Iconicity mediates semantic networks of sound symbolisma).

One speech sound can be associated with multiple meanings through iconicity, indexicality, and/or systematicity. It was not until recently that this "pluripotentiality" of sound symbolism attracted serious attention, and it remains uninvestigated how pluripotentiality may arise. In the current study, Japanese, Korean, Mandarin, and English speakers rated unfamiliar jewel names on three semantic scales: size, brightness, and hardness. The results showed language-specific and cross-linguistically shared pluripotential sound symbolism. Japanese speakers associated voiced stops with large and dark jewels, whereas Mandarin speakers associated [i] with small and bright jewels. Japanese, Mandarin, and English speakers also associated lip rounding with darkness and softness. These sound-symbolic meanings are unlikely to be obtained through metaphorical or metonymical extension, nor are they reported to colexify. Notably, in a purely semantic network without the mediation of lip rounding, softness can instead be associated with brightness, as illustrated by synesthetic metaphors such as yawaraka-na hizashi /jawaɾakanaçizaɕi/ "a gentle (lit. soft) sunshine" in Japanese. These findings suggest that the semantic networks of sound symbolism may not coincide with those of metaphor or metonymy. The current study summarizes the findings in the form of (phono)semantic maps to facilitate cross-linguistic comparisons of pluripotential sound symbolism.

Visualization and exploration of linked data using virtual reality.

In this report, we analyse the use of virtual reality (VR) as a method to navigate and explore complex knowledge graphs. Over the past few decades, linked data technologies [Resource Description Framework (RDF) and Web Ontology Language (OWL)] have shown to be valuable to encode such graphs and many tools have emerged to interactively visualize RDF. However, as knowledge graphs get larger, most of these tools struggle with the limitations of 2D screens or 3D projections. Therefore, in this paper, we evaluate the use of VR to visually explore SPARQL Protocol and RDF Query Language (SPARQL) (construct) queries, including a series of tutorial videos that demonstrate the power of VR (see Graph2VR tutorial playlist: https://www.youtube.com/playlist?list=PLRQCsKSUyhNIdUzBNRTmE-_JmuiOEZbdH). We first review existing methods for Linked Data visualization and then report the creation of a prototype, Graph2VR. Finally, we report a first evaluation of the use of VR for exploring linked data graphs. Our results show that most participants enjoyed testing Graph2VR and found it to be a useful tool for graph exploration and data discovery. The usability study also provides valuable insights for potential future improvements to Linked Data visualization in VR.

Sniffing out meaning: Chemosensory and semantic neural network changes in sommeliers.

Wine tasting is a very complex process that integrates a combination of sensation, language, and memory. Taste and smell provide perceptual information that, together with the semantic narrative that converts flavor into words, seem to be processed differently between sommeliers and naïve wine consumers. We investigate whether sommeliers' wine experience shapes only chemosensory processing, as has been previously demonstrated, or if it also modulates the way in which the taste and olfactory circuits interact with the semantic network. Combining diffusion-weighted images and fMRI (activation and connectivity) we investigated whether brain response to tasting wine differs between sommeliers and nonexperts (1) in the sensory neural circuits representing flavor and/or (2) in the neural circuits for language and memory. We demonstrate that training in wine tasting shapes the microstructure of the left and right superior longitudinal fasciculus. Using mediation analysis, we showed that the experience modulates the relationship between fractional anisotropy and behavior: the higher the fractional anisotropy the higher the capacity to recognize wine complexity. In addition, we found functional differences between sommeliers and naïve consumers affecting the flavor sensory circuit, but also regions involved in semantic operations. The former reflects a capacity for differential sensory processing, while the latter reflects sommeliers' ability to attend to relevant sensory inputs and translate them into complex verbal descriptions. The enhanced synchronization between these apparently independent circuits suggests that sommeliers integrated these descriptions with previous semantic knowledge to optimize their capacity to distinguish between subtle differences in the qualitative character of the wine.

Analysis of the impact of COVID-19 on Scotland's care-homes from March 2020 to October 2021: national linked data cohort analysis.

BACKGROUND: The impact of the COVID-19 pandemic on long-term care residents remains of wide interest, but most analyses focus on the initial wave of infections. OBJECTIVE: To examine change over time in: (i) The size, duration, classification and pattern of care-home outbreaks of COVID-19 and associated mortality and (ii) characteristics associated with an outbreak. DESIGN: Retrospective observational cohort study using routinely-collected data. SETTING: All adult care-homes in Scotland (1,092 homes, 41,299 places). METHODS: Analysis was undertaken at care-home level, over three periods. Period (P)1 01/03/2020-31/08/2020; P2 01/09/2020-31/05/2021 and P3 01/06/2021-31/10/2021. Outcomes were the presence and characteristics of outbreaks and mortality within the care-home. Cluster analysis was used to compare the pattern of outbreaks. Logistic regression examined care-home characteristics associated with outbreaks. RESULTS: In total 296 (27.1%) care-homes had one outbreak, 220 (20.1%) had two, 91 (8.3%) had three, and 68 (6.2%) had four or more. There were 1,313 outbreaks involving residents: 431 outbreaks in P1, 559 in P2 and 323 in P3. The COVID-19 mortality rate per 1,000 beds fell from 45.8 in P1, to 29.3 in P2, and 3.5 in P3. Larger care-homes were much more likely to have an outbreak, but associations between size and outbreaks were weaker in later periods. CONCLUSIONS: COVID-19 mitigation measures appear to have been beneficial, although the impact on residents remained severe until early 2021. Care-home residents, staff, relatives and providers are critical groups for consideration and involvement in future pandemic planning.

Impact of low birth weight on academic attainment during adolescence: A comprehensive retrospective cohort study using linked data.

BACKGROUND: This study addresses a critical knowledge gap by exploring the intricate relationship between low birth weight (LBW) and the heightened risk of suboptimal academic achievement during adolescence through a comprehensive retrospective cohort design. METHODS: In this registry-based cohort study, meticulously linked health and curriculum-based test data for individuals born in New South Wales (NSW), Australia, between 2003 and 2005 were employed. Birth weight data were carefully sourced from the NSW perinatal data collection (PDC). The educational performance of offspring was thoroughly evaluated using the National Assessment Program for Literacy and Numeracy (NAPLAN) during grade 9, approximately at 14 years of age. RESULTS: After rigorous adjustments for potential confounders, findings revealed a compelling narrative: LBW adolescents demonstrated an elevated susceptibility to not meeting national minimum standards across all domains, encompassing spelling [OR, 1.59 (95%CI 1.48-1.69)], writing [OR, 1.51 (95%CI 1.41-1.61)], reading [OR, 1.38 (95%CI 1.29-1.48)], and numeracy [OR, 1.52 (95%CI 1.40-1.63)]. Notably, LBW boys exhibited a more pronounced inclination towards diminished academic performance compared to their female counterparts. CONCLUSIONS: This comprehensive retrospective cohort study, based on linked data, unequivocally establishes LBW as significantly associated with an increased vulnerability to substandard educational achievement during adolescence. Particularly robust effects were observed in females across all outcomes. Aimed at investigating whether LBW serves as a predictive factor for later academic difficulties, this study underscores the imperative for the adoption and fortification of preventative and early intervention strategies to curtail the prevalence of LBW-associated academic underachievement in later adolescence.

Does use of GP and specialist services vary across areas and according to individual socioeconomic position? A multilevel analysis using linked data in Australia.

OBJECTIVE: Timely access to primary care and supporting specialist care relative to need is essential for health equity. However, use of services can vary according to an individual's socioeconomic circumstances or where they live. This study aimed to quantify individual socioeconomic variation in general practitioner (GP) and specialist use in New South Wales (NSW), accounting for area-level variation in use. DESIGN: Outcomes were GP use and quality-of-care and specialist use. Multilevel logistic regression was used to estimate: (1) median ORs (MORs) to quantify small area variation in outcomes, which gives the median increased risk of moving to an area of higher risk of an outcome, and (2) ORs to quantify associations between outcomes and individual education level, our main exposure variable. Analyses were adjusted for individual sociodemographic and health characteristics and performed separately by remoteness categories. SETTING: Baseline data (2006-2009) from the 45 and Up Study, NSW, Australia, linked to Medicare Benefits Schedule and death data (to December 2012). PARTICIPANTS: 267 153 adults aged 45 years and older. RESULTS: GP (MOR=1.32-1.35) and specialist use (1.16-1.18) varied between areas, accounting for individual characteristics. For a given level of need and accounting for area variation, low education-level individuals were more likely to be frequent users of GP services (no school certificate vs university, OR=1.63-1.91, depending on remoteness category) and have continuity of care (OR=1.14-1.24), but were less likely to see a specialist (OR=0.85-0.95). CONCLUSION: GP and specialist use varied across small areas in NSW, independent of individual characteristics. Use of GP care was equitable, but specialist care was not. Failure to address inequitable specialist use may undermine equity gains within the primary care system. Policies should also focus on local variation.

The organization of the semantic network as reflected by the neural correlates of six semantic dimensions.

Multiple sensory-motor and non-sensory-motor dimensions have been proposed for semantic representation, but it remains unclear how the semantic system is organized along them in the human brain. Using naturalistic fMRI data and large-scale semantic ratings, we investigated the overlaps and dissociations between the neural correlates of six semantic dimensions: vision, motor, socialness, emotion, space, and time. Our findings revealed a more complex semantic atlas than what is predicted by current neurobiological models of semantic representation. Brain regions that are selectively sensitive to specific semantic dimensions were found both within and outside the brain networks assumed to represent multimodal general and/or abstract semantics. Overlaps between the neural correlates of different semantic dimensions were mainly found inside the default mode network, concentrated in the left anterior superior temporal gyrus and angular gyrus, which have been proposed as two connector hubs that bridge the multimodal experiential semantic system and the language-supported semantic system.

Voxel representation of brain images inpainting via Regional Pixel Semantic Network and pyramidal attention AE - Quantile differential mechanism model.

Medical image inpainting holds significant importance in enhancing the quality of medical images by restoring missing areas, thereby rendering them suitable for diagnostic purposes. While several techniques have been previously proposed for medical image inpainting, they are not suitable for distorted images containing metallic implants due to their limited consideration of known shaped masking. To overcome this limitation, a novel Vectorized Box Interpolation with Arbitrary Auto-Rand Augment Masking technique has been proposed which involves scaling and vectorizing images to expand their details and generating asymmetrically shaped masking in an automatic random format. One of the challenging tasks in this regard is the precise detection of lost regions, which is addressed through the introduction of the Regional Pixel Semantic Network. This technique employs the locally shared features (LSF) based region sensing with FCN (fully convolutional network) segmentation, which performs automatic segmentation based on neighboring pixel local dependency and regional features to determine the location of masked regions. During the reconstruction of missing parts, a significant challenge posed is the inability to recognize proximity in encoding owing to the generation of shadow-like regions on the feature map. To address this issue, a novel Multilayered DRC Regularized Pyramidal Attention AE Model has been proposed which employs dilated convolution with coherent pyramidal attention for feature extraction and improves image resolution using a Laplacian convolutional layer. Moreover, the realness of the generated image is determined using the Quantile Differential Mechanism model, where in the Quantile Differential Partial Convolutional Discriminator utilizes the hyperbolic tangent activation function in the partial convolutional layer to calculate recognition accuracy. As a result, the proposed method achieves high percentages for accuracy (98 %), precision (97 %), sensitivity (96 %), recall (95 %), and F-measure (96 %) thereby outperforming existing methods. Overall, this proposed method effectively handles distorted images with metallic implants, accurately detects lost regions, and improves the reconstructed image quality.

Development and quality appraisal of a new English breast screening linked data set as part of the age, test threshold, and frequency of mammography screening (ATHENA-M) study.

OBJECTIVES: To build a data set capturing the whole breast cancer screening journey from individual breast cancer screening records to outcomes and assess data quality. METHODS: Routine screening records (invitation, attendance, test results) from all 79 English NHS breast screening centres between January 1, 1988 and March 31, 2018 were linked to cancer registry (cancer characteristics and treatment) and national mortality data. Data quality was assessed using comparability, validity, timeliness, and completeness. RESULTS: Screening records were extracted from 76/79 English breast screening centres, 3/79 were not possible due to software issues. Data linkage was successful from 1997 after introduction of a universal identifier for women (NHS number). Prior to 1997 outcome data are incomplete due to linkage issues, reducing validity. Between January 1, 1997 and March 31, 2018, a total of 11 262 730 women were offered screening of whom 9 371 973 attended at least one appointment, with 139 million person-years of follow-up (a median of 12.4 person years for each woman included) with 73 810 breast cancer deaths and 1 111 139 any-cause deaths. Comparability to reference data sets and internal validity were demonstrated. Data completeness was high for core screening variables (>99%) and main cancer outcomes (>95%). CONCLUSIONS: The ATHENA-M project has created a large high-quality and representative data set of individual women's screening trajectories and outcomes in England from 1997 to 2018, data before 1997 are lower quality. ADVANCES IN KNOWLEDGE: This is the most complete data set of English breast screening records and outcomes constructed to date, which can be used to evaluate and optimize screening.

Channel semantic mutual learning for visible-thermal person re-identification.

Visible-infrared person re-identification (VI-ReID) is a cross-modality retrieval issue aiming to match the same pedestrian between visible and infrared cameras. Thus, the modality discrepancy presents a significant challenge for this task. Most methods employ different networks to extract features that are invariant between modalities. While we propose a novel channel semantic mutual learning network (CSMN), which attributes the difference in semantics between modalities to the difference at the channel level, it optimises the semantic consistency between channels from two perspectives: the local inter-channel semantics and the global inter-modal semantics. Meanwhile, we design a channel-level auto-guided double metric loss (CADM) to learn modality-invariant features and the sample distribution in a fine-grained manner. We conducted experiments on RegDB and SYSU-MM01, and the experimental results validate the superiority of CSMN. Especially on RegDB datasets, CSMN improves the current best performance by 3.43% and 0.5% on the Rank-1 score and mINP value, respectively. The code is available at https://github.com/013zyj/CSMN.

Quality of ethnicity data within Scottish health records and implications of misclassification for ethnic inequalities in severe COVID-19: a national linked data study.

BACKGROUND: We compared the quality of ethnicity coding within the Public Health Scotland Ethnicity Look-up (PHS-EL) dataset, and other National Health Service datasets, with the 2011 Scottish Census. METHODS: Measures of quality included the level of missingness and misclassification. We examined the impact of misclassification using Cox proportional hazards to compare the risk of severe coronavirus disease (COVID-19) (hospitalization & death) by ethnic group. RESULTS: Misclassification within PHS-EL was higher for all minority ethnic groups [12.5 to 69.1%] compared with the White Scottish majority [5.1%] and highest in the White Gypsy/Traveller group [69.1%]. Missingness in PHS-EL was highest among the White Other British group [39%] and lowest among the Pakistani group [17%]. PHS-EL data often underestimated severe COVID-19 risk compared with Census data. e.g. in the White Gypsy/Traveller group the Hazard Ratio (HR) was 1.68 [95% Confidence Intervals (CI): 1.03, 2.74] compared with the White Scottish majority using Census ethnicity data and 0.73 [95% CI: 0.10, 5.15] using PHS-EL data; and HR was 2.03 [95% CI: 1.20, 3.44] in the Census for the Bangladeshi group versus 1.45 [95% CI: 0.75, 2.78] in PHS-EL. CONCLUSIONS: Poor quality ethnicity coding in health records can bias estimates, thereby threatening monitoring and understanding ethnic inequalities in health.

Costs of breast cancer recurrence after initial treatment for HR+, HER2-, high-risk early breast cancer: estimates from SEER-Medicare linked data.

OBJECTIVE: To assess the costs of treated recurrence and survival in elderly patients with early breast cancer (EBC) at high risk of recurrence using Surveillance Epidemiology and End Results (SEER) registry-Medicare linked claims data. METHODS: This retrospective study included patients aged ≥65 years with hormone receptor-positive (HR+), human epidermal growth factor receptor 2 negative (HER2-), node-positive EBC at high risk of recurrence. Treated recurrences were defined based on treatment events/procedure codes from claims. Primary outcomes were monthly total extra costs and cumulative extra costs of treated recurrence relative to patients with non/untreated recurrence. Costs were calculated using a Kaplan-Meier sampling average estimator method and inflated to 2021 US$. Secondary outcomes included analysis by recurrence type and overall survival (OS) after recurrence. Subgroup analysis evaluated costs in patients with Medicare Part D coverage. RESULTS: Among 3,081 eligible patients [mean (SD) age at diagnosis was 74.5 (7.1) years], the majority were females (97.4%) and white (87.8%). Treated recurrence was observed in 964 patients (31.3%). The monthly extra cost of treated recurrence was highest at the beginning of the first treated recurrence episode, with 6-year cumulative cost of $117,926. Six-year cumulative extra costs were higher for patients with distant recurrences ($168,656) than for patients with locoregional recurrences ($96,465). Median OS was 4.34 years for all treated recurrences, 1.92 years for distant recurrence, and 6.78 years for locoregional recurrence. Similar cumulative extra cost trends were observed in the subgroup with Part D coverage as in the overall population. LIMITATIONS: This study utilizes claims data to identify treated recurrence. Due to age constraints of the dataset, results may not extrapolate to a younger population where EBC is commonly diagnosed. CONCLUSION: EBC recurrence in this elderly population has substantial costs, particularly in patients with distant recurrences. Therapies that delay or prevent recurrence may reduce long-term costs significantly.

Cox regression with linked data.

Record linkage is increasingly used, especially in medical studies, to combine data from different databases that refer to the same entities. The linked data can bring analysts novel and valuable knowledge that is impossible to obtain from a single database. However, linkage errors are usually unavoidable, regardless of record linkage methods, and ignoring these errors may lead to biased estimates. While different methods have been developed to deal with the linkage errors in the generalized linear model, there is not much interest on Cox regression model, although this is one of the most important statistical models in clinical and epidemiological research. In this work, we propose an adjusted estimating equation for secondary Cox regression analysis, where linked data have been prepared by a third-party operator, and no information on matching variables is available to the analyst. Through a Monte Carlo simulation study, the proposed method is shown to lead to substantial bias reductions in the estimation of the parameters of the Cox model caused by false links. An asymptotically unbiased variance estimator for the adjusted estimators of Cox regression coefficients is also proposed. Finally, the proposed method is applied to a linked database from the Brest stroke registry in France.

The SIB Swiss Institute of Bioinformatics Semantic Web of data.

The SIB Swiss Institute of Bioinformatics (https://www.sib.swiss/) is a federation of bioinformatics research and service groups. The international life science community in academia and industry has been accessing the freely available databases provided by SIB since its inception in 1998. In this paper we present the 11 databases which currently offer semantically enriched data in accordance with the FAIR principles (Findable, Accessible, Interoperable, Reusable), as well as the Swiss Personalized Health Network initiative (SPHN) which also employs this enrichment. The semantic enrichment facilitates the manipulation of large data sets from public databases and private data sets. Examples are provided to illustrate that the data from the SIB databases can not only be queried using precise criteria individually, but also across multiple databases, including a variety of non-SIB databases. Data manipulation, be it exploration, extraction, annotation, combination, and publication, is possible using the SPARQL query language. Providing documentation, tutorials and sample queries makes it easier to navigate this web of semantic data. Through this paper, the reader will discover how the existing SIB knowledge graphs can be leveraged to tackle the complex biological or clinical questions that are being addressed today.

Research trends in ergonomics, industrial safety and health: semantic network and topic analyses.

We evaluated the research trends in ergonomics, industrial safety and health from the 1980s to the present. In the ergonomics area, keywords and abstracts from five journals were analyzed. For industrial safety and health, six journal databases were evaluated. A frequency analysis, a semantic network of keywords and a topic network of abstracts were conducted. The results of ergonomics showed that 'macro-ergonomics' and 'manual material handling' were the most popular topics, and 'ergonomic' and 'electromyography' were the most cited keywords. 'Posture' and 'biomechanics' were the most frequently used with high centrality. The results of industrial safety and health showed that 'job stress' and 'organizational safety' were the most popular topics, and 'occupational exposure' and 'occupational health' were the most cited keywords. 'Dust' and 'exposure' were frequently used with high centrality. The results would be helpful in understanding the trends of research efforts and foreseeing trends of future research.

Effects of short-term second language learning on the development of individual semantic networks in written and spoken language.

Previous studies have primarily focused on the relationship between native language (L1) and second language (L2) in the brain, specifically in one language modality, such as written or spoken language. However, there is limited research on how L2 proficiency impacts both modalities. This study aimed to investigate the functional networks involved in reading and speech comprehension for both L1 and L2, and observe changes in these networks as L2 proficiency improves. The dataset used in this study was obtained from a previous research conducted by Gurunandan et al., which involved Spanish-English bilingual participants undergoing a three-month English training program. Participants underwent fMRI scanning and performed a semantic animacy judgment task in both spoken and written language before and after training. Through analysis, distinct neural networks associated with spoken and written language were found between individuals' L1 and L2, both before and after training. Moreover, as L2 proficiency improved, the spoken and written networks for L2 remained distinct from those of the L1. These findings suggest that short-term L2 learning experiences can modify neural networks, but may not be enough to achieve native-like proficiency, supporting the accommodation hypothesis. These results have important implications for language learning and education, indicating that additional short-term training and exposure alone may not bridge the gap between L1 and L2 processing networks.

Fertility treatment pathways and births for women with and without polycystic ovary syndrome-a retrospective population linked data study.

OBJECTIVE: To study the fertility treatment pathways used by women with and without polycystic ovary syndrome (PCOS) and which pathways were more likely to result in a birth. DESIGN: This retrospective national community-based cohort study used longitudinal self-report survey data (collected 1996-2022; aged 18-49 years) from women born in 1973-1978 who are participants in the Australian Longitudinal Study on Women's Health. The study also used linked administrative data on fertility treatments (1996-2021). PATIENTS: Of the 8,463 eligible women, 1,109 accessed fertility treatment and were included. EXPOSURE: Polycystic ovary syndrome diagnosis was self-reported. MAIN OUTCOME MEASURE: use of ovulation induction (OI), intrauterine insemination, and/or in vitro fertilization (IVF) was established through linked administrative data. Births were self-reported. RESULTS: One in 10 of the eligible participants had PCOS (783/7,987, 10%) and 1 in 4 of the women who used fertility treatment had PCOS (274/1,109, 25%). Women with PCOS were 3 years younger on average at first fertility treatment (M = 31.4 years, SD = 4.18) than women without PCOS (M = 34.2 years, SD = 4.56). Seven treatment pathways were identified and use differed by PCOS status. Women with PCOS were more likely to start with OI (71%; odds ratio [OR] 4.20, 95% confidence interval [CI]: 2.91, 6.07) than women without PCOS (36%). Of the women with PCOS who started with OI, 46% required additional types of treatment. More women without PCOS ended up in IVF (72% vs. 51%). Overall, 63% (701/1,109) had an attributed birth, and in adjusted regressions births did not vary by last type of treatment (IVF: 67%, reference; intrauterine insemination: 67%, OR 0.94 95% CI: 0.56, 1.58; OI: 61%, OR 0.71, 95% CI: 0.52, 0.98), or by PCOS status (OR 1.27, 95% CI: 0.91, 1.77). By age, 74% of women under 35 years (471/639) and 49% of women 35 years or older had a birth. CONCLUSION: More women with PCOS used fertility treatment but births were equivalent to women without PCOS. Most women followed clinical recommendations. Births did not differ between pathways, so there was no disadvantage in starting with less invasive treatments (although there may be financial or emotional disadvantages).

Predicting risk of suicidal behavior from insurance claims data vs. linked data from insurance claims and electronic health records.

PURPOSE: Observational studies assessing effects of medical products on suicidal behavior often rely on health record data to account for pre-existing risk. We assess whether high-dimensional models predicting suicide risk using data derived from insurance claims and electronic health records (EHRs) are superior to models using data from insurance claims alone. METHODS: Data were from seven large health systems identified outpatient mental health visits by patients aged 11 or older between 1/1/2009 and 9/30/2017. Data for the 5 years prior to each visit identified potential predictors of suicidal behavior typically available from insurance claims (e.g., mental health diagnoses, procedure codes, medication dispensings) and additional potential predictors available from EHRs (self-reported race and ethnicity, responses to Patient Health Questionnaire or PHQ-9 depression questionnaires). Nonfatal self-harm events following each visit were identified from insurance claims data and fatal self-harm events were identified by linkage to state mortality records. Random forest models predicting nonfatal or fatal self-harm over 90 days following each visit were developed in a 70% random sample of visits and validated in a held-out sample of 30%. Performance of models using linked claims and EHR data was compared to models using claims data only. RESULTS: Among 15 845 047 encounters by 1 574 612 patients, 99 098 (0.6%) were followed by a self-harm event within 90 days. Overall classification performance did not differ between the best-fitting model using all data (area under the receiver operating curve or AUC = 0.846, 95% CI 0.839-0.854) and the best-fitting model limited to data available from insurance claims (AUC = 0.846, 95% CI 0.838-0.853). Competing models showed similar classification performance across a range of cut-points and similar calibration performance across a range of risk strata. Results were similar when the sample was limited to health systems and time periods where PHQ-9 depression questionnaires were recorded more frequently. CONCLUSION: Investigators using health record data to account for pre-existing risk in observational studies of suicidal behavior need not limit that research to databases including linked EHR data.