In the Clinical Challenge, “Treatment of First-Episode Schizophrenia in a Young Woman,”1 published online October 30, 2019, there was an error in a quiz answer choice. Answer A was corrected to “Consider the risk of adverse events and which drugs are approved for use” instead of “Commence treatment with oral amisulpride.” In addition, the Discussion and Patient Outcome sections have been corrected to note that the drug amisulpride is not approved for use in the United States but is approved for use in the United Kingdom where the patient was treated. This article has been corrected online.

In Reply In our study,1 we used 3 waves of data from the longitudinal Population Assessment of Tobacco and Health study to show that adolescents who report more frequent social media use report greater symptoms of psychopathology 1 year later. In their letter, Keyes and Kreski replicated our findings. They then showed that, after further adjusting for psychopathology measured concurrently with social media use, the prospective association we identified disappears, except for among the most frequent users of social media use.

To the Editor Riehm et al1 report an association between social media use and increased risk of internalizing and combined internalizing/externalizing problems 1 year later (between waves 2 and 3 of Population Assessment of Tobacco and Health study data), controlling for psychopathology and alcohol/marijuana use at wave 1. However, the analytic strategy is vulnerable to substantial residual confounding.

Evidence has accumulated over the last 40 years indicating increased rates of violent perpetration and being the target of violence among people with mental illness. Landmark data collected in the early 1980s by the Epidemiologic Catchment Area study found a 12-month prevalence of 12% for any type of violence among people with mental disorders, which dropped to 7% if people with comorbid substance abuse (DSM-III) were excluded, compared with 2% in the general population.1 Similar magnitudes of increased risk have been found in subsequent studies2 and meta-analyses,3 although the absolute rates of violence vary, probably because of differences in the populations sampled, outcome measures, and means of ascertainment.4 However, despite the greater risk associated with mental disorders, the proportion of violence accounted for by mental disorder is small, with 1-year attributable risk estimated at 4% in the Epidemiologic Catchment Area analyses.1

Cannabidiol (CBD), a phytocannabinoid with potential medicinal properties, has recently hit all levels of society with tidal wave force that very few saw coming. After being virtually unknown by most people not even a decade ago, CBD is now in the general lexicon with numerous companies riding the CBD wave: more than 1000 products are being sold through the internet, dispensaries, pharmacies, large national retail stores, boutique shops, and local bodegas. This, along with extensive media coverage, has made CBD a household name and spawned an international business with an estimated market value of more than $60 billion within the next few years.

Antipsychotic medications are the primary pharmacologic treatment for schizophrenia and were initially discovered fortuitously in the early 1950s. Chlorpromazine was the prototype of this class of medications, retroactively noted as first-generation antipsychotics, which are primarily characterized by their high-affinity antagonism at the dopamine D2 receptor. The rediscovery of clozapine in 1989 heralded the initial era of second-generation antipsychotics, which extended through 2002. In general, second-generation antipsychotics have lower D2 affinity, along with high serotonin 5-HT2A antagonist affinity. Subsequently, several new second-generation drugs were introduced: asenapine and iloperidone in 2009, lurasidone hydrochloride in 2010, and cariprazine and brexpiprazole in 2015.

In early 2019, representatives of the US Department of Justice sued to prevent a nonprofit group from opening a supervised consumption facility in Philadelphia, Pennsylvania. This case, United States v Safehouse, has drawn considerable attention given its implications for US drug policies; however, one exchange during an August 2019 hearing was particularly disconcerting. After Safehouse president Jose Benitez testified that staff affiliated with a syringe exchange program had reversed more than 500 opioid overdoses with naloxone in 2018, William McSwain, the US Attorney for the Eastern District of Pennsylvania, contested whether these overdose reversals counted as “saved lives.” Arguing that a naloxone recipient might use substances later and die anyways, McSwain said, “You’re saying you saved a life when, in fact, someone just died.”

In the Original Investigation titled “Association of Maternal Eating Disorders With Pregnancy and Neonatal Outcomes,”1 published online November 20, 2019, the surname of the second author was inaccurate. Angelica Lindén Hirschberg should be referred to as Dr Hirschberg, not Dr Lindén Hirschberg. The error has been corrected.

The Original Investigation, “Effect of Stellate Ganglion Block Treatment on Posttraumatic Stress Disorder Symptoms: A Randomized Clinical Trial,”1 published online November 6, 2019, had an error in the Abstract. A sentence in the Design, Setting, and Participants subsection that referred to “anesthesiologists performing the procedures” should have instead said “physicians performing the procedures”; the physicians involved were not all anesthesiologists. The article was corrected online.

For the Original Investigation titled “Association of Trauma Type, Age of Exposure, and Frequency in Childhood and Adolescence With Psychotic Experiences in Early Adulthood,”1 published online November 21, 2018, and in the January 2019 print issue of JAMA Psychiatry, the authors requested a change to the publication license from standard license to Open Access license. The article is now published under the Open Access license and has been corrected online. The article was previously corrected.2

Medication selection in depression still relies primarily on trial and error, frustrating both patients and clinicians. Diagnostic tests that predict treatment response in advance could facilitate an informed approach and reduce suffering. Ideally, those tests would directly measure brain function. The primary research test of brain function, functional magnetic resonance imaging (fMRI), is too expensive to scale to routine care. Electroencephalography (EEG), however, is comparatively inexpensive and fast and could be made available in many settings. Plus, as a direct measure of localized brain activity, EEG is ideal for dissecting the connections between neurotransmission, symptoms, and pharmacologic response.

As a fan of popular music and sports, I am intrigued by top 10 rankings. Thus, I was glad to have a chance to select the 10 articles published in JAMA Psychiatry in 2019 that I think are particularly noteworthy. Of course, any such list, by necessity, excludes many important articles. What follows, with apologies to authors whose articles are not included, is a highly idiosyncratic view that reflects my thinking about the field and my teaching of the next generation of psychiatrists. Articles are grouped in 3 thematic areas for organization but are not a top 10 countdown or absolute ranking. In the sections below, I provide a brief summary of the selected articles and note that many of these articles had accompanying editorials that did a great job of discussing the work in more detail. Also, several articles used complex and novel methods, and some editorials discussed the benefits and limitations of these methods. I would also add that this list excludes articles from my own institution.

Mission Statement: JAMA Psychiatry strives to publish original, state-of-the-art studies and commentaries of general interest to clinicians, scholars, and research scientists in psychiatry, mental health, behavioral science, and allied fields. The journal seeks to inform and to educate its readers as well as to stimulate debate and further exploration into the nature, causes, treatment, and public health importance of mental illness.

In the Original Investigation titled “Comparison of the Association Between Goal-Directed Planning and Self-reported Compulsivity vs Obsessive-Compulsive Disorder Diagnosis,” published online October 9 in JAMA Psychiatry,1 errors occurred when 2 graphs in Figure 2 were inadvertently mislabeled. In part B, the part label above the graph and the x-axis label below the graph should have been “Distress”; in part C, the part label above the graph and the x-axis label below the graph should have been “Compulsivity.” This article was corrected online.

In the Original Investigation titled “Prediction of Sex-Specific Suicide Risk Using Machine Learning and Single-Payer Health Care Registry Data From Denmark,”1 published online October 23, 2019, there were errors in the Abstract, Statistical Analysis, and Results. In the Abstract, the following sentence should have been removed: “For example, stress disorders among unmarried men older than 30 years were important factors for suicide risk in the presence of depression (risk, 0.54).” In the second paragraph of the Statistical Analysis, the number of predictor variables given as 2554 should have been 2564. In the first and second paragraphs of the Random Forest section of the Results, predictors enumerated as “Seventeen” and “Nineteen” should have been “Eighteen” and “Twenty-one,” respectively. In the second paragraph of the Results in CART Remodeling, the first 2 sentences were given as “Among men, the highest risk for suicide was found among those younger than 30 years who were diagnosed with schizophrenia in the past 2 years but without recorded prescriptions for antidepressants, antipsychotics, medications for addictions (eg, methadone), or hypnotics or sedatives (n = 26; risk, 0.58). Similarly, unmarried men older than 30 years who had been diagnosed with a stress disorder in the prior 4 years but did not have a recorded prescription for these medications had a risk of 0.54 (n = 37).” The sentences should have appeared as “Among men, the highest risk for suicide was found among those not being treated by pharmacotherapy (eg, antidepressants, antipsychotics, or anxiolytics) and with a prior suicide attempt in the prior 4 years, and being in the second income quartile (n = 18; risk, 1.0). Similarly, men who received a prior diagnosis of poisoning by adverse effects or underdosing of drugs but did not have a coded prescription for antidepressants, antipsychotics, medications for addictions (eg, methadone), or hypnotics/sedatives in the prior 4 years had a risk of 0.42 (n = 251).” This article was corrected online.

In Reply In their Letter, Leas et al, who, to our knowledge, published the first study on possible effects of 13 Reasons Why on suicide-related internet search behavior,1 point out the necessity to use new methodological approaches to improve public health surveillance and responses to emerging public health threats such as 13 Reasons Why. We wholeheartedly agree with the scientific points made in their Letter, although we are a bit puzzled that our reference to their work, including highlighting it in our discussion, however briefly, was viewed as a dismissal of their contribution to evidence on the impact of 13 Reasons Why.

To the Editor We applaud Niederkrotenthaler and colleagues1 for adding another layer of evidence that 13 Reasons Why is harming the public by pushing some individuals toward suicide. However, their dismissal of some of the earliest evidence on this subject deserves a revision not because it undermines their central claim but because it makes it even stronger and can make psychiatric epidemiology more actionable in the future.

In Reply We would like to thank Pennap and colleagues for their thoughtful letter. We completely agree that the use of diagnostic data are likely to greatly underestimate the prevalence of suicide attempts, and we should have stated that explicitly in our article1 as a possible limitation. To emphasize this point, we cite a recent study of postpartum women that demonstrated that detection of suicidal behavior through natural language processing of clinical notes resulted in an 11-fold greater estimate of suicide attempts than in those based solely on diagnostic data.2 However, as Pennap and colleagues note in their Letter, while this underestimate most assuredly affects an estimate of the risk difference, it may not affect the relative risk of suicide attempts. Therefore, we believe that it is safe to conclude that while the diagnostic data reported in our study1 most likely underestimate the true prevalence of suicide attempts and the risk difference, it is probably a reasonable estimate of the relative risk of suicide attempts in the children of parents who use opioids on a regular basis. We also note that our sample was based on parents and children with private health insurance and the rate is likely to be lower than in a general emergency department sample, which includes patients with public health insurance and no health insurance. For example the background rate of acute liver injury based on International Classification of Diseases, Ninth Revision code 570 was 4 times higher in Medicaid claims data relative to claims based on people with private health insurance (ie, the same MarketScan data used in our article).3 Future work that studies suicidal behavior using medical records may benefit from augmenting estimates of the prevalence of suicidal behavior from diagnostic codes with the use of natural language processing to identify cases of suicidal behavior that are documented in the medical record but are not reflected in the diagnostic code.

Sleep disturbance and neuropsychiatric illness have a bidirectional relationship, and while treatment of psychiatric illness often improves sleep, independently addressing sleep disturbance may also lead to better mental health (for a review, see study by Pigeon et al1). Furthermore, sleep complaints, whether they be insomnia, hypersomnia, restless legs syndrome, or nightmares, are often very distressing to the individual experiencing these events. Thus, their identification and treatment may not only improve the severity and course of psychiatric illness, but immediate quality of life as well. Psychiatrists and other mental health professionals are in a good position to evaluate and diagnose sleep disorders as they routinely ask about sleep in the diagnosis of psychiatric illness. Unfortunately, they often have limited training in how to assess these common complaints.

Suicide rates have continued to rise in the United States. Speculations for this rise proliferate but the causes for the increase remain unknown. While research focuses on identifying causes, the health care system is an important site for identification of patients at risk. Forty percent of individuals who die by suicide were seen in primary care in the month prior to suicide. The Zero Suicide model describes a comprehensive approach for health care systems to aid in identification and intervention of suicidal patients. While this model promises to improve care of suicidal patients, the need for innovation in our approach to understanding and caring for suicidal patients is pressing. Use of technology to enhance moment-to-moment monitoring of at-risk individuals offers promise and the possibility of intervening close to escalation of acute suicidal states. Further, once identified, suicidal individuals are often difficult to engage in treatment. Novel approaches to engagement and treatment that are effective and acceptable to suicidal patients ought to be developed. Specifically, males are much more likely than females to die by suicide. At the same time, males are less likely to seek and remain in the treatments we have to offer. Innovation should seek to identify strategies that are acceptable to males. Additionally, while about half of psychiatric inpatient admissions are suicide related, there is a paucity of suicide-specific psychosocial interventions available for inpatient settings. Innovation in monitoring and treatment offer promise in helping to reduce suicidal behavior in the United States.

Population-based health registries can be valuable resources for public health research, particularly for studying rare diseases and long-term health outcomes, where more traditional observational and experimental designs often have insufficient sample size or follow-up time to adequately address the research question(s) of interest. In this issue, Tith et al1 demonstrate the utility of one such health registry for asking questions that are difficult to answer using more traditional study designs. Using hospitalization records for Quebec, Canada, from 2006 to 2018, they found that women who had been hospitalized for bulimia nervosa had higher incidence rates of cardiovascular disease (CVD) and death than women without bulimia who had experienced a pregnancy-related hospitalization. This research highlights the potential strength of registry data for informing interventions to enhance public health. Here, we offer thoughts on how to design and interpret registry-based studies to enhance validity.

Machine learning is on the rise. According to Scopus (www2.scopus.com), the number of publications in medicine with machine learning in the title, abstract, or as a keyword during 2016 to 2018 increased from 1658 to 3904. In psychiatry, applications of machine learning are proposed to improve the accuracy of diagnosis and prognosis and determine treatment choice. At the same time, much of this research has given insufficient attention to high-quality methods, clinical applications, and ethical aspects. This is compounded by poor reporting of performative measures and misleading claims about the high accuracy of such approaches. In this issue of JAMA Psychiatry, the article by Gradus and colleagues1 raises important questions about the place of machine learning in research and practice.

In this issue of JAMA Psychiatry, Hakulinen and colleagues1 assembled a comprehensive longitudinal data set using a range of Danish administrative data to study the link between household income during childhood and the likelihood that children will develop schizophrenia. They offer persuasive evidence of an association between persistent low household income and the likelihood of developing schizophrenia. They also find that upward income mobility at all starting income levels is negatively associated with the risk of schizophrenia. Thus, it is persistent low income that is especially damaging to mental health.

Advances in the last 4 decades have enormously improved the lives of persons with obsessive-compulsive disorder (OCD), a condition once thought to be both rare and untreatable. We now know that OCD is both common and treatable. Most individuals with OCD experience symptom relief after receiving evidence-based psychological and/or pharmacological treatments that were initially developed throughout the 1970s and 1990s (namely, cognitive-behavior therapy [CBT] and serotonin reuptake inhibitors). However, challenges remain.

The Alzheimer community was recently confronted with sobering news: a phase III trial of aducanumab, an anti-amyloid compound that represented the most recent hope for a disease-modifying therapy (DMT), was discontinued when a futility analysis determined that no significant likelihood of therapeutic efficacy existed. This was a disheartening outcome, given promising phase Ib data that showed dose-dependent disease slowing and plaque clearance.1 Thus, the world is still facing an Alzheimer disease (AD) epidemic without any DMTs to combat the staggering financial and human costs.2 Symptomatic treatments (eg, cholinesterase inhibitors) that temporarily slow decline are available, but these do not modify disease progression. Earlier hope of addressing AD neuropathology by reducing buildup or enhancing removal of β-amyloid (Aβ) plaques has been moderated by the failure of more than a dozen anti-amyloid compounds. In this Viewpoint, we explore potential reasons for these failures, describe the response of the AD community, and highlight reasons for continued optimism.

Psychiatry has seen significant advances in our understanding of the neuroscience of development and the evolution of mental disorders in adolescents and younger adults. This allows us to educate our communities, combat stigma, identify individuals at high risk of developing mental disorders, and provide evidence-based interventions. These efforts are critical and may reduce chronicity and recurrence of mental disorders that contribute to accelerated biological aging and persistent disability. This allows psychiatry to re-envision its goal of preventing the development of mental disorders in early life, thereby reducing disease burden and service utilization across the lifespan.1 These recent advances hold tremendous potential benefit for future generations.

Parental socioeconomic position may be linked to offspring’s schizophrenia risk, but parental socioeconomic mobility has not been investigated. Using a Danish national birth cohort of more than 1 million individuals, Hakulinen and colleagues examined parental income measured at birth year and at child ages 5, 10, and 15 years and found that regardless of parental income level at birth, upward income mobility was associated with significantly lower schizophrenia risk compared with downward mobility. In an Editorial, Frank discusses the implications of this work for the field. Editorial Continuing Medical Education

In Reply We thank Mueser and McGurk for the interest in this Danish randomized clinical trial investigating the effects of individual placement and support (IPS) and IPS supplemented with cognitive remediation and social skills training among 720 participants with severe mental illness.1 First, the authors state that the rates of competitive employment are low compared with previous IPS trials. This is correct, but the primary aim in our trial was competitive employment and education, which cannot be directly compared with trials exclusively having employment as a primary aim. We consider that 38% in competitive employment is a success because an additional 31% enrolled in education. It is obvious that the employment rates would have been higher if we had excluded all participants who had an aim of education at baseline. Second, the authors raise a critique of the rates of noncompetitive work. In Denmark and other Scandinavian countries, companies usually recruit people who receive social benefits by offering a period of unpaid internship and then afterward competitive employment. This is a culture that has been promoted by the Danish welfare system for decades, and it was not possible to completely avoid. The same tendency was seen in the Swedish IPS trials.2 However, the employment specialists had a very strict focus on minimizing the duration of internships and in accordance with the IPS principles. Third, the average time to first employment in our trial also includes participants who did not obtain employment. They were all rated for 548 days, which was the number of days from baseline to the 18-month follow-up. The average time to competitive employment among participants who obtained at least 1 job was 210 days for the IPS group and 232 days for the group of IPS supplemented with cognitive remediation and social skills training. Fourth, the choice of changing the computer software from COGPACK to Computerized Interactive Remediation of Cognition—a Training for Schizophrenia (CIRCUITS), Danish version, was based on thorough testing. The conclusion of this testing was that CIRCUITS would be more appealing to the young population in our trial. Because CIRCUITS was newly developed, it obviously had less empirical support than COGPACK. Fifth, it is a misunderstanding that 6 of the 12 sessions in the Thinking Skills for Work program were replaced with social skills training. The 6 sessions were a supplement that took place after the participants had finished Thinking Skills for Work.3 Sixth, it is correct that participants had less individual sessions with the cognitive specialist compared with the original Thinking Skills for Work manual, but it was possible for the participants to receive individualized counseling both from the psychologist and the employment specialists who were cotherapists. In addition, we believe that the trial answered several important questions. We demonstrated that IPS can be implemented and be effective in a Scandinavian welfare model and that support for education can be included and work along with the IPS principles. However, we cannot draw firm conclusions on the benefits of augmenting IPS with cognitive remediation. We can conclude that it should not be implemented in the form used in this trial. Effects may be found among participants who are motivated and who do not respond to IPS after 18 months of intervention, as it was demonstrated in the trial by McGurk et al.4

To the Editor We would like to point out several limitations of the study by Christensen et al1 comparing the effects of the individual placement and support (IPS) model of supported employment with IPS enhanced by cognitive remediation and social skills training (IPSE) and service as usual on work in persons with severe mental illness that preclude drawing firm conclusions about the potential benefits of augmenting IPS with cognitive remediation. Three sets of findings suggest problems implementing the IPS program.

US veterans face severe mental health needs; there have been more than 6000 suicides annually from 2008 to 2016, a 25.9% jump in suicides from 2005 to 2016, a suicide rate 1.5-fold higher than nonveterans, and higher suicide rates for veterans accessing Department of Veterans Affairs (VA) facilities vs those not using VA services, veterans overall, and nonveterans.1 On June 6, 2018, the Veterans Affairs Maintaining Internal Systems and Strengthening Integrated Outside Networks (VA MISSION) Act of 2018 came into law as S 2372.2 The VA MISSION Act passed after former VA Secretary David Shulkin, MD, called suicide prevention his top clinical priority, stating that just 6 of 20 veterans who have committed suicide each day accessed VA care in the past year. In January 2018, President Donald Trump issued an executive order for the VA and the Departments of Defense and Homeland Security to coordinate seamless access to treatment for service members transitioning into veteran status to prevent suicides.

In Reply We thank King and Ehrlich for their thoughtful Letter critiquing our meta-analysis1 that examined delay discounting as a transdiagnostic process in psychiatric disorders. In our article, we proposed that delay discounting, an index of impulsive decision-making, falls on a continuum. At the aggregate level, most disorders examined (eg, major depressive disorder, bipolar disorder, schizophrenia, borderline personality disorder, binge-eating disorder, and bulimia nervosa) exhibited steeper, more impulsive, delay discounting compared with controls, whereas people with anorexia nervosa exhibited shallower, less impulsive, discounting compared with controls. We discussed that excessively shallow discounting in anorexia nervosa is consistent with clinical symptoms of the disorder, including excessive control over food intake. King and Ehrlich raised an important concern about analyzing anorexia nervosa as a unitary disorder instead of examining 2 potentially important factors: (1) clinical subtype (eg, restricting vs binge eating/purging) and (2) association of current weight status (eg, acute underweight vs weight restored) with outcomes. King and Ehrlich suggested that a more specific analysis would reveal that shallow discounting is limited to patients with restricting subtype who have acute underweight.

To the Editor Amlung and colleagues1 published a meta-analysis of delay discounting (DD), a behavioral measure of impulsivity and self-control, including 57 effect sizes from 43 studies across 8 psychiatric diagnostic categories. Results indicated a greater preference for smaller immediate monetary rewards (ie, more impulsive decision-making) in 6 conditions (major depression, schizophrenia, bipolar disorder, borderline personality disorder, bulimia nervosa, and binge-eating disorder) compared with controls but greater preference for larger delayed rewards (ie, more self-controlled choice) in anorexia nervosa. The findings have important implications for the National Institute of Mental Health’s Research Domain Criteria initiative, which seeks to elucidate neurocognitive mechanisms that transcend categorical diagnoses.

In this issue of JAMA Psychiatry, Suokas et al1 reinforce 2 observations emerging from the rich literature on the epidemiology of mental health problems. First, Scandinavian countries with established nationwide registries that permit the linking of diverse databases offer unparalleled opportunities to investigate the temporal associations between risk factors and mental health problems at the level of an individual’s life course and at the level of populations over time. Second, poverty, in this case defined by personal income, is a major determinant for poor mental health, in this case represented by the outcome of first psychiatric admission. In summary, this study shows that (1) the poorer you are in Finland, the higher the likelihood of experiencing a first psychiatric admission; (2) this risk gradient is seen across income strata; and (3) despite overall reductions in admission rates in adults during the 19 years of observation, which is probably a reflection of the changing orientation of psychiatric care, inequalities have persisted and even grown larger across the study period.

The last decade has witnessed a rapid increase in private equity (PE) acquisitions of health care organizations.1 This phenomenon has been described in 3 recent articles.2-4 However, very little is known about PE investment in behavioral health5 despite substantial energy in this sphere. In 2016, PE transactions accounted for 60% of all sales in behavioral health care, and the 24 large behavioral health investments that year far exceeded the 11 such investments, at that time a record, made in 2013.6 This Viewpoint is intended to provide early insights into this phenomenon.

In this issue of JAMA Psychiatry, DeRubeis et al1 present the second phase of an elegantly designed study that has yielded results that will inform patient care and future research. A paucity of studies is available to inform the long-term treatment of persistent major depressive disorder (MDD), either chronic or recurrent.2 Most studies have focused on short-term treatment with single interventions, and many longer-term studies have not been controlled.