Association of early hypotension in pediatric sepsis with development of new or persistent acute kidney injury.

OBJECTIVE: To determine how hypotension in the first 48 h of sepsis management impacts acute kidney injury (AKI) development and persistence. STUDY DESIGN: Retrospective study of patients > 1 month to < 20 years old with sepsis in a pediatric ICU between November 2012 and January 2015 (n = 217). All systolic blood pressure (SBP) data documented within 48 h after sepsis recognition were collected and converted to percentiles for age, sex, and height. Time below SBP percentiles and below pediatric advanced life support (PALS) targets was calculated by summing elapsed time under SBP thresholds during the first 48 h. The primary outcome was new or persistent AKI, defined as stage 2 or 3 AKI present between sepsis day 3-7 using Kidney Disease: Improving Global Outcomes creatinine definitions. Secondary outcomes included AKI-free days (days alive and free of AKI) and time to kidney recovery. RESULTS: Fifty of 217 sepsis patients (23%) had new or persistent AKI. Patients with AKI spent a median of 35 min under the first SBP percentile, versus 4 min in those without AKI. After adjustment for potential confounders, the odds of AKI increased by 9% with each doubling of minutes spent under this threshold (p = 0.03). Time under the first SBP percentile was also associated with fewer AKI-free days (p = 0.02). Time spent under PALS targets was not associated with AKI. CONCLUSIONS: The duration of severe systolic hypotension in the first 48 h of pediatric sepsis management is associated with AKI incidence and duration when defined by age, sex, and height norms, but not by PALS definitions. Graphical abstract.

Development of a Clinical Prediction Model for Central Line-Associated Bloodstream Infection in Children Presenting to the Emergency Department.

OBJECTIVE: The majority of the children with a central line who present to the emergency department with fever or other signs of bacteremia do not have a central line-associated bloodstream infection (CLABSI). Our objective was to develop a clinical prediction model for CLABSI among this group of children in order to ultimately limit unnecessary hospital admissions and antibiotic use. METHODS: We performed a nested case-control study of children with a central line who presented to the emergency department of an urban, tertiary care children's hospital between January 2010 and March 2015 and were evaluated for CLABSI with a blood culture. RESULTS: The final multivariable model developed to predict CLABSI consisted of 12 factors: age younger than 5 years, black race, use of total parenteral nutrition, tunneled central venous catheter, double-lumen catheter, absence of other bacterial infection, absence of viral upper respiratory tract infection symptoms, diarrhea, emergency department temperature greater than 39.5°C, fever prior to presentation, neutropenia, and spring/summer season. The clinical prediction score had good discrimination for CLABSI with a c-statistic of 0.81 (confidence interval, 0.77-0.85). A cut point less than 6 was associated with a sensitivity of 98.5% and a negative predictive value of 99.2% for CLABSI. CONCLUSIONS: We were able to identify risk factors and develop a clinical prediction model for CLABSI in children presenting to the emergency department. Once validated in future study, this clinical prediction model could be used to assess the need for hospitalization and/or antibiotics among this group of patients.

Risk maps for cities: Incorporating streets into geostatistical models.

Vector-borne diseases commonly emerge in urban landscapes, and Gaussian field models can be used to create risk maps of vector presence across a large environment. However, these models do not account for the possibility that streets function as permeable barriers for insect vectors. We describe a methodology to transform spatial point data to incorporate permeable barriers, by distorting the map to widen streets, with one additional parameter. We use Gaussian field models to estimate this additional parameter, and develop risk maps incorporating streets as permeable barriers. We demonstrate our method on simulated datasets and apply it to data on Triatoma infestans, a vector of Chagas disease in Arequipa, Peru. We found that the transformed landscape that best fit the observed pattern of Triatoma infestans infestation, approximately doubled the true Euclidean distance between neighboring houses on different city blocks. Our findings may better guide control of re-emergent insect populations.

Risk factors and inpatient outcomes associated with acute kidney injury at pediatric severe sepsis presentation.

BACKGROUND: Little data exist on acute kidney injury (AKI) risk factors in pediatric sepsis. We identified risk factors and inpatient outcomes associated with AKI at sepsis recognition in children with severe sepsis. METHODS: Retrospective, cross-sectional study with inpatient outcome description of 315 patients > 1 month to < 20 years old with severe sepsis in a pediatric intensive care unit over 3 years. Exposures included demographics, vitals, and laboratory data. The primary outcome was kidney disease: Improving Global Outcomes creatinine-defined AKI within 24 h of sepsis recognition. Factors associated with AKI and AKI severity were identified using multivariable Poisson and multinomial logistic regression, respectively. RESULTS: AKI was present in 42% (133/315) of severe sepsis patients, and 26% (83/315) had severe (stage 2/3) AKI. In multivariable-adjusted analysis, hematologic/immunologic comorbidities, malignancies, chronic kidney disease (CKD), abdominal infection, admission illness severity, and minimum systolic blood pressure (SBP) ≤ 5th percentile for age and sex within 24 h of sepsis recognition were associated with AKI. Factors associated with mild AKI were CKD and abdominal infection, while factors associated with severe AKI were younger age, hematologic/immunologic comorbidities, malignancy, abdominal infection, and minimum SBP ≤ 5th percentile. Patients with AKI had increased hospital mortality (17 vs. 8%, P = 0.02) and length of stay [median 20 (IQR 10-47) vs. 16 days (IQR 7-37), P = 0.03]. CONCLUSIONS: In pediatric severe sepsis, AKI is associated with age, comorbidities, infection characteristics, and hypotension. Future evaluation of risk factors for AKI progression during sepsis is warranted to minimize AKI progression in this high-risk population.

Electronic health record (EHR) based postmarketing surveillance of adverse events associated with pediatric off-label medication use: A case study of short-acting beta-2 agonists and arrhythmias.

PURPOSE: Use electronic health record (EHR) data to (1) estimate the risk of arrhythmia associated with inhaled short-acting beta-2 agonists (SABA) in pediatric patients and (2) determine whether risk varied by on-label versus off-label prescribing. METHODS: Retrospective cohort study of 335 041 children ≤18 years using EHR primary care data from 2 pediatric health systems (2011-2013). A series of monthly pseudotrials were created, using propensity score methodology to balance baseline characteristics between SABA-exposed (identified by prescription) and SABA-unexposed children. Association between SABA and subsequent arrhythmia for each health system was estimated through pooled logistic regression with separate estimates for children initiating under and over 4 years old (off-label and on-label, respectively). RESULTS: Eleven percent of the cohort received a SABA prescription, 57% occurred under the age of 4 years (off-label). During the follow-up period, there were 283 first arrhythmia events, most commonly atrial tachyarrhythmias and premature ventricular/atrial contractions. In 1 health system, adjusted risk for arrhythmia was increased among exposed children (OR 1.89, 95% CI 1.31-2.73) without evidence of interaction between label status and risk. The absolute adjusted rate difference was 3.6/10 000 person-years of SABA exposure. The association between SABA exposure and arrhythmias was less strong in the second system (OR 1.26, 95% CI 0.30-5.33). CONCLUSION: Using EHR data, we could estimate the risk of a rare event associated with medication use and determine difference in risk related to on-label versus off-label status. These findings support the value of EHR-based data for postmarketing drug studies in the pediatric population.

Urology Consultation and Emergency Department Revisits for Children with Urinary Stone Disease.

PURPOSE: We determined the association between urology consultation and emergency department revisits for children with urinary stones. MATERIALS AND METHODS: This retrospective cohort study included patients 18 years old or younger who presented to an emergency department in South Carolina with a urinary stone from 1997 to 2015. The primary exposure was urology consultation during the index emergency department visit. The primary outcome was a stone related emergency department revisit occurring within 180 days of discharge from an index emergency department visit. Secondary outcomes included computerized tomography use, inpatient admission and emergent surgery. RESULTS: Of 5,642 index emergency department visits for acute urinary stones 11% resulted in at least 1 stone related emergency department revisit within 180 days. Of revisits 59% occurred within 30 days of discharge and 39% were due to pain. The odds of emergency department revisit were highest within the first 48 hours of discharge home (OR 22.6, 95% CI 18.0 to 28.5) and rapidly decreased thereafter. Urology consultation was associated with a 37% lower adjusted odds of emergency department revisit (OR 0.63, 95% CI 0.44 to 0.90) and 68% lower odds of computerized tomography use across all emergency department visits (OR 0.32, 95% CI 0.15 to 0.69). Among patients who revisited the emergency department the frequency of pain complaints was 27% in those with a urological consultation at the index visit and 39% in those without. CONCLUSIONS: Urology consultation was associated with decreased emergency department revisits and computerized tomography use in pediatric patients with urinary stones. Future studies should identify patients who benefit most from urology consultation and ascertain processes of care that decrease emergency department revisits among high risk patients.

Assessment of the combination of temperature and relative humidity on kidney stone presentations.

Temperature and relative humidity have opposing effects on evaporative water loss, the likely mediator of the temperature-dependence of nephrolithiasis. However, prior studies considered only dry-bulb temperatures when estimating the temperature-dependence of nephrolithiasis. We used distributed lag non-linear models and repeated 10-fold cross-validation to determine the daily temperature metric and corresponding adjustment for relative humidity that most accurately predicted kidney stone presentations during hot and cold periods in South Carolina from 1997 to 2015. We examined three metrics for wet-bulb temperatures and heat index, both of which measure the combination of temperature and humidity, and for dry-bulb temperatures: (1) daytime mean temperature; (2) 24-h mean temperature; and (3) most extreme 24-h temperature. For models using dry-bulb temperatures, we considered four treatments of relative humidity. Among 188,531 patients who presented with kidney stones, 24-h wet bulb temperature best predicted kidney stone presentation during summer. Mean cross-validated residuals were generally lower in summer for wet-bulb temperatures and heat index than the corresponding dry-bulb temperature metric, regardless of type of adjustment for relative humidity. Those dry-bulb models that additionally adjusted for relative humidity had higher mean residuals than other temperature metrics. The relative risk of kidney stone presentations at the 99th percentile of each temperature metric compared to the respective median temperature in summer months differed by temperature metric and relative humidity adjustment, and ranged from an excess risk of 8-14%. All metrics performed similarly in winter. The combination of temperature and relative humidity determine the risk of kidney stone presentations, particularly during periods of high heat and humidity. These results suggest that metrics that measure moist heat stress should be used to estimate the temperature-dependence of kidney stone presentations, but that the particular metric is relatively unimportant.

Incidence of Pneumocystis jirovecii and Adverse Events Associated With Pneumocystis Prophylaxis in Children Receiving Glucocorticoids.

BACKGROUND: Antimicrobial prophylaxis is indicated to prevent Pneumocystis jirovecii pneumonia (PJP) in profoundly immunosuppressed children. The incidence of PJP infection in children with chronic glucocorticoid exposure is unknown, and PJP prophylaxis has been associated with adverse events. We hypothesized that PJP infection is rare in children without human immunodeficiency virus (HIV)/acquired immunodeficiency syndrome (AIDS), cancer, or a transplant history who are using chronic glucocorticoids and that those exposed to PJP prophylaxis are more likely to experience a cutaneous hypersensitivity reaction or myelosuppression than unexposed patients. METHODS: This study involved a retrospective cohort from the Clinformatics Data Mart Database (OptumInsight, Eden Prairie, MN). We identified patients ≤18 years of age who received at least 2 prescriptions for a systemic glucocorticoid within a 60-day period and excluded patients with a history of PJP infection, an oncologic diagnosis, transplant, or HIV/AIDS. PJP prophylaxis exposure was identified by using national drug codes. Cutaneous hypersensitivity reaction or myelosuppression was identified by using International Classification of Diseases, 9th Revision (ICD-9), codes. We used a discrete time-failure model to examine the association between exposure and outcome. RESULTS: We identified 119399 children on glucocorticoids, 10% of whom received PJP prophylaxis. The incidences of PJP were 0.61 and 0.53 per 10000 patient-years in children exposed and those unexposed to PJP prophylaxis, respectively. In a multivariable model, trimethoprim-sulfamethoxazole was associated with cutaneous hypersensitivity reaction (odds ratio, 3.20; 95% confidence interval, 2.62-3.92) and myelosuppression (odds ratio, 1.85; 95% confidence interval, 1.56-2.20). CONCLUSIONS: PJP infection was rare in children using glucocorticoids chronically, and PJP prophylaxis-associated cutaneous hypersensitivity reactions and myelosuppression are more common. The use of PJP chemoprophylaxis in children without HIV/AIDS, cancer, or a transplant history who are taking glucocorticoids chronically should be considered carefully.

Automated identification of implausible values in growth data from pediatric electronic health records.

OBJECTIVE: Large electronic health record (EHR) datasets are increasingly used to facilitate research on growth, but measurement and recording errors can lead to biased results. We developed and tested an automated method for identifying implausible values in pediatric EHR growth data. MATERIALS AND METHODS: Using deidentified data from 46 primary care sites, we developed an algorithm to identify weight and height values that should be excluded from analysis, including implausible values and values that were recorded repeatedly without remeasurement. The foundation of the algorithm is a comparison of each measurement, expressed as a standard deviation score, with a weighted moving average of a child's other measurements. We evaluated the performance of the algorithm by (1) comparing its results with the judgment of physician reviewers for a stratified random selection of 400 measurements and (2) evaluating its accuracy in a dataset with simulated errors. RESULTS: Of 2 000 595 growth measurements from 280 610 patients 1 to 21 years old, 3.8% of weight and 4.5% of height values were identified as implausible or excluded for other reasons. The proportion excluded varied widely by primary care site. The automated method had a sensitivity of 97% (95% confidence interval [CI], 94-99%) and a specificity of 90% (95% CI, 85-94%) for identifying implausible values compared to physician judgment, and identified 95% (weight) and 98% (height) of simulated errors. DISCUSSION AND CONCLUSION: This automated, flexible, and validated method for preparing large datasets will facilitate the use of pediatric EHR growth datasets for research.

Dietary Zinc and Incident Calcium Kidney Stones in Adolescence.

PURPOSE: We determined the association between dietary zinc intake and incident calcium kidney stones in adolescents. We also examined the relationship between dietary zinc intake and urinary zinc excretion between cases and controls. MATERIALS AND METHODS: We conducted a nested case-control study within a large pediatric health care system. Three 24-hour dietary recalls and spot urine chemistry analyses were obtained for 30 participants 12 to 18 years old with a first idiopathic calcium based kidney stone and 30 healthy controls matched for age, sex, race and month of enrollment. Conditional logistic regression models were used to estimate the association between daily zinc intake and incident calcium kidney stones, adjusting for dietary phytate, protein, calcium, sodium and oxalate. Multivariable linear regression was used to estimate the association between dietary and urine zinc, adjusting for urine creatinine and dietary phytate and calcium. RESULTS: Cases had lower daily zinc intake (8.1 mg) than controls (10 mg, p = 0.029). Daily zinc intake of boys and girls with calcium stones was 2 mg and 1.2 mg less, respectively, than the daily intake recommended by the Institute of Medicine. Odds of incident stones were reduced by 13% for every 1 mg increase in daily zinc intake (OR 0.87, 95% CI 0.75-0.99). There was an estimated 4.5 µg/dl increase in urine zinc for every 1 mg increase in dietary zinc (p = 0.009), with weak evidence of a smaller increase in urine zinc in cases than in controls (interaction p = 0.08). CONCLUSIONS: Decreased dietary zinc intake was independently associated with incident calcium nephrolithiasis in this population of adolescents.

Preschool ADHD Diagnosis and Stimulant Use Before and After the 2011 AAP Practice Guideline.

OBJECTIVE: To evaluate the change in the diagnosis of attention-deficit/hyperactivity disorder (ADHD) and prescribing of stimulants to children 4 to 5 years old after release of the 2011 American Academy of Pediatrics guideline. METHODS: Electronic health record data were extracted from 63 primary care practices. We included preventive visits from children 48 to 72 months old receiving care from January 2008 to July 2014. We compared rates of ADHD diagnosis and stimulant prescribing before and after guideline release using logistic regression with a spline and clustering by practice. Patterns of change (increase, decrease, no change) were described for each practice. RESULTS: Among 87 067 children with 118 957 visits before the guideline and 56 814 with 92 601 visits after the guideline, children had an ADHD diagnosis at 0.7% (95% confidence interval [CI], 0.7% to 0.8%) of visits before and 0.9% (95% CI, 0.8% to 0.9%) after guideline release and had stimulant prescriptions at 0.4% (95% CI, 0.4% to 0.4%) of visits in both periods. A significantly increasing preguideline trend in ADHD diagnosis ended after guideline release. The rate of stimulant medication use remained constant before and after guideline release. Patterns of change from before to after the guideline varied significantly across practices. CONCLUSIONS: Release of the 2011 guideline that addressed ADHD in preschoolers was associated with the end of an increasing rate of diagnosis, and the rate of prescribing stimulants remained constant. These are reassuring results given that a standardized approach to diagnosis was recommended and stimulant treatment is not first-line therapy for this age group.

Using a Mystery-Caller Approach to Examine Access to Prostate Cancer Care in Philadelphia.

PURPOSE: Prior work suggests that access to health care may influence the diagnosis and treatment of prostate cancer. Mystery-caller methods have been used previously to measure access to care for health services such as primary care, where patients' self-initiate requests for care. We used a mystery-caller survey for specialized prostate cancer care to assess dimensions of access to prostate cancer care. MATERIALS AND METHODS: We created an inventory of urology and radiation oncology practices in southeastern Pennsylvania. Using a 'mystery caller' approach, a research assistant posing as a medical office scheduler in a primary care office, attempted to make a new patient appointment on behalf of a referred patient. Linear regression was used to determine the association between time to next available appointment with practice and census tract characteristics. RESULTS: We successfully obtained information on new patient appointments from 198 practices out of the 223 in the region (88.8%). Radiation oncology practices were more likely to accept Medicaid compared to urology practices (91.3% vs 36.4%) and had shorter mean wait times for new patient appointments (9.0 vs 12.8 days). We did not observe significant differences in wait times according to census tract characteristics including neighborhood socioeconomic status and the proportion of male African American residents. CONCLUSIONS: Mystery-caller methods that reflect real-world referral processes from primary care offices can be used to measure access to specialized cancer care. We observed significant differences in wait times and insurance acceptance between radiation oncology and urology practices.

Adoption of a Portal for the Primary Care Management of Pediatric Asthma: A Mixed-Methods Implementation Study.

BACKGROUND: Patient portals may improve communication between families of children with asthma and their primary care providers and improve outcomes. However, the feasibility of using portals to collect patient-reported outcomes from families and the barriers and facilitators of portal implementation across diverse pediatric primary care settings have not been established. OBJECTIVE: We evaluated the feasibility of using a patient portal for pediatric asthma in primary care, its impact on management, and barriers and facilitators of implementation success. METHODS: We conducted a mixed-methods implementation study in 20 practices (11 states). Using the portal, parents of children with asthma aged 6-12 years completed monthly surveys to communicate treatment concerns, treatment goals, symptom control, medication use, and side effects. We used logistic regression to evaluate the association of portal use with child characteristics and changes to asthma management. Ten clinician focus groups and 22 semistructured parent interviews explored barriers and facilitators of use in the context of an evidence-based implementation framework. RESULTS: We invited 9133 families to enroll and 237 (2.59%) used the portal (range by practice, 0.6%-13.6%). Children of parents or guardians who used the portal were significantly more likely than nonusers to be aged 6-9 years (vs 10-12, P=.02), have mild or moderate/severe persistent asthma (P=.009 and P=.04), have a prescription of a controller medication (P<.001), and have private insurance (P=.002). Portal users with uncontrolled asthma had significantly more medication changes and primary care asthma visits after using the portal relative to the year earlier (increases of 14% and 16%, respectively). Qualitative results revealed the importance of practice organization (coordinated workflows) as well as family (asthma severity) and innovation (facilitated communication and ease of use) characteristics for implementation success. CONCLUSIONS: Although use was associated with higher treatment engagement, our results suggest that achieving widespread portal adoption is unlikely in the short term. Implementation efforts should include workflow redesign and prioritize enrollment of symptomatic children. CLINICALTRIAL: Clinicaltrials.gov NCT01966068; https://clinicaltrials.gov/ct2/show/NCT01966068 (Archived by WebCite at http://www.webcitation.org/6i9iSQkm3).

Variations in Mental Health Diagnosis and Prescribing Across Pediatric Primary Care Practices.

BACKGROUND: Primary care pediatricians increasingly care for children's mental health problems, but little is known about practice-level variation in diagnosis and psychotropic medication prescribing practices. METHODS: This retrospective review of electronic heath records from 43 US primary care practices included children aged 4 to 18 years with ≥1 office visit from January 1, 2009, to June 30, 2014. We examined variability in diagnosis and psychotropic prescribing across practices using logistic regression with practice fixed effects and evaluated associations of the availability of colocated or community-based mental health providers or the proportion of children in foster care with diagnosis and prescribing using generalized linear mixed models. RESULTS: Among 294 748 children, 40 932 (15%) received a mental health diagnosis and 39 695 (14%) were prescribed psychotropic medication. Attention deficit/hyperactivity disorder was most commonly diagnosed (1%-16% per practice). The proportion of children receiving any psychotropic medication (4%-26%) and the proportion receiving ≥2 medication classes (1%-12%) varied across practices. Prescribing of specific medication classes also varied (stimulants, 3%-18%; antidepressants, 1%-12%; α-agonists, 0%-8%; second-generation antipsychotics, 0%-5%). Variability was partially explained by community availability of psychiatrists (significantly higher odds of a diagnosis or prescription when not available) but not by colocation of mental health professionals or percentage of children in foster care. CONCLUSIONS: The prevalence of mental health diagnosis and psychotropic medication prescribing varies substantially across practices and is only partially explained by psychiatrist availability. Research is needed to better define the causes of variable practice-level diagnosis and prescribing and implications for child mental health outcomes.

Annual Incidence of Nephrolithiasis among Children and Adults in South Carolina from 1997 to 2012.

BACKGROUND AND OBJECTIVES: The prevalence of nephrolithiasis in the United States has increased substantially, but recent changes in incidence with respect to age, sex, and race are not well characterized. This study examined temporal trends in the annual incidence and cumulative risk of nephrolithiasis among children and adults living in South Carolina over a 16-year period. DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS: We performed a population-based, repeated cross-sectional study using the US Census and South Carolina Medical Encounter data, which capture all emergency department visits, surgeries, and admissions in the state. The annual incidence of nephrolithiasis in South Carolina from 1997 to 2012 was estimated, and linear mixed models were used to estimate incidence rate ratios for age, sex, and racial groups. The cumulative risk of nephrolithiasis during childhood and over the lifetime was estimated for males and females in 1997 and 2012. RESULTS: Among an at-risk population of 4,625,364 people, 152,925 unique patients received emergency, inpatient, or surgical care for nephrolithiasis. Between 1997 and 2012, the mean annual incidence of nephrolithiasis increased 1% annually from 206 to 239 per 100,000 persons. Among age groups, the greatest increase was observed among 15-19 year olds, in whom incidence increased 26% per 5 years (incidence rate ratio, 1.26; 95% confidence interval, 1.22 to 1.29). Adjusting for age and race, incidence increased 15% per 5 years among females (incidence rate ratio, 1.15; 95% confidence interval, 1.14 to 1.16) but remained stable for males. The incidence among blacks increased 15% more per 5 years compared with whites (incidence rate ratio, 1.15; 95% confidence interval, 1.14 to 1.17). These changes in incidence resulted in doubling of the risk of nephrolithiasis during childhood and a 45% increase in the lifetime risk of nephrolithiasis for women over the study period. CONCLUSIONS: The incidence of kidney stones has increased among young patients, particularly women, and blacks.

Differential Child Maltreatment Risk Across Deployment Periods of US Army Soldiers.

OBJECTIVES: We described the risk for maltreatment among toddlers of US Army soldiers over different deployment cycles to develop a systematic response within the US Army to provide families appropriate supports. METHODS: We conducted a person-time analysis of substantiated maltreatment reports and medical diagnoses among children of 112,325 deployed US Army soldiers between 2001 and 2007. RESULTS: Risk of maltreatment was elevated after deployment for children of soldiers deployed once but not for children of soldiers deployed twice. During the 6 months after deployment, children of soldiers deployed once had 4.43 substantiated maltreatment reports and 4.96 medical diagnoses per 10,000 child-months. The highest maltreatment rate among children of soldiers deployed twice occurred during the second deployment for substantiated maltreatment (4.83 episodes per 10,000 child-months) and before the first deployment for medical diagnoses of maltreatment (3.78 episodes per 10,000 child-months). CONCLUSIONS: We confirmed an elevated risk for child maltreatment during deployment but also found a previously unidentified high-risk period during the 6 months following deployment, indicating elevated stress within families of deployed and returning soldiers. These findings can inform efforts by the military to initiate and standardize support and preparation to families during periods of elevated risk.

Propensity Score Methods for Analyzing Observational Data Like Randomized Experiments: Challenges and Solutions for Rare Outcomes and Exposures.

Randomized controlled trials are the "gold standard" for estimating the causal effects of treatments. However, it is often not feasible to conduct such a trial because of ethical concerns or budgetary constraints. We expand upon an approach to the analysis of observational data sets that mimics a sequence of randomized studies by implementing propensity score models within each trial to achieve covariate balance, using weighting and matching. The methods are illustrated using data from a safety study of the relationship between second-generation antipsychotics and type 2 diabetes (outcome) in Medicaid-insured children aged 10-18 years across the United States from 2003 to 2007. Challenges in this data set include a rare outcome, a rare exposure, substantial and important differences between exposure groups, and a very large sample size.

Changing patterns of alpha agonist medication use in children and adolescents 2009-2011.

OBJECTIVES: The purpose of this study was to describe rates and patterns of long- and short-acting alpha agonist use for behavioral problems in a primary care population following Food and Drug Administration (FDA) approval of the long-acting alpha agonists guanfacine and clonidine. METHODS: Children and adolescents 4-18 years of age, who received an alpha agonist prescription between 2009 and 2011, were identified from a sample of 45 United States primary care practices in two electronic health record-based research networks. Alpha agonist receipt was identified using National Drug Codes and medication names. The proportion of subjects receiving long- and short-acting prescriptions in each year was calculated and examined with respect to reported mental health diagnoses, and whether indications for use were on-label, had evidence from clinical trials, or had no trial evidence. RESULTS: In a cohort of 282,875 subjects, 27,671 (10%) received any psychotropic medication and only 4,227 subjects (1.5%) received at least one prescription for an alpha agonist, most commonly a short-acting formulation (83%). Only 20% of alpha agonist use was on-label (use of long-acting formulations for attention-deficit/hyperactivity disorder [ADHD]). Most subjects (68%) received alpha agonists for indications with evidence of efficacy from clinical trials but no FDA approval, primarily short-acting formulations for ADHD and autism; 12% received alpha agonists for diagnoses lacking randomized clinical trial evidence in children, including sleep disorders and anxiety, or for which there was no documented mental health diagnosis. Rates of long-acting alpha agonist use increased more than 20-fold from 0.2% to 4%, whereas rates of short-acting alpha agonist use grew only slightly between 2009 and 2011 from 10.6% to 11.3%. CONCLUSIONS: Alpha agonist use was uncommon in this population, and most subjects received short-acting forms for conditions that were off-label, but with clinical trial evidence. The safety and efficacy of use for conditions, including sleep disorders and anxiety, lacking evidence from randomized trials, warrant further investigation.

Risk for incident diabetes mellitus following initiation of second-generation antipsychotics among Medicaid-enrolled youths.

IMPORTANCE: Second-generation antipsychotics (SGAs) have increasingly been prescribed to Medicaid-enrolled children, either singly or in a medication combination. Although metabolic adverse effects have been linked to SGA use in youths, estimating the risk for type 2 diabetes mellitus, a rarer outcome, has been challenging. OBJECTIVE: To determine whether SGA initiation was associated with an increased risk for incident type 2 diabetes mellitus. Secondary analyses examined the risk associated with multiple-drug regimens, including stimulants and antidepressants, as well as individual SGAs. DESIGN, SETTING, AND PARTICIPANTS: Retrospective national cohort study of Medicaid-enrolled youths between January 2003 and December 2007. In this observational study using national Medicaid Analytic eXtract data files, initiators and noninitiators of SGAs were identified in each month. Included in this study were US youths aged 10 to 18 years with a mental health diagnosis and enrolled in a Medicaid fee-for-service arrangement during the study. Those with chronic steroid exposure, a diagnosis of diabetes mellitus, or SGA use during a 1-year look-back period were ineligible. The mean follow-up time for all participants was 17.2 months. Youths were followed up until diagnosis of diabetes mellitus or end of follow-up owing to censoring caused by the transition into a Medicaid managed care arrangement or Medicaid ineligibility (the end of available data). Propensity weights were developed to balance observed demographic and clinical characteristics between exposure groups. Discrete failure time models were fitted using weighted logistic regression to estimate the risk for incident diabetes mellitus between initiators and noninitiators. EXPOSURE: A filled SGA prescription. MAIN OUTCOMES AND MEASURES: Incident type 2 diabetes mellitus identified through visit and pharmacy claims during the observation period. RESULTS: Among 107,551 SGA initiators and 1,221,434 noninitiators, the risk for incident diabetes mellitus was increased among initiators (odds ratio [OR], 1.51; 95% CI, 1.35-1.69; P < .001). Compared with youths initiating only SGAs, the risk was higher among SGA initiators who used antidepressants concomitantly at the time of SGA initiation (OR, 1.54; 95% CI, 1.17-2.03; P = .002) but was not significantly different for SGA initiators who were concomitantly using stimulants. As compared with a reference group of risperidone initiators, the risk was higher among those initiating ziprasidone (OR, 1.61; 95% CI, 0.99-2.64; P = .06) and aripiprazole (OR, 1.58; 95% CI, 1.21-2.07; P = .001) but not quetiapine fumarate or olanzapine. CONCLUSIONS AND RELEVANCE: The risk for incident type 2 diabetes mellitus was increased among youths initiating SGAs and was highest in those concomitantly using antidepressants. Compared with risperidone, newer antipsychotics were not associated with decreased risk.