Chapter 6 - An empirical model for strategic network formation

We develop and analyze a tractable empirical model for strategic network formation that can be estimated with data from a single network at a single point in time. We model the network formation as a sequential process where in each period a single randomly selected pair of agents has the opportunity to form a link. Conditional on such an opportunity, a link will be formed if both agents view the link as beneficial to them. They base their decision on their own characateristics, the characteristics of the potential partner, and on features of the current state of the network, such as whether the two potential partners already have friends in common. A key assumption is that agents do not take into account possible future changes to the network. This assumption avoids complications with the presence of multiple equilibria, and also greatly simplifies the computational burden of anlyzing these models. We use Bayesian markov-chainmonte-carlo methods to obtain draws from the posterior distribution of interest. We apply our methods to a social network of 669 high school students, with, on average, 4.6 friends. We then use the model to evaluate the effect of an alternative assignment to classes on the topology of the network.

We make four main contributions to the theory of network formation. (1) The problem of network formation with farsighted agents can be formulated as an abstract network formation game. (2) In any farsighted network formation game the feasible set of networks contains a unique, finite, disjoint collection of nonempty subsets having the property that each subset forms a strategic basin of attraction. These basins of attraction contain all the networks that are likely to emerge and persist if individuals behave farsightedly in playing the network formation game. (3) A von Neumann Morgenstern stable set of the farsighted network formation game is constructed by selecting one network from each basin of attraction. We refer to any such von Neumann-Morgebstern stable set as farsighted basis. (4) The core of the farsighted network formation games is constructed by selecting one network from each basin of attraction containing a single network. We call this notion of the core, the farsighted core. We conclude that the farsighted core is nonempty if and only if there exists one farsighted basin of attraction containing a single network. To relate our three equilibrium and stability notions (basins of attraction, farsighted basis and farsighted core) to recent work by Jackson and Wolinsky (1996), we define a notion of pairwise stability similar to the Jackson-Wolinsky notion and we show that a farsighted core is contained in the set of pairwise stable networks. Finally, we introduce, via an example, competitative contracting networks and highlight how the analysis of these networks requires the new features of our network formation model.

Guided by Kraemer et al.'s (Psychological Methods, 3:257-271, 1999) framework for measuring the potency of risk factors, we sought to improve on the classification accuracy reported in Petras et al. (Journal of the American Academy of Child and Adolescent Psychiatry 43:88-96, 2004a) and Petras et al. (Journal of the American Academy of Child and Adolescent Psychiatry 44:790-797, 2005) by using multiple as opposed to single point in time assessments of early aggressive and disruptive behavior in the classification of youth who would likely benefit from targeted preventive interventions. Different from Petras et al. (2004a, 2005), the outcome used in this study included serious antisocial behavior in young adulthood as well as in adolescence. Among males, the use of multiple time points did not yield greater classification accuracy than the highest single time points, that is, third and fifth grades. For females, although fifth grade represented the best single time point in terms of classification accuracy, no significant association was found between earlier time points and the later outcome, rendering a test of the multiple time points hypothesis moot. The findings presented in this study have strong implications for the design of targeted intervention for violence prevention, indicating that the screening quality based on aggression ratings during the elementary years is rather modest, particularly for females.

Background:Identification of AML patients at high risk of relapse after achieving a complete remission (CR) with induction chemotherapy enables the use of personalized post‐remission treatment strategies to prevent relapse. Detection of molecular measurable residual disease (mMRD) by conventional next‐generation sequencing (NGS) at a single time point post‐induction has previously been associated with a higher incidence of relapse and inferior overall survival (OS).Aims:In this study, we used error‐corrected NGS (EC‐NGS) to evaluate whether changes in mMRD levels between two time points during remission provide additional prognostic information over single time point assessments.Methods:88 AML patients who received standard induction chemotherapy and achieved a CR were evaluated. Targeted NGS of 54 genes associated with myeloid malignancies was performed at diagnosis. We collected peripheral blood (PB) samples upon count recovery following induction chemotherapy and each cycle of consolidation chemotherapy. To detect mMRD, we used a custom 37‐gene hybrid‐capture panel and EC‐NGS based on the Duplex Sequencing approach. PB samples collected at two different time points during remission were analyzed for each patient. The Cox proportional hazards model was used to relate predictor variables to time to relapse and OS. P‐values <0.05 were considered significant.Results:Sequencing analysis of diagnostic samples identified at least one putative oncogenic mutation in 82 of the 88 patients (93%). EC‐NGS of samples collected at the first remission time point (T1) identified at least one persistent mutation in 63 of the 82 patients (77%); 40% of the persistent mutations were in DNMT3A, TET2, and ASXL1 (DTA). The persistence of DTA or non‐DTA mutations, when considered separately, did not correlate with risk of relapse or OS. The co‐persistence of at least one DTA and one non‐DTA mutation was associated with a higher risk of relapse (HR: 2.61; 95% CI, 1.16 to 5.83; P = 0.02) but not OS (P = 0.56). To evaluate whether changes in mMRD level correlate with clinical outcomes, we analyzed the subset of mutations that were detected at diagnosis and persisted in both T1 and T2. We calculated the fold change (FC) in variant allele frequency (VAF) between T1 and T2 for each mutation and used the maximum FC among all mutations of each patient (maxFC) as a continuous variable for regression analysis. High maxFC values restricted to mutations found at diagnosis (maxFCD) were significantly correlated with a higher risk of relapse (P = 0.002) and inferior OS (P = 0.003). We extended the analysis to all non‐synonymous mutations that were detected in T1 and T2 including ones that were not identified at diagnosis. The mutations were predominantly in DNMT3A, TET2, ASXL1, and TP53. Higher maxFC values among mutations found in remission samples (maxFCR) was strongly associated with an increased risk of relapse (P < 0.0001; Fig 1A) and inferior OS (P < 0.0001; Fig 1B). In multivariable regression analysis, maxFCR remained an independent risk factor for relapse after adjustment for age, WBC, ELN 2017 risk group, maxFCD, MRD by flow cytometry, and persistence of mutations found at diagnosis (Table 1).Summary/Conclusion:Our analysis demonstrates that assessment of mMRD kinetics using EC‐NGS provides additional prognostic information over mMRD monitoring at a single time point. We developed an analytical framework to analyze mMRD results in remission and identified a novel parameter (maxFCR) that is independently associated with risk of relapse and OS.image

This paper proposes an empirical model of network formation, combining strategic and random networks features. Payoffs depend on direct links, but also link externalities. Players meet sequentially at random, myopically updating their links. Under mild assumptions, the network formation process is a potential game and converges to an exponential random graph model (ERGM), generating directed dense networks. I provide new identification results for ERGMs in large networks: if link externalities are nonnegative, the ERGM is asymptotically indistinguishable from an Erdős–Renyi model with independent links. We can identify the parameters only when at least one of the externalities is negative and sufficiently large. However, the standard estimation methods for ERGMs can have exponentially slow convergence, even when the model has asymptotically independent links. I thus estimate parameters using a Bayesian MCMC method. When the parameters are identifiable, I show evidence that the estimation algorithm converges in almost quadratic time.

The aim of this study was to compare the accuracy of two abbreviated approaches for estimating Gross Motor Function Measure 66 (GMFM-66) scores against the full GMFM-66 and to explore their strengths and limitations. An existing dataset (n=224) comprising children aged 1 to 13years (mean age 6y 11mo, SD 4y 6mo; 132 males, 92 females) with cerebral palsy (CP) of all Gross Motor Function Classification System (GMFCS) levels was used to compare the validity of the item set version (GMFM-66-IS) and the basal and ceiling version (GMFM-66-B&C) with the full GMFM-66 scores. Follow-up assessment at 1year (n=109) allowed evaluation of change scores and accuracy at a single point in time. The cross-sectional agreement was excellent for both abbreviated measures (all intraclass correlation coefficients [ICCs] >0.98). When measuring change over time, both the GMFM-66-IS and the GMFM-66-B&C showed good agreement for children with bilateral CP (ICCs >0.9). However, the GMFM-66-IS assessed change over 1year more accurately than the GMFM-66-B&C in children with unilateral CP (ICC=0.89 vs ICC=0.58; 95% confidence intervals do not overlap). Both approaches for estimating GMFM-66 scores are accurate at a single point in time. If the primary goal of assessment is to measure change, the full GMFM-66 should still be regarded as the criterion standard. The GMFM-66-IS should be the preferred shortened measure for children with unilateral CP.

Chromothripsis under the microscope: a cytogenetic perspective of two cases of AML with catastrophic chromosome rearrangement

Using cross-sectional data from the Framingham Heart Study including nondemented middle-aged adults, Echouffo-Tcheugui et al. identified an association between higher serum cortisol levels and lower brain volume and poorer performance on a variety of cognitive assessments. Although the serum cortisol measurement was acquired at a single point in time, and may not be reflective of chronic elevations in cortisol as Barbosa et al. assert, the data are strongly suggestive of a correlation. The investigators also agree that hair cortisol might be a better indicator of long-term cortisol levels; however, these data were not available in this observational cohort. Drs. Jimenez-Pavon and Carbonell-Baeza comment on the role that physical activity (PA) and cardiorespiratory fitness (CRF) may have on structural brain integrity and cognitive performance. Some of these associations from the Framingham Heart Study have been previously published in Neurology ® in 2016. Although analysis of these physiologic variables may be worthy of further investigation, Echouffo-Tcheugi et al. argue that PA and CRF may alter cortisol levels (and thereby, neural structure and function), but that such a mediation could lead to an overadjustment. Based on the results of this analysis, the investigators emphasize that cortisol levels may be the bridge the between PA/CRF and brain volume and cognitive performance. Using cross-sectional data from the Framingham Heart Study including nondemented middle-aged adults, Echouffo-Tcheugui et al. identified an association between higher serum cortisol levels and lower brain volume and poorer performance on a variety of cognitive assessments. Although the serum cortisol measurement was acquired at a single point in time, and may not be reflective of chronic elevations in cortisol as Barbosa et al. assert, the data are strongly suggestive of a correlation. The investigators also agree that hair cortisol might be a better indicator of long-term cortisol levels; however, these data were not available in this observational cohort. Drs. Jimenez-Pavon and Carbonell-Baeza comment on the role that physical activity (PA) and cardiorespiratory fitness (CRF) may have on structural brain integrity and cognitive performance. Some of these associations from the Framingham Heart Study have been previously published in Neurology ® in 2016. Although analysis of these physiologic variables may be worthy of further investigation, Echouffo-Tcheugi et al. argue that PA and CRF may alter cortisol levels (and thereby, neural structure and function), but that such a mediation could lead to an overadjustment. Based on the results of this analysis, the investigators emphasize that cortisol levels may be the bridge the between PA/CRF and brain volume and cognitive performance.

Identification and inference of network formation games with misclassified links

To address the inadequacy of traditional Data Envelopment Analysis (DEA) as it can only handle efficiency evaluation of a single point in time, this paper proposes a new efficiency evaluation model of knowledge innovation of Innovative Research Teams in University (IRTU) with DEA improved by taking into consideration the fact that IRTU knowledge innovation occurs over multiple time periods and is characterized by time sequentiality. The improved model can not only handle knowledge innovation at a single point in time, but also that at several points in time. An empirical research validates the improved model.KeywordsIRTUKnowledge innovationDEA modelImpetus system

177 Lu-PSMA-617-radioligand therapy (RLT) has shown promising therapeutic role in patients with metastatic castration-resistant prostate cancer. However, off-target action in salivary glands often presents with xerostomia. Personalized dosimetry can help in optimizing the treatment, however, has so far been tedious due to multiple time-point imaging. In this prospective study, we intended to estimate the absorbed dose delivered to the salivary glands in patients undergoing 177 Lu-PSMA-617-RLT using quantitative SPECT/CT at a single time point. Patients undergoing 177 Lu-PSMA-617 RLT were included in this prospective study. Post-therapy whole-body images and regional quantitative single time-point SPECT/CT were acquired at 24 h with high-energy collimator. The data was processed and analyzed using Q.Metrix software. A scaling factor, that is, the time-integrated activity conversion factor was applied for the image acquired at 24 h. Absorbed doses were computed using MIRD scheme and OLINDA software. A total of 21 patients (mean age: 66 ± 9 years) were included. The value of mean absorbed dose for the parotid glands was 1.90 ± 1.31Gy (range: 0.26-6.23) and that for the submandibular glands was 1.37 ± 0.94Gy (range: 0.16-3.65). The mean absorbed doses per administered activity for the parotid and submandibular glands were 0.26 ± 0.18 Gy/GBq and 0.19 ± 0.12 Gy/GBq, respectively. The absorbed doses were estimated for one cycle of therapy and were well within acceptable limits. None of the patients experienced dryness of mouth. Single time-point dosimetry with quantitative SPECT/CT is feasible and can be standardized to estimate the absorbed dose to salivary glands instead of multiple time-point acquisitions.

Life-course socioeconomic positions and subclinical atherosclerosis in the multi-ethnic study of atherosclerosis

The purpose of this study was to define and document the interior design profession's body of knowledge at a single point in time. This was done using a career cycle approach and a health, safety, and welfare framework. The method and framework used to define the body of knowledge are presented in the article. The body of knowledge was defined from a career cycle approach using the four stages of a professional interior designer's career cycle: education, experience, examination, and legal regulation (NCIDQa, 2003). A content analysis was conducted of the written documents of the organizations that represent each stage in the cycle. Eighty–one knowledge areas were identified from this content analysis and placed into one of seven categories: Codes; Communication; Design; Furnishings, Fixtures, and Equipment; Human Needs; Interior Building Construction; and Professional Practice. These categories and knowledge areas are what defined the interior design profession's body of knowledge based on this approach. Next, each knowledge area was analyzed using a health, safety, and welfare framework to determine its benefit to the public. Finally, a review of literature was conducted to document that the knowledge areas comprise the specialized knowledge necessary for the professional interior designer to protect the public's health, safety, and welfare. The method used to define the interior design profession's body of knowledge assessed several limited bodies of knowledge that had been developed for a specific purpose, such as education or examination. While this comprehensive body of knowledge reflects a single point in time, it provides a venue for dialogue from which revision can occur and updating can continue, leading to further development of the profession.

A significant heritability has been documented for articular osteochondrosis. Selection against osteochondrosis has historically been based on macroscopic evaluation, but as computed tomography (CT) now is used to select boars with optimal body composition it can potentially also be used to screen for osteochondrosis. False negative diagnosis will occur if defects have not developed or have resolved prior to screening at a single time point. The aim of the current study was to assess the suitability of the use of CT at a single point in time as a screening tool in piglets for articular osteochondrosis, which is known to be a highly dynamic condition in which lesions develop and resolve over time. Male Landrace piglets (n = 18) were serial CT scanned from 2-8 times at biweekly intervals from 70-180 days of age. At each interval, 1-2 piglets were euthanased and the left distal femur processed for histological validation. A total of 795 defects were identified in the 112 available CT scans. Within the hind and fore limbs, the incidence of defects was highest in the stifle (n = 321) and elbow joints (n = 110), respectively. Ninety-eight per cent of the defects in the stifle and elbow joints had developed by the 7th examination interval when the piglets were a mean age of 159 days old. The proportion of defects that resolved was lowest in the stifle joint at 51% and highest in the elbow joint at 69%. Scanning of the current piglets at an age of 159 days resulted in detection of 98% of the total number of defects that developed up to the maximum age of 180 days. The proportion of defects that resolved ranged from 51-69% for different joints, but may not adversely affect prevalence as this category of false negative diagnosis will result in selection of pigs that are disposed for healing. Optimally timed CT is a powerful screening tool for osteochondrosis.

Participant noncompliance, in which participants do not follow their assigned treatment protocol, has long complicated the interpretation of randomized clinical trials. No gold standard has been identified for detecting noncompliance, but in some trials participants' biomarkers can provide objective information that suggests exposure to non-study treatments. However, existing methods are limited to retrospectively detecting noncompliance at a single time point based on a single biomarker measurement. We propose a novel method that can leverage participants' full biomarker history to detect noncompliance across multiple time points. Conditional on longitudinal biomarker data, our method can estimate the probability of compliance at (1) a single time point of the trial, (2) all time points, and (3) a future time point. Across time points, we model the biomarker as a mixture density with (latent) components corresponding to longitudinal patterns of compliance. To estimate the mixture density, we fit mixed effects models for both compliance and the biomarker. We use the mixture density to derive compliance probabilities that condition on the longitudinal biomarker data. We evaluate our compliance probabilities by simulation and apply them to a trial in which current smokers were asked to only smoke low nicotine study cigarettes (Center for the Evaluation of Nicotine in Cigarettes Project 1 Study 2). In the simulation, we investigated three different effects of compliance on the biomarker, as well as the effect of misspecification of the covariance structures. We compared probability estimators (1) and (2) to those that ignore the longitudinal correlation in the data according to area under the receiver operating characteristic curve. We evaluated estimator (3) by plotting its calibration lines. For Center for the Evaluation of Nicotine in Cigarettes Project 1 Study 2, we compared estimators (1) and (3) to a probability estimator of compliance at the last time point that ignores the longitudinal correlation. In the simulation, for both compliance at the last time point and at all time points, conditioning on the longitudinal biomarker data uniformly raised area under the receiver operating characteristic curve across all three compliance effect scenarios. The gains in area under the receiver operating characteristic curve were smaller under misspecification. The calibration lines for the prediction of compliance closely followed 45°, though with additional variability under misspecification. For compliance at the last time point of Center for the Evaluation of Nicotine in Cigarettes Project 1 Study 2, conditioning on participants' full biomarker history boosted area under the receiver operating characteristic curve by three percentage points. The prediction probabilities somewhat accurately approximated the non-longitudinal compliance probabilities. Compared to existing methods that only use a single biomarker measurement, our method can account for the longitudinal correlation in the biomarker and compliance to more accurately identify noncompliant participants. Our method can also use participants' biomarker history to predict compliance at a future time point.