Truncation Bias Research Articles

Best practices for estimating and reporting epidemiological delay distributions of infectious diseases.

Epidemiological delays are key quantities that inform public health policy and clinical practice. They are used as inputs for mathematical and statistical models, which in turn can guide control strategies. In recent work, we found that censoring, right truncation, and dynamical bias were rarely addressed correctly when estimating delays and that these biases were large enough to have knock-on impacts across a large number of use cases. Here, we formulate a checklist of best practices for estimating and reporting epidemiological delays. We also provide a flowchart to guide practitioners based on their data. Our examples are focused on the incubation period and serial interval due to their importance in outbreak response and modeling, but our recommendations are applicable to other delays. The recommendations, which are based on the literature and our experience estimating epidemiological delay distributions during outbreak responses, can help improve the robustness and utility of reported estimates and provide guidance for the evaluation of estimates for downstream use in transmission models or other analyses.

Associations of Serum Perfluoroalkyl Substances and Placental Human Chorionic Gonadotropin in Early Pregnancy, Measured in the UPSIDE Study in Rochester, New York.

Per- and polyfluoroalkyl substances (PFAS) are widely detected in pregnant women and associated with adverse outcomes related to impaired placental function. Human chorionic gonadotropin (hCG) is a dimeric glycoprotein hormone that can indicate placental toxicity. Our aim was to quantify the association of serum PFAS with placental hCG, measured as an intact molecule (hCG), as free alpha-() and beta-subunits (), and as a hyperglycosylated form (h-hCG), and evaluate effect measure modification by social determinants and by fetal sex. Data were collected from 326 pregnant women enrolled from 2015 to 2019 in the UPSIDE study in Rochester, New York. hCG forms were normalized for gestational age at the time of blood draw in the first trimester [multiple of the median (MoM)]. Seven PFAS were measured in second-trimester maternal serum. Multivariate imputation by chained equations and inverse probability weighting were used to evaluate robustness of linear associations. PFAS mixture effects were estimated by Bayesian kernel machine regression. Perfluorohexane sulfonic acid (PFHxS) [: 0.29 log MoM units per log PFHxS; 95% confidence interval (CI): 0.08, 0.51] and perfluorodecanoic acid (PFDA) (hCG: ; 95% CI: , ) were associated with hCG in the single chemical and mixture analyses. The PFAS mixture was negatively associated with and positively with . Subgroup analyses revealed that PFAS associations with hCG differed by maternal race/ethnicity and education. Perfluoropentanoic acid (PFPeA) was associated with only in Black participants (; 95% CI: , ) and in participants with high school education or less (; 95% CI: , ); conversely, perfluorononanoic acid (PFNA) was negatively associated with only in White participants (; 95% CI: , ) and with only in participants with a college education or greater (; 95% CI: , ). These findings were robust to testing for selection bias, confounding bias, and left truncation bias where PFAS detection frequency was . Two associations were negative in male (and null in female) pregnancies: Perfluoroundecanoic acid (PFUnDA) with , and PFNA with h-hCG. Evidence was strongest for the association between PFHxS and PFDA with hCG in all participants and for PFPeA and PFNA within subgroups defined by social determinants and fetal sex. PFAS mixture associations with and differed, suggesting subunit-specific types of toxicity and/or regulation. Future studies will evaluate the biological, clinical and public health significance of these findings. https://doi.org/10.1289/EHP12950.

Is there a secular decline in disruptive patents? Correcting for measurement bias

Despite tremendous growth in the volume of new scientific and technological knowledge, the popular press has recently raised concerns that disruptive innovation is slowing. These dire prognoses were driven in part by Park et al. (2023), a Nature publication that uses decades of data and millions of observations coupled with a novel quantitative metric (the CD index) that characterizes innovation in science and technology as either consolidating or disruptive. We challenge the (Park et al., 2023) patent findings, principally around concerns of truncation bias and exclusion bias. We show that 88 percent of the decrease in the average CD index over 1980–2010 reported by the authors can be explained by their truncation of all backward patent citations before 1976. We also show that this truncation bias varies by technology class. We further account for a change in U.S. patent law that allows for citations to patent applications in addition to patent grants—something ignored by the authors in their analysis—and update the analysis to 2016. We show that the number of highly disruptive patents has increased since 1980—particularly since 2008. Our results suggest caution in using the (Park et al., 2023) patent findings and conclusions as a basis for research and decision-making in public policy, industry restructuring or firm reorganization aimed at altering the current innovation landscape.

Genomic Characterization and Clinical Outcomes of Patients with Peritoneal Metastases from the AACR GENIE Biopharma Collaborative Colorectal Cancer Registry.

Utilizing the GENIE BPC registry, this study found that PM in patients with colorectal cancer occur more frequently in females and right-sided primary tumors and are associated with worse OS. In addition, we found a lower frequency of APC alterations and a higher frequency in MED12 alterations in patients with PM.

Estimates of congenital cytomegalovirus-attributable infant mortality in high-income countries: A review.

As many as 5%-10% of infants with symptomatic congenital cytomegalovirus (cCMV) disease, or 0.4%-0.8% of all liveborn infants with cCMV infection, die in early infancy in high-income countries. However, estimates are uncertain due to several potential biases that can result from data limitations and study designs. First, infants with cCMV infections who die prior to diagnosis, which usually occurs at 1-4weeks after birth, may be excluded from both the count of deaths and the denominator of cCMV births, resulting in left truncation and immortal time biases. These 'biases' are features of the data and do not reflect bias on the part of researchers, but understanding the potential existence of threats to validity can help with interpretation of findings. Left truncation of infant deaths occurring prior to diagnosis of cCMV can result in understatement of the burden of infant deaths due to cCMV. Conversely, overestimation of infant deaths associated with symptomatic cCMV may occur in clinical case series owing to greater representation of relatively severely affected infants owing to ascertainment and referral biases. In this review, we summarise the characteristics of 26 studies that reported estimates of cCMV-associated infant deaths, including potential biases or limitations to which those estimates may have been subject. We discuss study designs whose implementation might generate improved estimates of infant deaths attributable to cCMV. More complete estimates of the overall public health impact of cCMV could inform current and future screening, prevention, and vaccine research.

Abstract 1177: Genomic characteristics and clinical outcomes of early onset colorectal cancer (EOCRC): Findings from AACR Project GENIE Biopharma Collaborative registry

Abstract Background: By 2030, the incidence of EOCRC is predicted to increase by more than 140%. A better understanding of the molecular features of EOCRC is needed. Methods: The AACR GENIE Biopharma Collaborative CRC 1.2-consortium dataset consists of clinical and genomic data from a cohort of CRC patients (pts) with next generation tumor sequencing (55-500 genes) from 2015 to 2017 at three academic centers. Imaging reports and medical oncology notes were retrospectively curated. EOCRC was defined as age at diagnosis < 50 years; while later onset CRC (LOCRC) as ≥ 50 years. Association with clinical variables was analyzed using Chi-Square or Fisher’s exact test. Associations with genomic alterations were analyzed using the Benjamini-Hochberg procedure for multiple testing; the associations were evaluated among the overall study population and by stage. Kaplan-Meier method was used to estimate the median overall survival (OS) and was adjusted to left truncation bias introduced by the inclusion criteria for this project. Results: A total of 1507 pts with CRC, 519 (34%) with EOCRC, were analyzed. EOCRC pts were more likely to be diagnosed with Stage IV than LOCRC (51% vs 44%, p<0.01). Left-sided primary tumors were more frequent in EOCRC (73% vs 57%, p<0.01). Among stage I-III, EOCRC was associated with longer OS compared to LOCRC (HR:1.43, 95%CI 1.07-1.91, p=0.01). Among stage IV CRC, no differences were observed in OS between EOCRC and LOCRC in a multivariate analysis including RAS/BRAF mutation (mt), MSI and liver, lung and peritoneal metastasis (HR: 1.01, 95% CI 0.8-1.28, p=0.92). The proportion of BRAF mt was higher in LOCRC (15 vs 7%, p<0.01), mainly in stage I-III (17% vs 7%, p=0.01). This association is significant after adjusting for tumor sidedness (OR: 1.77 95% CI 1.19-2.63, p<0.01). BRAF V600E mt is more frequent in LOCRC (11 vs 3%, p<0.01). APC and PI3KCA mt are more common in EOCRC, but only among right-sided tumors. There were not significant differences in copy number alterations or structural variants. APC, TP53, KRAS, BRAF and PI3KCA were analyzed according to stage. Among EOCRC, TP53 mt was more frequent in stage IV vs I-III (79 vs 70%, p=0.02). In LOCRC, both TP53 (76% vs 69%, p=0.01) and KRAS mt (47 vs 38%, p<0.01) were more frequent in stage IV vs I-III. Effect modification of the association between these genes and OS by age of onset was explored using an interaction test in stage IV and I-III. APC mt in stage I-III conferred better prognosis in EOCRC (HR 0.31 (0.18-0.51) for EOCRC; HR 0.72 (0.51-1.01) for LOCRC; interaction p<0.01). Conclusions: EOCRC does not show significant genomic differences from LOCRC except for lower frequency of BRAF mt. For right-sided tumors, APC and PI3KCA mt are more common in EOCRC. APC mt in stage I-III is associated with a better prognosis in EOCRC. Further characterization of the biology of EOCRC beyond genomic profiling is needed. Citation Format: Enrique Sanz Garcia, Eric Chen, Marios Giannakis, Gregory J. Riely, Jeremy L. Warner, Michele L. LeNoue-Newton, Jessica Weiss, Katrina Hueniken, Kenneth L. Kehl, Deborah Schrag, Andrea Cercek, Kimmie Ng, AACR GENIE BPC Core Team . Genomic characteristics and clinical outcomes of early onset colorectal cancer (EOCRC): Findings from AACR Project GENIE Biopharma Collaborative registry [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 1177.

Rigour and reproducibility in perinatal and paediatric epidemiologic research using big data.

Rigour and reproducibility in perinatal and paediatric epidemiologic research using big data.

Estimating the serial intervals of SARS-CoV-2 Omicron BA.4, BA.5, and BA.2.12.1 variants in Hong Kong.

Empirical evidence on the epidemiological characteristics of the emerged SARS-CoV-2 variants could shed light on the transmission potential of the virus and strategic outbreak control planning. In this study, by using contact tracing data collected during an Omicron-predominant epidemic phase in Hong Kong, we estimated the mean serial interval of SARS-CoV-2 Omicron BA.4, BA.5, and BA.2.12.1 variants at 2.8days (95% credible interval [CrI]: 1.5, 6.7), 2.7days (95% CrI: 2.1, 3.6), and 4.4days (95% CrI: 2.6, 7.5), respectively, with adjustment for right truncation and sampling bias. The short serial interval for the current circulating variant indicated that outbreak mitigations through contact tracing and case isolation would be quite challenging.

The challenges of time for studies on the population effects of the COVID-19 pandemic on perinatal outcomes.

The challenges of time for studies on the population effects of the COVID-19 pandemic on perinatal outcomes.

Evaluating methodological approaches to assess the severity of infection with SARS-CoV-2 variants: scoping review and applications on Belgian COVID-19 data

BackgroundDifferences in the genetic material of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants may result in altered virulence characteristics. Assessing the disease severity caused by newly emerging variants is essential to estimate their impact on public health. However, causally inferring the intrinsic severity of infection with variants using observational data is a challenging process on which guidance is still limited. We describe potential limitations and biases that researchers are confronted with and evaluate different methodological approaches to study the severity of infection with SARS-CoV-2 variants.MethodsWe reviewed the literature to identify limitations and potential biases in methods used to study the severity of infection with a particular variant. The impact of different methodological choices is illustrated by using real-world data of Belgian hospitalized COVID-19 patients.ResultsWe observed different ways of defining coronavirus disease 2019 (COVID-19) disease severity (e.g., admission to the hospital or intensive care unit versus the occurrence of severe complications or death) and exposure to a variant (e.g., linkage of the sequencing or genotyping result with the patient data through a unique identifier versus categorization of patients based on time periods). Different potential selection biases (e.g., overcontrol bias, endogenous selection bias, sample truncation bias) and factors fluctuating over time (e.g., medical expertise and therapeutic strategies, vaccination coverage and natural immunity, pressure on the healthcare system, affected population groups) according to the successive waves of COVID-19, dominated by different variants, were identified. Using data of Belgian hospitalized COVID-19 patients, we were able to document (i) the robustness of the analyses when using different variant exposure ascertainment methods, (ii) indications of the presence of selection bias and (iii) how important confounding variables are fluctuating over time.ConclusionsWhen estimating the unbiased marginal effect of SARS-CoV-2 variants on the severity of infection, different strategies can be used and different assumptions can be made, potentially leading to different conclusions. We propose four best practices to identify and reduce potential bias introduced by the study design, the data analysis approach, and the features of the underlying surveillance strategies and data infrastructure.

Estimation of the serial interval of monkeypox during the early outbreak in 2022.

With increased transmissibility and novel transmission mode, monkeypox poses new threats to public health globally in the background of the ongoing COVID-19 pandemic. Estimates of the serial interval, a key epidemiological parameter of infectious disease transmission, could provide insights into the virus transmission risks. As of October 2022, little was known about the serial interval of monkeypox due to the lack of contact tracing data. In this study, public-available contact tracing data of global monkeypox cases were collected and 21 infector-infectee transmission pairs were identified. We proposed a statistical method applied to real-world observations to estimate the serial interval of the monkeypox. We estimated a mean serial interval of 5.6 days with the right truncation and sampling bias adjustedand calculated the reproduction number of 1.33 for the early monkeypox outbreaks at a global scale. Our findingsprovided a preliminary understanding of the transmission potentials of the current situation of monkeypox outbreaks. We highlighted the needfor continuous surveillance of monkeypox for transmission risk assessment.

Transmission dynamics of monkeypox in the United Kingdom: contact tracing study

ObjectiveTo analyse the transmission dynamics of the monkeypox outbreak in the UK, declared a Public Health Emergency of International Concern in July 2022.DesignContact tracing study, linking data on case-contact pairs...

Is the lack of prior exposure to sperm antigens associated with worse neonatal and maternal outcomes? A 10-year single-center experience comparing ICSI-TESE pregnancies to ICSI pregnancies.

Nowadays, pathogenesis of preeclampsia (PE) is still unknown. Among the different etiological hypotheses, some authors proposed that it might be because of an abnormal immunologic response to a foreign fetal antigen derived from the father's spermatozoa. Indeed, the fetus is considered a semi-allograft, being one half paternally derived in its antigenicity, and the first pathogenic insult of PE may be an abnormal maternal immune response toward this semi-allogenic implant. In the context of artificial reproductive techniques, it has been shown that the use of donor and surgically retrieved spermatozoa (e.g., testicular sperm extraction [TESE]) increases the risk of PE, confirming the protective effect of sperm exposure on maternal complications. Determining whether the lack of exposure to sperm antigens is associated with worse maternal and neonatal outcomes in pregnancies obtained through intracytoplasmic sperm injection after TESE (ICSI-TESE) for obstructive azoospermia (OA). This is a single-center case-control retrospective study, focusing on all first pregnancies obtained through ICSI-TESE for OA at Humanitas Fertility Center between January 1, 2010 and December 31, 2019. Controls included patients that achieved their first pregnancy with ICSI and ejaculated spermatozoa, for a diagnosis other than azoospermia, in the same time period. Cases were matched with controls in a 1:2 ratio, considering female age, female BMI, and year of controlled ovarian stimulation. The primary outcome measure was the delivery rate, defined as the number of deliveries divided by the total number of clinical pregnancies. Secondary outcome measures focused on maternal and neonatal complications, such as miscarriage rate, rate of main obstetric complications, prematurity rate, and rate of congenital malformations. By analyzing overall 113 pregnancies among cases and 214 pregnancies among controls, this study showed that the delivery rate was higher in controls with respect to cases (92.06% vs. 84.07%, p=0.026); among deliveries, live births were 98.95% and 100%, respectively, whereas only one stillbirth occurred in cases. The first trimester miscarriage rate was higher in the cases than controls (13.27% vs. 6.07%, p=0.027), whereas no difference was found among the rate of second trimester miscarriages, therapeutic abortions, and ectopic pregnancies. There was no difference regarding the rate of maternal complications, including gestational hypertension, PE, HELLP syndrome, gestational diabetes, placenta previa, placental abruption, and premature rupture of the membranes. Considering neonatal complications, it was shown that twins belonging to controls had a higher prematurity rate with respect to cases (65.79% vs. 50.00%) but without a statistical relevance. Lastly, the rate of congenital malformations did not differ among the two groups. This study showed that, once couples diagnosed with OA achieve a pregnancy, they have a much higher risk of miscarriage in the first trimester in respect to non-azoospermic patients. Moreover, controls had a higher delivery rate in respect to cases; however, when the fetal status at birth was compared, no difference was found between live births and stillbirths. Differently from the findings in the literature, no association with PE was found. This might be related to a collider bias/left truncation bias: As azoospermic patients are at higher risk of early termination of pregnancy, it results that they do not have the possibility to develop PE and other adverse outcomes.

Estimating survival parameters under conditionally independent left truncation.

Databases derived from electronic health records (EHRs) are commonly subject to left truncation, a type of selection bias that occurs when patients need to survive long enough to satisfy certain entry criteria. Standard methods to adjust for left truncation bias rely on an assumption of marginal independence between entry and survival times, which may not always be satisfied in practice. In this work, we examine how a weaker assumption of conditional independence can result in unbiased estimation of common statistical parameters. In particular, we show the estimability of conditional parameters in a truncated dataset, and of marginal parameters that leverage reference data containing non‐truncated data on confounders. The latter is complementary to observational causal inference methodology applied to real‐world external comparators, which is a common use case for real‐world databases. We implement our proposed methods in simulation studies, demonstrating unbiased estimation and valid statistical inference. We also illustrate estimation of a survival distribution under conditionally independent left truncation in a real‐world clinico‐genomic database.

Under the same (Chole)sky: DNK models, timing restrictions and recursive identification of monetary policy shocks

Recent structural VAR studies of the monetary transmission mechanism have voiced concerns about the use of recursive identification schemes based on short-run exclusion restrictions. This paper characterizes the effects on impulse propagation of informational constraints embodying classical Cholesky-type timing restrictions in otherwise standard DSGE models. We formally show that timing restrictions can produce non-trivial moving average components of rational expectations solutions, or even serve as an independent source of model-based nonfundamentalness, thereby hampering impulse response analysis via VAR procedures. We then derive population conditions for existence of VAR representations of DSGE economies exhibiting timing restrictions, and numerically explore their bearing on shock identification in a range of monetary models of the business cycle. Our analysis reveals that dynamic New Keynesian models admit invertible equilibrium representations as well as fast-converging VAR coefficient matrices under empirically tenable parameterizations. This alleviates concerns about identification and lag truncation bias: low-order Cholesky-VARs do well at retrieving the true aggregate effects of monetary policy shocks in a Cholesky world.

Patent Valuation Using Citations: A Review and Sensitivity Analysis

Abstract As a result of studies demonstrating a correlation between a patent’s value and its forward citation count, patent valuation using forward citations has been increasingly used by practitioners when a patent’s value has not been otherwise established. Although potential limitations of patent citation analysis have been discussed in the past, there is little empirical research demonstrating the sensitivity of estimated patent values to various assumptions embedded within the method. We first summarize an approach that has been used by prior practitioners to estimate the relative value of patents within a portfolio using forward citations, and then perform various analyses to investigate the sensitivity of the approach to certain assumptions. We find that some concerns of prior literature are well-founded, while others are less so. For example, we confirm that biased valuations will result from failure to properly control for patent age and technology. Our analysis also finds that truncation bias is a problem when analyzing recently issued patents, which confirms findings from existing literature. We estimate the rate at which such truncation bias dissipates as patents age and find that the bias for the median patent is reduced to below 10% within five years from the date of publication, although additional variation can remain on an individualized level. Regarding the issue of self-citations, we find that the valuation approach using forward citation analysis can be (but is not always) sensitive to the issue of self-citations, with a median difference of 16.8%. Finally, the valuation approach using forward citation analysis appears to be robust to assumptions underlying patent cohort construction.

Under the same (Chole)sky: DNK models, timing restrictions and recursive identification of monetary policy shocks

Recent structural VAR studies of the monetary transmission mechanism have voiced concerns about the use of recursive identification schemes based on short-run exclusion restrictions. We trace out the effects on impulse propagation of informational constraints embodying classical Cholesky-timing restrictions in otherwise standard Dynamic New Keynesian (DNK) models. By reinforcing internal propagation mechanisms and enlarging a model's equilibrium state space, timing restrictions may produce a non-trivial moving average component of the equilibrium representation, making finite order VARs a poor approximation of true adjustment paths to monetary impulses, albeit correctly identified. They can even serve as an independent source of model-based nonfundamentalness, thereby hampering shock identification via VAR methods. This notwithstanding, restricted DNK models are shown to feature (i) invertible equilibrium representations for the observables and (ii) fast-converging VAR coefficient matrices under empirically tenable parameterizations. This alleviates concerns about identification and lag truncation bias: low-order Cholesky-VARs do well at uncovering the transmission of monetary impulses in a truly Cholesky world.

Public‐to‐Private Buyouts and Innovation

AbstractWe study the effect of public‐to‐private buyout transactions on investments in innovation using an international sample over the 1997–2017 period. We use patent counts and citations to proxy for the quantity, quality and economic importance of innovation. Our results are based on time analysis and matched sample regressions. The data indicate that buyouts are associated with a significant reduction in patents and patent citations, including a reduction in radical (i.e. more scientific) patents. When we split the sample into institutional and management buyouts, the negative effect of buyouts is confirmed only for institutional buyouts. This suggests that only institutional buyouts prevent target firms from adopting long‐term investments. This finding is confirmed by reductions in innovator employment and innovation efficiency subsequent to going private. Moreover, the data indicate that the negative effect is most prevalent for transactions where the cost of the deal's debt financing is higher than that of the debt post‐buyout. We rule out some alternative explanations for these findings, including but not limited to outliers, truncation bias and endogeneity.

Occupational Dualism and Intergenerational Educational Mobility in the Rural Economy: Evidence from China and India

This paper extends the Becker-Tomes model of intergenerational educational mobility to a rural economy characterized by farm-nonfarm occupational dualism and provides a comparative analysis of rural China and rural India. The model builds a micro-foundation for the widely used linear-in-levels estimating equation. Returns to education for parents and productivity of financial investment in children’s education determine relative mobility, as measured by the slope, while the intercept depends, among other factors, on the degree of persistence in nonfarm occupations. Unlike many existing studies based on coresident samples, our estimates of intergenerational mobility do not suffer from truncation bias. The sons in rural India faced lower educational mobility compared with the sons in rural China in the 1970s to 1990s. To understand the role of genetic inheritance, Altonji et al. (2005) biprobit sensitivity analysis is combined with the evidence on intergenerational correlation in cognitive ability in economics and behavioral genetics literature. The observed persistence can be due solely to genetic correlations in China, but not in India. Father’s nonfarm occupation was complementary to his education in determining a sons’ schooling in India, but separable in China. There is evidence of emerging complementarity for the younger cohorts in rural China. Structural change in favor of the nonfarm sector contributed to educational inequality in rural India. Evidence from supplementary data on economic mechanisms suggests that the model provides plausible explanations for the contrasting roles of occupational dualism in intergenerational educational mobility in rural India and rural China. (Stone Center on Socio-Economic Inequality Working Paper Series)

Do incentives work? Option-based compensation and corporate innovation

Numerous scholars have investigated the link between option-based compensation and innovation in public firms, generally concluding options motivate innovative activities better than alternative compensation methods. In most studies, the evidence of causality is weak or suffers from econometric concerns, leading to questions regarding the conclusions. We use the implementation of FAS 123R as a quasi-natural experiment to explore the role option compensation has in incentivizing innovation. FAS 123R increased the cost of granting options as compensation, resulting in a plausibly exogenous shift away from options. Using an updated sample of patent applications, we do not observe any decline in innovation. Despite a similar empirical design, our results differ from those of Mao and Zhang (2018), due to truncation bias inherent in their patent data. Overall, using our updated data, our findings are inconsistent with conclusions from prior studies and suggest option-based compensation does not drive innovation.