Boards, Advisory Committees, Professional Organizations


  • Resident physician, University of California San Frnacisco (2019 - Present)

Professional Education


  • Master of Applied Statistics, University of California San Francisco (2017)
  • Doctor of Medicine, Ewha Womans University (2007)
  • MAS, University of California San Francisco, Epidemiology and Biostatistics
  • MD, Ewha Womans University, Medicine

Stanford Advisors


Lab Affiliations


All Publications


  • Severity of substance use as an indicator of suicide risk among U.S. military veterans. Addictive behaviors Na, P. J., Nichter, B., Hill, M. L., Kim, B., Norman, S. B., Pietrzak, R. H. 2021; 122: 107035

    Abstract

    Substance use is a strong risk factor for suicidal behavior. To date, however, no known study has evaluated optimal cut-off scores on substance use severity measures to identify individuals who may be at elevated risk for suicide.Data were analyzed from the 2019-2020 National Health and Resilience in Veterans Study, which surveyed a nationally representative sample of 4069 U.S. veterans. Receiver operating characteristic curve analyses were conducted to identify cut-off scores on the Alcohol Use Disorders Identification Test (AUDIT) and Screen of Drug Use (SDU) that optimally differentiated veterans with and without past-year suicidal ideation (SI) in the full sample, as well a high-risk subsample with histories of major depressive disorder (MDD) and/or posttraumatic stress disorder (PTSD).The prevalence of past-year SI was 11.7% (n = 387); A total of 10.5% (n = 360) of the sample screened positive for past-year alcohol use disorder, 9.1% (n = 314) for past-year drug use disorder, and 22.4% (n = 833) for lifetime MDD/PTSD. The optimal AUDIT cut-off score that differentiated suicide ideators and non-ideators was ≥14 in the full sample, and ≥18 in the high-risk subsample; these scores are indicative of moderate-to-severe alcohol use disorder. The optimal SDU cut-off score was ≥1 for both the full sample and high-risk subsample, which is lower than the recommend cut-off score for probable drug use disorder (≥7).Results identify high-specificity thresholds on the AUDIT and SDU that may have clinical utility in suicide risk assessment in veterans. Veterans with moderate-to-severe alcohol use disorder or who use non-prescription drugs may warrant further suicide risk assessment.

    View details for DOI 10.1016/j.addbeh.2021.107035

    View details for PubMedID 34246987

  • Prenatal exposure to paternal smoking and likelihood for autism spectrum disorder. Autism : the international journal of research and practice Kim, B., Ha, M., Kim, Y. S., Koh, Y. J., Dong, S., Kwon, H. J., Kim, Y. S., Lim, M. H., Paik, K. C., Yoo, S. J., Kim, H., Hong, P. S., Sanders, S. J., Leventhal, B. L. 2021; 25 (7): 1946-1959

    Abstract

    What is Already Known about This Subject: Genetics, (including de novo mutations), environmental factors (including toxic exposures), and their interactions impact autism spectrum disorder etiology. Paternal smoking is a candidate risk for autism spectrum disorder due to biological plausibility, high prevalence, and potential intervention.What This Study Adds: This original study and its replication confirms that paternal factors can substantially contribute to autism spectrum disorder risk for their offspring. It specifically indicates that paternal smoking both before and during pregnancy contributes significantly to autism spectrum disorder risk.Implications for practice, research, or policy: Smoking prevention, especially in pregnancy planning, may decrease autism spectrum disorder risk in offspring.

    View details for DOI 10.1177/13623613211007319

    View details for PubMedID 33878952

    View details for PubMedCentralID PMC8419001

  • The influence of poverty attribution on attitudes toward suicide and suicidal thought: A cross-national comparison between South Korean, Japanese, and American populations. Comprehensive psychiatry Lee, H., Park, C. H., Rhee, S. J., Kim, B., Lee, S. S., Ha, K., Baik, C. J., Ahn, Y. M. 2021; 109: 152259

    Abstract

    Previous studies report that income inequality is an important risk factor for depression and suicide, and an increasing income gap appears inevitable. However, little study to date has investigated associations between the attribution of poverty and suicide. Though we previously reported associations between socio-cultural factors, including income, and suicide, we tried to explore more focused associations between income, attribution of poverty (individualistic, societal), permissive attitude toward suicide, and suicidal thought using a structural equation model.A total of 2213 participants from each of three nations (South Korea, Japan, and the United States) completed an online survey. Participants without a history of psychological disorders or suicide attempts completed scales measuring attributions of poverty, attitudes toward suicide, and severity of suicidal thoughts.We established a structural equation model, which exhibited a good fit for all nations, and compared significant path coefficients by country. South Korea had the highest severity of suicidal thought and societal attribution of poverty, followed by Japan and America. In all nations, a permissive attitude was positively related to the severity of suicidal thought and individualistic attribution of poverty was positively related to a permissive attitude toward suicide. Societal attribution of poverty was positively associated with a permissive attitude in Japan and the United States. Income was negatively associated with the severity of suicide in South Korea and the United States.Through an established structural equation model, we found the influence of poverty on suicide and identify the common and distinctive factors associated with suicide in each country.

    View details for DOI 10.1016/j.comppsych.2021.152259

    View details for PubMedID 34273607

  • An integrated model for the relationship between socio-cultural factors, Attitudes Toward Suicide, and intensity of suicidal ideation in Korean, Japanese, and American populations. Journal of affective disorders Lee, H., Park, C. H., Rhee, S. J., Kim, J., Kim, B., Lee, S. S., Ha, K., Baik, C. J., Ahn, Y. M. 2021; 280 (Pt A): 203-210

    Abstract

    Although many studies have identified the risk factors for suicide, the absence of a statistical model that integrates several risk areas makes it difficult to understand the abnormally high suicide rate in South Korea. Therefore, we aimed to establish a multidimensional structural equation model of suicide incorporating socio-cultural and psychiatric factors. We performed cross-national comparisons to identify the unique factors influencing South Korea's suicide rate.We conducted a web survey inviting adults aged 20- to 59-years from South Korea, Japan, and the United States to respond to questionnaires. A total of 2,213 subjects were included. We then developed a structural equation model, exhibiting a good fit in all countries.A permissive attitude was the factor that most strongly predicted the intensity of suicidal ideation in all countries. In South Korea, a low income was directly associated with both permissive attitude and the intensity of suicidal ideation. South Korea's highly fatalistic attitude related to more receptive attitudes toward suicide. Individual's resilience to stress provided significant protection against suicidal ideation in the United States and Japan, but not in South Korea.Since our sample excluded adults over 60 years, thus reducing the generalizability of our results. Furthermore, we employed a cross-sectional design; a longitudinal study is needed to draw causal inferences about suicidal death CONCLUSIONS: This study developed the first multidimensional, integrated statistical model of suicidal ideation. Our findings explain the causes of South Korea's high suicide rate and can be used to develop new interventions.

    View details for DOI 10.1016/j.jad.2020.10.042

    View details for PubMedID 33220555

  • Prevalence of and Factors Associated with School Bullying in Students with Autism Spectrum Disorder: A Cross-Cultural Meta-Analysis. Yonsei medical journal Park, I., Gong, J., Lyons, G. L., Hirota, T., Takahashi, M., Kim, B., Lee, S. Y., Kim, Y. S., Lee, J., Leventhal, B. L. 2020; 61 (11): 909–22

    Abstract

    Through this meta-analysis, we sought to examine the prevalence of, risks for, and factors associated with bullying involvement (victimization, perpetration, perpetration-victimization) among students with autism spectrum disorder (ASD). Additionally, we attempted to examine sources of variance in the prevalence and effect sizes of bullying in students with ASD across studies. Systematic database and literature review identified 34 relevant studies (31 for Western countries, three for Eastern countries). Pooled prevalence estimates for victimization, perpetration, and perpetration-victimization in general were 67%, 29%, and 14%, respectively. The risk of victimization in students with ASD was significantly higher than that in typically developing students and students with other disabilities. Further, deficits in social interaction and communication, externalizing symptoms, internalizing symptoms, and integrated inclusive school settings were related to higher victimization, and externalizing symptoms were related to higher perpetration. Finally, moderation analyses revealed significant variations in the pooled prevalences thereof depending on culture, age, school settings, and methodological quality and in the pooled effect sizes according to publication year and methodological quality. Our results highlight needs for bullying intervention for students with ASD, especially those who are younger, are in an inclusive school setting, and have higher social difficulties and externalizing/internalizing symptoms; for intensive research of bullying experiences among students with ASD in Eastern countries; and for efforts to improve the methodological quality of such research.

    View details for DOI 10.3349/ymj.2020.61.11.909

    View details for PubMedID 33107234

  • Identifying the Medical Lethality of Suicide Attempts Using Network Analysis and Deep Learning: Nationwide Study. JMIR medical informatics Kim, B., Kim, Y., Park, C. H., Rhee, S. J., Kim, Y. S., Leventhal, B. L., Ahn, Y. M., Paik, H. 2020; 8 (7): e14500

    Abstract

    Suicide is one of the leading causes of death among young and middle-aged people. However, little is understood about the behaviors leading up to actual suicide attempts and whether these behaviors are specific to the nature of suicide attempts.The goal of this study was to examine the clusters of behaviors antecedent to suicide attempts to determine if they could be used to assess the potential lethality of the attempt. To accomplish this goal, we developed a deep learning model using the relationships among behaviors antecedent to suicide attempts and the attempts themselves.This study used data from the Korea National Suicide Survey. We identified 1112 individuals who attempted suicide and completed a psychiatric evaluation in the emergency room. The 15-item Beck Suicide Intent Scale (SIS) was used for assessing antecedent behaviors, and the medical outcomes of the suicide attempts were measured by assessing lethality with the Columbia Suicide Severity Rating Scale (C-SSRS; lethal suicide attempt >3 and nonlethal attempt ≤3).Using scores from the SIS, individuals who had lethal and nonlethal attempts comprised two different network nodes with the edges representing the relationships among nodes. Among the antecedent behaviors, the conception of a method's lethality predicted suicidal behaviors with severe medical outcomes. The vectorized relationship values among the elements of antecedent behaviors in our deep learning model (E-GONet) increased performances, such as F1 and area under the precision-recall gain curve (AUPRG), for identifying lethal attempts (up to 3% for F1 and 32% for AUPRG), as compared with other models (mean F1: 0.81 for E-GONet, 0.78 for linear regression, and 0.80 for random forest; mean AUPRG: 0.73 for E-GONet, 0.41 for linear regression, and 0.69 for random forest).The relationships among behaviors antecedent to suicide attempts can be used to understand the suicidal intent of individuals and help identify the lethality of potential suicide attempts. Such a model may be useful in prioritizing cases for preventive intervention.

    View details for DOI 10.2196/14500

    View details for PubMedID 32673253

    View details for PubMedCentralID PMC7380907

  • Childhood vaccination as a protective factor for developmental psychopathology. Research in Autism Spectrum Disorders Kim, B., Ha, M., Kim, Y., Leventhal, B. B. 2020; 79: 101653
  • Predicting Planned Suicide Attempts With the Columbia-Suicide Severity Rating Scale: A Subanalysis of the 2013 Korea National Suicide Survey on Emergency Department Visitors. The Journal of nervous and mental disease Park, C. H., Kim, H., Kim, B., Kim, E. Y., Lee, H. J., Kim, D., Ahn, Y. M. 2019; 207 (2): 59-68

    Abstract

    Identifying predictors of planned suicide attempts (PSA) is critical because these are associated with grave consequences. Using data of suicide attempters visiting emergency departments, we investigated whether the Columbia-Suicide Severity Rating Scale (C-SSRS) subscales, by retrospectively evaluating ideation before an attempt, could predict the occurrence of PSA versus unplanned suicide attempts using logistic regression analyses. The severity subscale was used as a continuous (model A) and a categorical (model B) variable. In model A, higher scores on each subscale were associated with increased risk of PSA. In model B, the highest score on the severity subscale and a higher intensity subscale score predicted PSA. The severity and intensity subscales had areas under receiver operating curves of 0.712 and 0.688 with optimum cutoff points of 4/5 and 15/16, respectively. In addition, being aged 30 to 49 and 50 to 69 years, being male, interpersonal stress, and depressive and adjustment disorders increased PSA risk. The C-SSRS subscales, along with sociodemographic and clinical risk factors, can predict PSA.

    View details for DOI 10.1097/NMD.0000000000000921

    View details for PubMedID 30672875

  • Classification of attempted suicide by cluster analysis: A study of 888 suicide attempters presenting to the emergency department. Journal of affective disorders Kim, H., Kim, B., Kim, S. H., Park, C. H., Kim, E. Y., Ahn, Y. M. 2018; 235: 184-190

    Abstract

    It is essential to understand the latent structure of the population of suicide attempters for effective suicide prevention. The aim of this study was to identify subgroups among Korean suicide attempters in terms of the details of the suicide attempt.A total of 888 people who attempted suicide and were subsequently treated in the emergency rooms of 17 medical centers between May and November of 2013 were included in the analysis. The variables assessed included demographic characteristics, clinical information, and details of the suicide attempt assessed by the Suicide Intent Scale (SIS) and Columbia-Suicide Severity Rating Scale (C-SSRS). Cluster analysis was performed using the Ward method.Of the participants, 85.4% (n = 758) fell into a cluster characterized by less planning, low lethality methods, and ambivalence towards death ("impulsive"). The other cluster (n = 130) involved a more severe and well-planned attempt, used highly lethal methods, and took more precautions to avoid being interrupted ("planned"). The first cluster was dominated by women, while the second cluster was associated more with men, older age, and physical illness.We only included participants who visited the emergency department after their suicide attempt and had no missing values for SIS or C-SSRS.Cluster analysis extracted two distinct subgroups of Korean suicide attempters showing different patterns of suicidal behaviors. Understanding that a significant portion of suicide attempts occur impulsively calls for new prevention strategies tailored to differing subgroup profiles.

    View details for DOI 10.1016/j.jad.2018.04.001

    View details for PubMedID 29656265

  • Factors Affecting Suicide Method Lethality Among Suicide Attempters in the Korea National Suicide Survey. The Journal of nervous and mental disease Yeum, T. S., Kim, B., Kim, E. Y., Kim, S. H., Ha, K., Ahn, Y. M. 2018; 206 (3): 202-210

    Abstract

    This is the first national survey study in South Korea investigating the relationship between suicide lethality and the clinical information of suicide attempters. An interview questionnaire was used to assess their sociodemographic factors, medical and psychiatric information, and two suicide scales, the Columbia-Suicide Severity Rating Scale and the Suicide Intent Scale. Suicide methods were categorized as low and high lethality; low lethality covered drug overdose or self-cutting behavior, and high lethality covered all other methods. High and low lethality suicide method groups were significantly different in demographic, medical, and psychiatric factors. The two scale score distributions differed significantly across two groups, and the difference was also valid for the subcategory analyses of the Suicide Intent Scale. Multiple factors such as older age, male sex, no previous psychiatric history, and previous suicide attempt, as well as high suicide intent by means of suicide scale, affect selection of suicide method of high lethality.

    View details for DOI 10.1097/NMD.0000000000000735

    View details for PubMedID 28902068

  • Sex difference in risk period for completed suicide following prior attempts: Korea National Suicide Survey (KNSS). Journal of affective disorders Kim, B., Lee, J., Kim, E., Kim, S. H., Ha, K., Kim, Y. S., Leventhal, B. L., Ahn, Y. M. 2018; 227: 861-868

    Abstract

    We provide an opportunity for implementing preventive interventions to decrease suicide mortality among prior suicide attempters. We aim to identify sex-specific high risk periods and factors for later suicide death among suicide attempters.8537 suicide attempters of Korea National Suicide Survey were collected from January 1, 2007 to December 31, 2011 and data on suicide death was obtained as of December 31, 2012. The risk period and risk factors for later suicide death was computed by Kaplan-Meier survival estimates and by plotting the hazard function using the Epanechnikov Kernal smoothing method and cox proportional hazard regression modeling.The hazard for later suicide death was significant up to 10 months for females and 20 months for males. Age 50-69 years (HR, 3.29; [CI: 1.80-6.02] and not being intoxicated with alcohol (HR, 1.94 [1.27-2.97])) in male attempters were significant risk factors for later suicide death.Risk for later suicide death was significantly increased during the first full year following index attempts for all with an addition 8 months of risk for males, especially those of advanced age who were sober at the time of attempt.

    View details for DOI 10.1016/j.jad.2017.11.013

    View details for PubMedID 29310206

  • Sex-specific association between the albumin D-element binding protein gene and metabolic syndrome in patients with bipolar disorder and schizophrenia. Psychiatry research Kim, E. Y., Kim, S. H., Lee, H. J., Kim, B., Kim, Y. S., Ahn, Y. M. 2016; 240: 47-52

    View details for DOI 10.1016/j.psychres.2016.03.040

    View details for PubMedID 27084990

  • Characteristics of methods of suicide attempts in Korea: Korea National Suicide Survey (KNSS). Journal of affective disorders Kim, B., Ahn, J. H., Cha, B., Chung, Y. C., Ha, T. H., Hong Jeong, S., Jung, H. Y., Ju, G., Kim, E. Y., Kim, J. M., Kim, M. D., Kim, M. H., Kim, S. I., Lee, K. U., Lee, S. H., Lee, S. J., Lee, Y. J., Moon, E., Ahn, Y. M. 2015; 188: 218-25

    Abstract

    Because the method used for a suicide attempt is an important determinant of outcome, these methods should be explored. The present study was a nationwide investigation of suicide attempts and the characteristics of suicidal behavior.To compare the suicide methods used in attempted suicides with those used in completed suicides and to examine the factors associated with each phenomenon.The present study reviewed the medical charts of subjects who had attempted suicide and subsequently visited the emergency rooms of 17 medical centers from May 1, 2013 to November 7, 2013. All subjects completed a full psychiatric interview conducted by trained psychiatric residents. Suicide-attempt methods were divided into the following six categories: drug poisoning, pesticide poisoning, gassing, cutting, hanging, and others. The associations among demographic variables, related psychiatric variables, and suicide-attempt methods were analyzed using a multinomial regression analysis.Of the 1359 suicide attempts or instrumental suicide-related behaviors with/without injuries and the 14,160 completed suicides, drug poisoning and cutting were the most common suicidal behaviors with/without injuries, but they were the least frequent method of completed suicides. In contrast, hanging and jumping from a height were less common among failed suicide attempts but resulted in a higher percentage of fatalities. Being male, age, and area of residence were associated with pesticide poisoning, whereas previous suicide attempts were associated with cutting, pesticide poisoning, and gassing.A previous suicide attempt is a risk factor for suicide; thus, assessing the characteristics of suicide attempts or instrumental suicide-related behaviors with/without injuries is necessary to prevent these attempts. The present findings showed that the methods of suicide used by individuals who only attempted suicide differed from those used by individuals who completed. Of the suicide methods, pesticide poisoning was related to age, residential area (urban), and a history of previous suicide attempts.

    View details for DOI 10.1016/j.jad.2015.08.050

    View details for PubMedID 26368946

  • Subjective depressive symptoms and metabolic syndrome among the general population. Progress in neuro-psychopharmacology & biological psychiatry Rhee, S. J., Kim, E. Y., Kim, S. H., Lee, H. J., Kim, B., Ha, K., Yoon, D. H., Ahn, Y. M. 2014; 54: 223-30

    Abstract

    The evidence of the association between depression and metabolic syndrome is increasing, but the existence of sex differences in this association remains controversial. The aim of this study was to investigate the association between subjective depressive symptoms and metabolic syndrome and each of its components by sex in the Korean population.The study sample comprised 15,073 men and 15,034 women who underwent routine health examinations. They completed the Beck Depression Inventory for depressive symptoms, and medical examinations provided data regarding metabolic syndrome. Adjustments for age, marriage, cigarette smoking, alcohol use, exercise, education, cancer, stroke, angina, and thyroid disease were performed. The association between depressive symptoms and metabolic syndrome and each of its components was analyzed by multiple logistic regression.In women, depressive symptoms were associated with metabolic syndrome (OR=1.35, 95% CI=1.11-1.64, p=0.002) and the high-density lipoprotein cholesterol component (OR=1.26, 95% CI=1.09-1.46, p=0.002) of metabolic syndrome. There was also an association between the severity of depressive symptoms and metabolic syndrome in women (OR=1.046, 95% CI=1.002-1.091, p=0.039). In men, depressive symptoms were inversely associated with the hypertension component of metabolic syndrome (OR=0.73, 95% CI=0.58-0.91, p=0.005).Subjective depressive symptoms were associated with metabolic syndrome only in women. Further research should consider sex differences and dyslipidemia.

    View details for DOI 10.1016/j.pnpbp.2014.06.006

    View details for PubMedID 24975752

  • Follicle-stimulating hormone (FSH), current suicidal ideation and attempt in female patients with major depressive disorder. Psychiatry research Kim, B., Kang, E. S., Fava, M., Mischoulon, D., Soskin, D., Yu, B. H., Lee, D., Lee, D. Y., Park, H. D., Jeon, H. J. 2013; 210 (3): 951-6

    Abstract

    Current suicidal ideation and attempts are more commonly found in female patients with major depressive disorder (MDD) than in males. However, little is known about the relationship between activity of female reproductive hormones and suicide. The study population consisted of 490 female MDD patients of age ≥18. They were assessed by the Mini-International Neuropsychiatric Interview. At the same visit, we measured blood Follicle-Stimulating Hormone (FSH), Luteinizing Hormone (LH), estradiol, progesterone, Adrenocorticotropic Hormone (ACTH), cortisol, thyroid hormones, and prolactin. Blood FSH showed a significant difference among female MDD patients with suicide attempt, those with ideation, and those without within the previous month. Post-hoc analysis also showed that FSH was significantly lower in MDD patients with suicide attempt and ideation than those without, whereas other hormones showed no differences between those with and without attempt. FSH was negatively associated with current suicidality scores after adjustment for age and education years in all age groups. FSH was significantly lower in those with current suicide ideation or attempt than those without in age 45 years or under, but not in other age groups. In conclusion, blood FSH is significantly lower in female MDD patients with current suicide attempt or ideation than those without, especially in age 45 years or under.

    View details for DOI 10.1016/j.psychres.2013.08.057

    View details for PubMedID 24080256

  • Temperament affects sympathetic nervous function in a normal population. Psychiatry investigation Kim, B., Lee, J. H., Kang, E. H., Yu, B. H. 2012; 9 (3): 293-7

    Abstract

    Although specific temperaments have been known to be related to autonomic nervous function in some psychiatric disorders, there are few studies that have examined the relationship between temperaments and autonomic nervous function in a normal population. In this study, we examined the effect of temperament on the sympathetic nervous function in a normal population.Sixty eight healthy subjects participated in the present study. Temperament was assessed using the Korean version of the Cloninger Temperament and Character Inventory (TCI). Autonomic nervous function was determined by measuring skin temperature in a resting state, which was recorded for 5 minutes from the palmar surface of the left 5th digit using a thermistor secured with a Velcro® band. Pearson's correlation analysis and multiple linear regression were used to examine the relationship between temperament and skin temperature.A higher harm avoidance score was correlated with a lower skin temperature (i.e. an increased sympathetic tone; r=-0.343, p=0.004) whereas a higher persistence score was correlated with a higher skin temperature (r=0.433, p=0.001). Hierarchical linear regression analysis revealed that harm avoidance was able to predict the variance of skin temperature independently, with a variance of 7.1% after controlling for sex, blood pressure and state anxiety and persistence was the factor predicting the variance of skin temperature with a variance of 5.0%.These results suggest that high harm avoidance is related to an increased sympathetic nervous function whereas high persistence is related to decreased sympathetic nervous function in a normal population.

    View details for DOI 10.4306/pi.2012.9.3.293

    View details for PubMedID 22993530

    View details for PubMedCentralID PMC3440480