated with the Healthy Eating Food Index-2019 score (B = 0.05 ± 0.02; P = 0.02).ConclusionsThe Locavore-I-SF, a short questionnaire based on 3 locally produced foods in Quebec, measures the behavior of local food procurement with good reliability and acceptable validity metrics.

Comparative Analysis of the Ability of Machine Learning Models in Predicting In-hospital Postoperative Outcomes After Total Hip Arthroplasty.

Machine learning (ML) methods have shown promise in a wide range of applications including the development of patient-specific predictive models before surgical interventions. The purpose of this study was to develop, test, and compare four distinct ML models to predict postoperative parameters after primary total hip arthroplasty. Data from the Nationwide Inpatient Sample were used to identify patients undergoing total hip arthroplasty from 2016 to 2017. Linear support vector machine (LSVM), random forest (RF), neural network (NN), and extreme gradient boost trees (XGBoost) predictive of mortality, length of stay, and discharge disposition were developed and validated using 15 predictive patient-specific and hospital-specific factors. Area under the curve of the receiver operating characteristic (AUCROC) curve and accuracy were used as validity metrics, and the strongest predictive variables under each model were assessed. A total of 177,442 patients were included in this analysis. For mortality, the XGBoost, NN, and LSVM models all had excellent responsiveness during validation while RF had fair responsiveness. LSVM had the highest responsiveness with an AUCROC of 0.973 during validation. For the length of stay, the LSVM and NN models had fair responsiveness while the XGBoost and random forest models had poor responsiveness. LSVM had the highest responsiveness with an AUCROC of 0.744 during validation. For the discharge disposition outcome, LSVM had good responsiveness while the XGBoost, NN, and RF models all had fair responsiveness. LSVM had the highest responsiveness with an AUCROC of 0.801. The ML methods tested demonstrated a range of poor-to-excellent responsiveness and accuracy in the prediction of the assessed metrics, with LSVM being the best performer. Such models should be further developed, with eventual integration into clinical practice to inform patient discussions and management decision making, with the potential for integration into tiered bundled payment models.

The Identification of Subphenotypes and Associations with Health Outcomes in Patients with Opioid-Related Emergency Department Encounters Using Latent Class Analysis.

The emergency department (ED) is a critical setting for the treatment of patients with opioid misuse. Detecting relevant clinical profiles allows for tailored treatment approaches. We sought to identify and characterize subphenotypes of ED patients with opioid-related encounters. A latent class analysis was conducted using 14,057,302 opioid-related encounters from 2016 through 2017 using the National Emergency Department Sample (NEDS), the largest all-payer ED database in the United States. The optimal model was determined by face validity and information criteria-based metrics. A three-step approach assessed class structure, assigned individuals to classes, and examined characteristics between classes. Class associations were determined for hospitalization, in-hospital death, and ED charges. The final five-class model consisted of the following subphenotypes: Chronic pain (class 1); Alcohol use (class 2); Depression and pain (class 3); Psychosis, liver disease, and polysubstance use (class 4); and Pregnancy (class 5). Using class 1 as the reference, the greatest odds for hospitalization occurred in classes 3 and 4 (Ors 5.24 and 5.33, p < 0.001) and for in-hospital death in class 4 (OR 3.44, p < 0.001). Median ED charges ranged from USD 2177 (class 1) to USD 2881 (class 4). These subphenotypes provide a basis for examining patient-tailored approaches for this patient population.

On the validity of interpreting functional analyses of inappropriate mealtime behavior using structured criteria.

Visual inspection is the traditional method behavior analysts use to interpret functional-analysis results. Limitations of visual inspection include lack of standardized rules, subjectivity, and inconsistent interrater reliability (Fisch, 1998). To address these limitations, researchers have developed, evaluated, and refined structured criteria to aid interpretation of functional analyses of destructive behavior (Hagopian et al., 1997; Roane et al., 2013; Saini et al., 2018). The current study applied the structured criteria Saini et al. (2018) described to functional analyses of inappropriate mealtime behavior. We assessed its predictive validity and evaluated its efficiency relative to 3 post hoc visual inspection procedures. Validity metrics were lower than those in Saini et al. however, ongoing visual inspection increased the efficiency of functional analyses by more than 30%. We discuss these findings relative to the procedural differences between functional analyses of destructive behavior and inappropriate mealtime behavior.

On the application of clustering for extracting driving scenarios from vehicle data

If we want to extract test cases from driving data for the purpose of testing vehicles, we want to avoid using similar test cases. In this paper, we focus on this topic. We provide a method for extracting driving episodes from data utilizing clustering algorithms. This method starts with clustering driving data. Afterward, data points representing time-ordered sequences are obtained from the cluster forming a driving episode. Besides outlying the foundations, we present the results of an experimental evaluation where we considered six different clustering algorithms and available driving data from three German cities. To evaluate the cluster quality, we utilize three cluster validity metrics. In addition, we introduce a measure for the quality of extracted episodes relying on the Pearson coefficient. Experimental evaluation showed that the Pearson coefficient can rank clustering algorithms better than the three cluster validity metrics. We can extract meaningful episodes from driving data using any clustering algorithm considering four to eight clusters. Combining k-means clustering with auto-encoders leads to the best Pearson correlation. SOM is the slowest clustering method, and Canopy is the fastest.

Validation studies of rheumatoid arthritis patient-reported outcome measures in populations at risk for inequity: A systematic review and analysis using the OMERACT summary of measurement properties equity table

BackgroundExisting patient-reported outcome measures (PROMs) in rheumatoid arthritis (RA) may be limited in their applicability to populations that are at risk for inequities. We conducted a systematic review to identify and rate evidence in the validation studies for PROMs in populations at risk for inequity. MethodsA systematic review of MEDLINE and EMBASE was completed. The search strategy was developed to identify measurement property studies for PROMs of interest (selected pain, disease activity, global evaluation and quality of life scales) in patients with RA. We identified experimental, observational, and qualitative studies reporting analysis of feasibility, construct validity and discriminant ability metrics for populations at risk for inequity by various factors including race, ethnicity, culture or language; employment status; sex and gender identity; education level; socioeconomic status; social support; age; health literacy and disability. These were rated based on the OMERACT Summary of Measurement Properties Equity table. ResultsFrom 19,786 titles and abstracts screened, we identified 14 unique studies reporting validation metrics for pain (n = 3), DAS28-ESR or DAS28-CRP (n = 2), ACR20 (n = 1), patient global assessment (n = 2), EQ5D (n = 4), and PROMIS® (n = 3) by race (n = 10 studies), age (n = 6 studies), sex (n = 5 studies), education level (n = 2 studies), and disability, literacy, employment status, social support level and socioeconomic status (n = 1 study each). Five studies reported on feasibility, 12 reported construct validity metrics, and 4 studies reported on discriminant validity metrics. All studies by culture or language were rated as having good measurement property metrics. There was limited assessment of measurement property metrics for other populations at risk for inequity. ConclusionOur study highlights important gaps in patient representation in rheumatology research for accepted outcome measures. New outcome measures being developed for research purposes and clinical practice should ensure and report representation of patients from populations at risk for inequities in the testing of metrics of feasibility, construct validity and discriminant ability metrics.

An Enhanced K-means Firefly for Health Care Cluster Analysis of Philippines COVID-19 Datasets

The development of digital health technologies during the COVID-19 outbreak sets the foundation for different research initiatives, including cluster analysis to better present and interpret the COVID-19 dataset. One of the cluster analysis algorithms often used is the K-means that various researchers applied in different applications. The application includes, but is not limited to, analyzing tourism attractions and restaurants, developing a machine learning model to track the virus's progression, and categorizing a country's cities or provinces based on its COVID-19 records. This paper intends to extend the cluster analysis of the prior research by enhancing a novel clustering algorithm known as K-means Firefly, which has improved the conventional K-means. The proponents achieved their objectives and were able to solve the limitation of the algorithm by improving the initialization of the algorithm's data and clusters, specifically by employing Principal Component Analysis for dimension reduction of data, the Calinski-Harabasz Index for automatically determining the number of clusters, and K-means++ for obtaining the initial clusters. The enhanced algorithm clustered the cities in the Philippines based on their COVID-19 datasets containing 32 healthcare-related features. As a result, the algorithm can handle real-world datasets with multiple features by reducing its dimensionality. It can automatically determine the optimal number of clusters and the initial location of the centroids. In terms of internal validity metrics, the enhanced algorithm also performed better than the previous implementation, with a percentage difference of 90.16% for the Silhouette Coefficient, 72.99% for the Davies-Bouldin Index, and 68.98 for the Calinski-Harabasz Index. The proposed algorithm may use for various applications, such as data dashboards and real-time tracking websites, due to its extended dynamic features in handling datasets and producing substantial clusters.

Predictive capacity of four machine learning models for in-hospital postoperative outcomes following total knee arthroplasty

BackgroundMachine learning (ML) methods have shown promise in the development of patient-specific predictive models prior to surgical interventions. The purpose of this study was to develop, test, and compare four distinct ML models to predict postoperative parameters following primary total knee arthroplasty (TKA). MethodsData from the Nationwide Inpatient Sample was used to identify patients undergoing TKA during 2016–2017. Four distinct ML models predictive of mortality, length of stay (LOS), and discharge disposition were developed and validated using 15 predictive patient and hospital-specific factors. Area under the curve of the receiver operating characteristic curve (AUCROC) and accuracy were used as validity metrics, and the strongest predictive variables under each model were assessed. ResultsA total of 305,577 patients were included. For mortality, the XGBoost, neural network (NN), and LSVM models all had excellent responsiveness during validation, while random forest (RF) had fair responsiveness. For predicting LOS, all four models had poor responsiveness. For the discharge disposition outcome, the LSVM, NN, and XGBoost models had good responsiveness, while the RF model had poor responsiveness. LSVM and XGBoost had the highest responsiveness for predicting discharge disposition with an AUCROC of 0.747. DiscussionThe ML models tested demonstrated a range of poor to excellent responsiveness and accuracy in the prediction of the assessed metrics, with considerable variability noted in the predictive precision between the models. The continued development of ML models should be encouraged, with eventual integration into clinical practice in order to inform patient discussions, management decision making, and health policy.

To call or not to call: exploring the validity of telephone interviews to derive maternal self-reports of experiences with facility childbirth care in northern Nigeria

BackgroundTo institutionalise respectful maternity care, frequent data on the experience of childbirth care is needed by health facility staff and managers. Telephone interviews have been proposed as a low-cost alternative...

P057 THE ACCURACY OF INCISIONAL HERNIA CODES IN ADMINISTRATIVE DATA: A VALIDATION STUDY

Abstract Aim Administrative databases contain valuable information for studying incisional hernia (IH) following intra-abdominal procedures at a large scale. We assessed the validity of billing codes for the identification of IH in patients following abdominal surgery. Material and Methods Using International Classification of Diseases, Ninth and Tenth Revision, Clinical Modification (ICD-9-CM and ICD-10-CM), a random sample of 1,000 patients who underwent abdominal operations between 2006-2020 within a large health system-wide administrative database and ≥1 year of follow-up were screened for eligibility. Validation of IH codes was performed using the electronic medical record as reference. Validity metrics of interest included sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy. Patient factors associated with false positive (FP) and false negative (FN) were also explored within univariate and multivariate analyses. Results 759 patients were included. The sensitivity of IH codes was 94.7% (CI 91.4 – 97.0), specificity was 94.9% (CI 92.6 – 96.7), PPV was 84.4% (CI 78.5 – 88.9), NPV was 98.4 (CI 97.4 – 99.0), accuracy was 94.9% (CI 93.0 – 96.3). Within adjusted analyses, patients admitted to a non-surgical service (relative to patients admitted to surgical services; OR 4.46 [95% CI 1.06-18.66]; p = 0.04) were associated with FP; whereas every one-year increase in age was associated with a 5.0% (95% CI 1.0%-10.0%) increase in FN (p = 0.03). Conclusions We have validated the use of ICD-9-CM and ICD-10-CM codes for accurately identifying IH following abdominal surgery. This method yields &amp;gt;94% for key validity measures.

Relevance of Complaint Severity in Predicting the Progression of Subjective Cognitive Decline and Mild Cognitive Impairment: A Machine Learning Approach.

The presence of subjective cognitive complaints (SCCs) is a core criterion for diagnosis of subjective cognitive decline (SCD); however, no standard procedure for distinguishing normative and non-normative SCCs has yet been established. To determine whether differentiation of participants with SCD according to SCC severity improves the validity of the prediction of progression in SCD and MCI and to explore validity metrics for two extreme thresholds of the distribution in scores in a questionnaire on SCCs. Two hundred and fifty-three older adults with SCCs participating in the Compostela Aging Study (CompAS) were classified as MCI or SCD at baseline. The participants underwent two follow-up assessments and were classified as cognitively stable or worsened. Severity of SCCs (low and high) in SCD was established by using two different percentiles of the questionnaire score distribution as cut-off points. The validity of these cut-off points for predicting progression using socio-demographic, health, and neuropsychological variables was tested by machine learning (ML) analysis. Severity of SCCs in SCD established considering the 5th percentile as a cut-off point proved to be the best metric for predicting progression. The variables with the main role in conforming the predictive algorithm were those related to memory, cognitive reserve, general health, and the stability of diagnosis over time. Moderate to high complainers showed an increased probability of progression in cognitive decline, suggesting the clinical relevance of standard procedures to determine SCC severity. Our findings highlight the important role of the multimodal ML approach in predicting progression.

Tracking changes in body composition: comparison of methods and influence of pre-assessment standardisation.

The present study reports the validity of multiple assessment methods for tracking changes in body composition over time and quantifies the influence of unstandardised pre-assessment procedures. Resistance-trained males underwent 6 weeks of structured resistance training alongside a hyperenergetic diet, with four total body composition evaluations. Pre-intervention, body composition was estimated in standardised (i.e. overnight fasted and rested) and unstandardised (i.e. no control over pre-assessment activities) conditions within a single day. The same assessments were repeated post-intervention, and body composition changes were estimated from all possible combinations of pre-intervention and post-intervention data. Assessment methods included dual-energy X-ray absorptiometry (DXA), air displacement plethysmography, three-dimensional optical imaging, single- and multi-frequency bioelectrical impedance analysis, bioimpedance spectroscopy and multi-component models. Data were analysed using equivalence testing, Bland-Altman analysis, Friedman tests and validity metrics. Most methods demonstrated meaningful errors when unstandardised conditions were present pre- and/or post-intervention, resulting in blunted or exaggerated changes relative to true body composition changes. However, some methods - particularly DXA and select digital anthropometry techniques - were more robust to a lack of standardisation. In standardised conditions, methods exhibiting the highest overall agreement with the four-component model were other multi-component models, select bioimpedance technologies, DXA and select digital anthropometry techniques. Although specific methods varied, the present study broadly demonstrates the importance of controlling and documenting standardisation procedures prior to body composition assessments across distinct assessment technologies, particularly for longitudinal investigations. Additionally, there are meaningful differences in the ability of common methods to track longitudinal body composition changes.