Background: Isotretinoin treatment for acne can reduce adverse psychiatric outcomes in adults, but there has been little investigation of the incidence of psychiatric outcomes in treated adolescents. Methods: This retrospective cohort study using the Rochester Epidemiology Project identified 606 patients aged 12-18 prescribed isotretinoin over a 10-year period between January 1, 2008 and December 31, 2017. Medical records were reviewed to identify psychiatric diagnoses before and during isotretinoin therapy, as well as psychiatric symptoms not captured by formal diagnoses and changes to isotretinoin dosing because of psychiatric diagnoses or symptoms. Results: One hundred seventy-seven (29.2%) had a psychiatric diagnosis prior to isotretinoin initiation, but 98 (16.2%) had a new psychiatric diagnosis or psychiatric symptom while taking isotretinoin. Patients with a psychiatric history were no more likely than those without to receive a new psychiatric diagnosis during treatment (4.5% vs. 3.7%; p = .650), but did experience more psychiatric symptoms, primarily low mood and mood swings (23.7% vs. 7.7%; p < .001). Only 25.5% of the 98 with a new psychiatric diagnosis or psychiatric symptom had a subsequent dose change. A dose change was more likely if patients received a new psychiatric diagnosis (41.7% vs. 20.3%; p = .037) or patients did not have a psychosocial explanation for psychiatric symptoms (34.4% vs. 10.8%; p = .009). Conclusions: A substantial proportion of adolescent patients prescribed isotretinoin had a prior psychiatric diagnosis. This predicts more psychiatric symptoms during isotretinoin treatment. Adolescents with a psychiatric history who have worsening symptoms and those with new-onset psychiatric symptoms would benefit from close monitoring while taking isotretinoin.

Introduction/Objectives: We aimed to investigate the effect of family structure on depression program outcomes for adolescents enrolled in a depression-focused, primary care-based collaborative care program. Methods: This was a retrospective study of primary care patients ages 12 to 18 years seen at a Midwestern academic center with data obtained by medical record review. We used logistic regression models to assess the effect of family structure on program graduation and achievement of a single Patient Health Questionnaire 9–Modified for Adolescents (PHQ-9M) score <5 at any time while enrolled. Results: Adolescents were divided into 2 groups, Both Parents in Household (n = 179) and Parents Not Together (n = 161). The Both Parents in Household group had higher rates of graduation (38.0% vs 23.6%, P = .005) and achieving single PHQ-9M scores <5 (64.1% vs 46.2%, P = .002) than the Parents Not Together group. Discussion: Youth residing with both parents had higher rates of successful outcomes in a depression-focused collaborative care program.

Purpose To examine associations of social support and social isolation with burnout, program satisfaction, and organization satisfaction among a large population of U.S. residents and fellows and to identify correlates of social support and social isolation. Method All residents and fellows enrolled in graduate medical education programs at Mayo Clinic sites were surveyed in February 2019. Survey items measured social support (emotional and tangible), social isolation, burnout, program satisfaction, and organization satisfaction. Factors of potential relevance to social support were collected (via the survey, institutional administrative records, and interviews with program coordinators and/or program directors) and categorized as individual, interpersonal, program, or work-related factors (duty hours, call burden, elective time, vacation days used before survey administration, required away rotations, etc.). Multivariable regression analyses were conducted to examine relationships between variables. Results Of 1,146 residents surveyed, 762 (66%) from 58 programs responded. In adjusted models, higher emotional and tangible support were associated with lower odds of burnout and higher odds of program and organization satisfaction, while higher social isolation scores were associated with higher odds of burnout and lower odds of program satisfaction and organization satisfaction. Independent predictors of social support and/or social isolation included age, gender, relationship status, parental status, postgraduate year, site, ratings of the program leadership team, ratings of faculty relationships and faculty professional behaviors, satisfaction with autonomy, and vacation days used before survey administration. Conclusions This study demonstrates that social support and social isolation are strongly related to burnout and satisfaction among residents and fellows. Personal and professional relationships, satisfaction with autonomy, and vacation days are independently associated with social support and/or social isolation, whereas most program and work-related factors are not. Additional studies are needed to determine if social support interventions targeting these factors can improve well-being and enhance satisfaction with training.

Background The treatment of depression in children and adolescents is a substantial public health challenge. This study examined artificial intelligence tools for the prediction of early outcomes in depressed children and adolescents treated with fluoxetine, duloxetine, or placebo. Methods The study samples included training datasets ( N = 271) from patients with major depressive disorder (MDD) treated with fluoxetine and testing datasets from patients with MDD treated with duloxetine ( N = 255) or placebo ( N = 265). Treatment trajectories were generated using probabilistic graphical models (PGMs). Unsupervised machine learning identified specific depressive symptom profiles and related thresholds of improvement during acute treatment. Results Variation in six depressive symptoms (difficulty having fun, social withdrawal, excessive fatigue, irritability, low self‐esteem, and depressed feelings) assessed with the Children’s Depression Rating Scale‐Revised at 4–6 weeks predicted treatment outcomes with fluoxetine at 10–12 weeks with an average accuracy of 73% in the training dataset. The same six symptoms predicted 10–12 week outcomes at 4–6 weeks in (a) duloxetine testing datasets with an average accuracy of 76% and (b) placebo‐treated patients with accuracies of 67%. In placebo‐treated patients, the accuracies of predicting response and remission were similar to antidepressants. Accuracies for predicting nonresponse to placebo treatment were significantly lower than antidepressants. Conclusions PGMs provided clinically meaningful predictions in samples of depressed children and adolescents treated with fluoxetine or duloxetine. Future work should augment PGMs with biological data for refined predictions to guide the selection of pharmacological and psychotherapeutic treatment in children and adolescents with depression.

The treatment of depression in children and adolescents is a substantial public health challenge. This study examined artificial intelligence tools for the prediction of early outcomes in depressed children and adolescents treated with fluoxetine, duloxetine, or placebo. The study samples included training datasets ( N  = 271) from patients with major depressive disorder (MDD) treated with fluoxetine and testing datasets from patients with MDD treated with duloxetine ( N  = 255) or placebo ( N  = 265). Treatment trajectories were generated using probabilistic graphical models (PGMs). Unsupervised machine learning identified specific depressive symptom profiles and related thresholds of improvement during acute treatment. Variation in six depressive symptoms (difficulty having fun, social withdrawal, excessive fatigue, irritability, low self-esteem, and depressed feelings) assessed with the Children’s Depression Rating Scale-Revised at 4–6 weeks predicted treatment outcomes with fluoxetine at 10–12 weeks with an average accuracy of 73% in the training dataset. The same six symptoms predicted 10–12 week outcomes at 4–6 weeks in (a) duloxetine testing datasets with an average accuracy of 76% and (b) placebo-treated patients with accuracies of 67%. In placebo-treated patients, the accuracies of predicting response and remission were similar to antidepressants. Accuracies for predicting nonresponse to placebo treatment were significantly lower than antidepressants. PGMs provided clinically meaningful predictions in samples of depressed children and adolescents treated with fluoxetine or duloxetine. Future work should augment PGMs with biological data for refined predictions to guide the selection of pharmacological and psychotherapeutic treatment in children and adolescents with depression.

Purpose Physician suicide rates are reportedly higher than those of the general population, but medical student suicide rates are not well studied. It is difficult to determine whether physician suicide rates can be predicted by medical student risk factors for suicide and difficult to identify those risk factors without knowing medical student suicide rates. The authors systematically reviewed the literature to collate data on medical student suicide rates. Method The authors searched the PubMed, Web of Science, and Library of Congress databases for papers published in any language before November 11, 2017. They identified 3,429 papers; after the initial screening process, they assessed 82 full-text articles for eligibility. Twelve ultimately met the full inclusion criteria; meta-analysis was not possible. Data regarding medical student suicide numbers and rates were extracted and compared with contemporaneous general population suicide rates using public epidemiological data, when available. Results Medical student suicide rates were infrequently reported in the historical and international literature, and data collection techniques were inconsistent. Generally, U.S. medical student suicide rates were lower than those of the contemporaneous general population. Proportionate mortality of medical students (number of deaths by a particular cause such as suicide divided by total number of deaths) was not reported in the literature. Conclusions Gaps exist in knowledge of medical student suicide rates, risk factors, and targets for intervention. Significant barriers have impeded information collection. Yet, more comprehensive data collection is needed to understand suicide risk in this population and to implement and improve effective intervention strategies.

OBJECTIVES The Patient Health Questionnaire-9 Modified (PHQ-9M) is a self-report tool used to assess the presence and severity of depressive symptoms in teenagers. Despite widespread use in primary care clinics and psychiatric settings, the PHQ-9M has not been validated nor are its psychometric properties adequately understood for the adolescent population. This study sought to examine the psychometrics of the PHQ-9M in treatment-seeking, depressed adolescents at a psychiatric psychopharmacology clinic who were concurrently assessed with the Children's Depression Rating Scale Revised (CDRS-R) and Quick Inventory of Depressive Symptomatology-Adolescent (17-item) Self-Report (QIDS-A17-SR). METHODS Adolescents (N = 160) aged 13 through 18 years with a diagnosis of major depressive disorder, determined on the basis of a clinical interview and semi-structured interview using the Kiddie Schedule for Affective Disorders and Schizophrenia-Present and Lifetime Version, were assessed for severity of depressive symptoms with the PHQ-9M, CDRS-R (adolescent interview only), and QIDS-A17-SR assessments at baseline, 4, and 8 weeks. Classical test theory analysis was used to evaluate the internal consistency and dimensionality of the PHQ-9M. Convergent validity was evaluated via intraclass correlations of the PHQ-9M with the CDRS-R and QIDS-A17-SR. Sensitivity to treatment response was also evaluated. RESULTS The internal consistency (Cronbach's coefficient α) at baseline, 4, and 8 weeks was 0.879, 0.859, and 0.827 for the PHQ-9M; 0.739, 0.835, and 0.867 for CDRS-R; and 0.712, 0.777, and 0.804 for QIDS-A17-SR, respectively. The PHQ-9M had moderate convergent validity with the CDRS-R but good convergent validity with the QIDS-A17-SR. The PHQ-9M was less sensitive to changes in symptom severity than the CDRS-R and QIDS-A17-SR. CONCLUSIONS The PHQ-9M appears to be a valid and reliable assessment tool for the severity of depressive symptoms in a psychiatric clinic setting. However, its utility as a treatment outcome measure may be limited compared with other available rating scales.

The collaborative care model for depression management in the primary care setting has been shown to both improve outcomes and reduce the overall cost of care in adults with depressive disorders. Limited information is available to guide implementation of this model with pediatric populations.

A 15-year-old boy, whose family moved from Sudan to the United States as refugees five years ago, is brought to the Emergency Department. His mother reports that he has always been healthy and does well in school, but over the last three days, he has been speaking to unseen presences, appearing intermittently agitated, sleeping erratically, and eating and drinking very little. The family says they have tried helping him with prayer, but it has made little difference, and they are concerned that he may be dehydrated.

We thank Dr. Ghaziuddin for his interest in our article and his clarification of important points. He emphasizes the heterogeneity of Muslim adolescents and families, and cautions the reader to manage potential bias so as not to negatively affect clinical judgment. He also rightly points out that attribution of psychiatric symptoms to spiritual phenomena occurs across many religious and cultural backgrounds. We fully agree, and we appreciate Dr. Ghaziuddin’s emphasis on individualizing assessment of each patient and family.

Many psychiatric and medical illnesses as well as normal reactions to stressors have symptoms that overlap with those of depressive disorders, including outwardly sad or dysphoric appearance, irritability, apathy or amotivation, fatigue, difficulty making decisions, social withdrawal, and sleep disturbances. This cluster of symptoms forms a readily observable behavioral phenotype that clinicians may label as before considering a broader differential diagnosis. To better understand what other conditions belong in the differential diagnosis, we reviewed a sample of 100 consecutive medical/ surgical inpatients referred to our consultationliaison psychiatry practice for evaluation of Ultimately, only 29 of these patients received a diagnosis. Many of the other diagnoses given in our sample required attention during inpatient medical or surgical care because they were potentially life-threatening if left unaddressed--such as delirium--or they interfered with managing the primary medical or surgical condition for which the patient was hospitalized. Hurried or uncertain primary care clinicians frequently use depression as a catch-all term when requesting psychiatric consultation for patients who seem depressed. A wide range of conditions can mimic depression, and the art of psychosomatic psychiatry includes considering protean possibilities when assessing a patient. We identified 7 diagnoses that mimic major and developed our 8 D differential to help clinicians properly diagnose depressed patients who have something other than a depressive disorder. Although our sample consisted of hospitalized patients, these mimics of may be found among patients referred from other clinical settings for evaluation of possible depression. [ILLUSTRATION OMITTED] [ILLUSTRATION OMITTED] [ILLUSTRATION OMITTED] The perils of misdiagnosis Depression is common among patients hospitalized with medical or surgical conditions. DSM-IV-TR diagnostic criteria for a major depressive episode (MDE) include the presence of low mood and/or anhedonia, plus [greater than or equal to]4 other depressive symptoms for[greater than or equal to]2 weeks. (1) Growing evidence suggests that the relationship between and morbidity and mortality in medical illness is bidirectional, and nonpsychiatrists are becoming increasingly aware of major depression's serious impact on their patients' physical health. (2-5) Although improving nonpsychiatrists' recognition of in medically ill patients is laudable, it comes with a high falsepositive rate. In a study of primary care outpatients, Berardi et al found that 45% of patients labeled depressed did not meet ICD-10 criteria for major depression, but >25% of those patients were prescribed an antidepressant.6 In a large retrospective study, Boland et al found that approximately 40% of patients referred to an inpatient psychiatric consultation service for did not meet criteria for a depressive illness, and primary medical services often confused organic syndromes such as delirium and dementia with depression.' Similarly, Clarke et al found that 26% of medical and surgical inpatients referred to psychiatry with depression had another diagnosis commonly delirium that better accounted for their symptoms. (8) What is the harm in overdiagnosing depression? Missing a serious or lifethreatening diagnosis is a primary concern. For example, unrecognized delirium, which frequently was misdiagnosed as in the Berardi, (6) Boland (7) and Clarke (8) studies, is associated with myriad difficulties, including higher morbidity and mortality. (9) Substance use disorders, which also commonly masquerade as depression, frequently are comorbid with medical illness. Delays in appropriate treatment of withdrawal syndromes particularly of alcohol and sedative/hypnotic medications--are risk factors for increased mortality in these illnesses. …

By design or by default, primary care providers (PCPs)are frequently the vanguard in the fight against suicide. Recent studies have highlighted programs to improve screening and prevention of suicidality in the medical home, particularly among high-risk patients, such as adolescents, the elderly, and veterans. Increasing efforts are also being paid to improving the PCP's skill in assessing for suicidality. However, it is becoming increasingly apparent that screening alone will not significantly lower suicide rates until it occurs within a well-integrated system that facilitates timely referral to more intensive mental health services for those patients who need them. Unfortunately, such systems are sorely lacking in many, if not most, areas of the USA.

Factors contributing to patients killing themselves while admitted to general hospital medical/surgical (med/surg) units have not been well described. These rare yet devastating suicides appear to have characteristics distinguishing them from suicides in psychiatric inpatients. This article emphasizes the importance of both agitation and readily available lethal means in suicides that are almost invariably impulsive. It also emphasizes how traditional risk factors such as past history of psychiatric illness, substance abuse, or suicidality typically are absent in this population, as are present depression and known suicidality. Caregivers seeking to prevent suicide in the med/surg environment therefore must appreciate the potential lethality of acute psychic and motoric agitation. Close surveillance of agitated patients, with interventions to calm them and secure their surroundings, will assure safety and save lives.

Ontological and Other Assumptions Lloyd A. Wells (bio) and Sandra J. Rackley (bio) Fahrenberg and Cheetham have conducted an immensely thought-provoking study of the assumptions about human nature made by 800 students and pose a question about the future impact of these assumptions on individuals’ practice in professions including medicine and psychotherapy. This work represents a branch of “philosophical anthropology,” which considers assumptions people make about human nature. The authors used a questionnaire, much of which was newly designed, to assess the assumptions made by university students studying psychology, philosophy, sciences, and other disciplines. Most of the respondents were studying at Freiburg, but some were sampled at several other universities in both the former East and West Germanies. Questions involved consciousness and the brain, evolution, free will, belief in God, the meaning of life, and theodicy. Given the importance of these beliefs and assumptions in so many peoples’ lives, it is quite surprising that there are so few modern data available, and this study ventures an important first step in assessing these assumptions. However, there are many problems inherent in this study (and any like it). The first, for us, is the validity of the questionnaire itself, an issue that was not specifically addressed in the paper. Face validity is a problem even in the few questions quoted directly in the paper. For example, some statements contain two clauses, and one could well agree with one without agreeing with another: “I do not know whether a God exists and I do not believe it is possible to know.” This is the most relevant answer for a true agnostic, but the assumption that “I do not know whether a God exists” does not necessarily link at all with “I do not believe it is possible to know.” Similarly, a theist might have trouble with the two clauses, “I know God really exists” and “I do not doubt this.” (Technically, knowledge does imply a lack of doubt, but belief certainly does not.) Particularly in the trilemmas, where various patterns of response were analyzed, these multiple-clause statements may have led students to respond in a manner inconsistent with their true beliefs and assumptions. Furthermore, many of the questions contain terms that may have been unfamiliar to first-year university students. The authors explain their rationale in not defining terms, but the lack of clarity about definitions may have led some of the students to not fully understand the questions asked and to perhaps influence the answers they chose. Misunderstanding of the questions may account for some of the disparities reported in the results. Other apparent inconsistencies, however, do not seem to be a result of unclear questions. For example, sixty percent of the students agreed with the statement that “I believe in … existence after death,” whereas forty-seven percent agreed with the statement that “when my brain is dead my consciousness and my person cease to be.” These [End Page 203] findings do not add up, and this study has several similar sets of data. People contradict themselves in long and complex questionnaires, and this may be a simple explanation for these findings, but is this the only explanation? How seriously did the students take the questionnaire, often given as part of their class? How honest were their responses? How generalizable are the findings of this study? It is very difficult even to speculate about this. The statistics are incomprehensible in light of the data given, and one must essentially take them on faith, wherever one falls on the God/theodicy questions. The design has many flaws, but we do learn how a subsample of German university students rate important factors about the meaning of life. Statements are made in the paper that the students’ results correlate with similar results for the general population, but this assertion seems to be based on a very few questions. One would like much more delineation of this apparent correlation, because we wonder about the juxtaposition of developmental factors with the answers the students gave to this questionnaire. First-year university students are at a critical developmental point, and it would be fascinating to follow this large cohort of students over time to see whether...