Sources of selection bias in evaluating social programs: An interpretation of conventional measures and evidence on the effectiveness of matching as a program evaluation method

This paper decomposes the conventional measure of selection bias in observational studies into three components. The first two components are due to differences in the distributions of characteristics between participant and nonparticipant (comparison) group members: the first arises from differences in the supports, and the second from differences in densities over the region of common support. The third component arises from selection bias precisely defined. Using data from a recent social experiment, we find that the component due to selection bias, precisely defined, is smaller than the first two components. However, selection bias still represents a substantial fraction of the experimental impact estimate. The empirical performance of matching methods of program evaluation is also examined. We find that matching based on the propensity score eliminates some but not all of the measured selection bias, with the remaining bias still a substantial fraction of the estimated impact. We find that the support of the distribution of propensity scores for the comparison group is typically only a small portion of the support for the participant group. For values outside the common support, it is impossible to reliably estimate the effect of program participation using matching methods. If the impact of participation depends on the propensity score, as we find in our data, the failure of the common support condition severely limits matching compared with random assignment as an evaluation estimator.

Diferencias regionales en las estructuras de ingresos laborales en Colombia

Dada la persistencia de las diferencias en ingresos laborales por regiones en Colombia, el presente artículo propone cuantificar la magnitud de este diferencial que es atribuida a la diferencia en estructuras de mercado laboral, entendiendo esta última como la diferencia en los retornos a las características de la fuerza laboral. Para ello se propone el uso de un método de descomposición del tipo Oaxaca- Blinder y se compara a Bogotá –la ciudad con mayores ingresos laborales- con otras ciudades principales. Los resultados obtenidos al conducir el ejercicio de descomposición muestran que las diferencias en estructura están a favor de Bogotá y que estas explican más de la mitad de la diferencia total, indicando que si se quieren reducir las disparidades de ingresos laborales entre ciudades no es suficiente con calificar la fuerza laboral y que es necesario indagar por las causas que hacen que los retornos a las características difieran entre ciudades.

Continuous Infusion of Piperacillin/Tazobactam in Septic Critically Ill Patients - a Multicenter Propensity Matched Analysis

The clinical efficacy of continuous infusion of piperacillin/tazobactam in critically ill patients with microbiologically documented infections is currently unknown. We conducted a retrospective multicenter cohort study in 7 Portuguese intensive care units (ICU). We included 569 critically ill adult patients with a documented infection and treated with piperacillin/tazobactam admitted to one of the participating ICU between 2006 and 2010. We successfully matched 173 pairs of patients according to whether they received continuous or conventional intermittent dosing of piperacillin/tazobactam, using a propensity score to adjust for confounding variables. The majority of patients received 16g/day of piperacillin plus 2g/day of tazobactam. The 28-day mortality rate was 28.3% in both groups (p = 1.0). The ICU and in-hospital mortality were also similar either in those receiving continuous infusion or intermittent dosing (23.7% vs. 20.2%, p = 0.512 and 41.6% vs. 40.5%, p = 0.913, respectively). In the subgroup of patients with a Simplified Acute Physiology Score (SAPS) II>42, the 28-day mortality rate was lower in the continuous infusion group (31.4% vs. 35.2%) although not reaching significance (p = 0.66). We concluded that the clinical efficacy of piperacillin/tazobactam in this heterogeneous group of critically ill patients infected with susceptible bacteria was independent of its mode of administration, either continuous infusion or intermittent dosing.

Prognostic Value of TIMI Score versus GRACE Score in ST-segment Elevation Myocardial Infarction

Background: The TIMI Score for ST-segment elevation myocardial infarction (STEMI) was created and validated specifically for this clinical scenario, while the GRACE score is generic to any type of acute coronary syndrome. Objective: Between TIMI and GRACE scores, identify the one of better prognostic performance in patients with STEMI. Methods: We included 152 individuals consecutively admitted for STEMI. The TIMI and GRACE scores were tested for their discriminatory ability (C-statistics) and calibration (Hosmer-Lemeshow) in relation to hospital death. Results: The TIMI score showed equal distribution of patients in the ranges of low, intermediate and high risk (39 %, 27 % and 34 %, respectively), as opposed to the GRACE Score that showed predominant distribution at low risk (80 %, 13 % and 7%, respectively). Case-fatality was 11%. The C-statistics of the TIMI score was 0.87 (95%CI = 0.76 to 0.98), similar to GRACE (0.87, 95%CI = 0.75 to 0.99) - p = 0.71. The TIMI score showed satisfactory calibration represented by χ2 = 1.4 (p = 0.92), well above the calibration of the GRACE score, which showed χ2 = 14 (p = 0.08). This calibration is reflected in the expected incidence ranges for low, intermediate and high risk, according to the TIMI score (0 %, 4.9 % and 25 %, respectively), differently to GRACE (2.4%, 25% and 73%), which featured middle range incidence inappropriately. Conclusion: Although the scores show similar discriminatory capacity for hospital death, the TIMI score had better calibration than GRACE. These findings need to be validated populations of different risk profiles.

Relationship between time of day and periprocedural myocardial infarction after elective angioplasty.

Objectives: To test if the time of day significantly influences the occurrence of type 4A myocardial infarction in elective patients undergoing percutaneous coronary intervention (PCI). Background: Recent studies have suggested an influence of circadian rhythms on myocardial infarction size and mortality among patients with ST-elevation myocardial infarction. The aim of the study is to investigate whether periprocedural myocardial infarction (PMI) is influenced by the time of day in elective patients undergoing PCI. Methods: All consecutive patients undergoing elective PCI between 2007 and 2011 at our institutions with known post-interventional troponin were retrospectively included. Patients (n = 1021) were divided into two groups according to the starting time of the PCI: the morning group (n = 651) between 07:00 and 11:59, and the afternoon group (n = 370) between 12:00 and 18:59. Baseline and procedural characteristics as well as clinical outcome defined as the occurrence of PMI were compared between groups. In order to limit selection bias, all analyses were equally performed in 308 pairs using propensity score (PS) matching. Results: In the overall population, the rate of PMI was statistically lower in the morning group compared to the afternoon group (20% vs. 30%, p < 0.001). This difference remained statistically significant after PS-matching (21% vs. 29%, p = 0.03). Multivariate analysis shows that being treated in the afternoon independently increases the risk for PMI with an odds ratio of 2.0 (95%CI: 1.1-3.4; p = 0.02). Conclusions: This observational PS-matched study suggests that the timing of an elective PCI influences the rate of PMI.

Effects of credit rating change on risk-taking

This thesis investigates whether there are changes in risk-taking behavior following an upgrade or downgrade in credit ratings. Research on effects of rating changes on capital markets is well-documented but the literature on how rating changes may affect firm behavior is sparse. Following, a downgrade in credit rating, managers may increase risk-taking to improve their overall performance or reduce risk-taking following upgrades to ensure that their performance is assessed more on the basis of what they may deem success in the form of an upgrade. Using a sample of firms trading in the U.S from 1994-2013, we find evidence of change in risk-taking behavior. We use cross-sectional regressions and matching using propensity scores and Barber and Lyon (1997) methodology to measure changes in risk-taking and we do find evidence of changes in managerial risk-taking behavior. Furthermore, we find that the direction of change (increase or decrease) in some cases is dependent on the type of measure rather than the type of rating change.

O impacto do trabalho infantil sobre o desempenho escolar: uma avaliação para o Brasil metropolitano

O objetivo deste trabalho é avaliar o seguinte contra factual: como teria sido o desempenho escolar das crianças que trabalham, caso tivessem sido efetivamente proibidas de trabalhar? Uma vez que não é possível observar as crianças que trabalham na situação de não-trabalho, a estratégia adotada consistiu em. construir um grupo de controle através do matching do propensity score (nearest-neighbor matching). Os dados utilizados neste estudo foram obtidos a partir de uma amostra da PME (Pesquisa Mensal de Emprego) para seis regiões metropolitanas do Brasil - constituída por crianças com idade entre 1 O e 14 anos acompanhadas por dois anos consecutivos ao longo do período de 1984 à 1997. Os resultados obtidos apontam para um impacto negativo do trabalho sobre o desempenho escolar das crianças que trabalham - embora em intensidade diferenciada segundo o indicador de desempenho utilizado. As estimativas para probabilidade de aprovação e para o progresso escolar sugerem um efeito negativo do trabalho, porém bem menor do que é geralmente observado. O que não ocorre com para a probabilidade de evasão - o trabalho explica quase a totalidade da diferença observada na probabilidade de evasão entre as crianças que trabalham e as demais. Os resultados sugerem, portanto, que se a legislação que proíbe o trabalho infantil fosse rigorosamente cumprida as crianças que trabalham apresentariam em média um melhor desempenho escolar

Análise de resultados intermediários das políticas de bônus em escolas públicas estaduais brasileiras

Programas de bônus para servidores de escolas públicas, visando estimular a melhoria da aprendizagem dos alunos, foram adotados por diferentes estados brasileiros entre 2008 e 2011. A presente pesquisa teve como objetivo avaliar o impacto desses programas em três dimensões: 1. práticas pedagógicas dos professores; 2. absenteísmo e rotatividade dos professores ou visão do diretor ou professor sobre a assiduidade e rotatividade do professor; 3. interlocução com familiares para garantir a frequência dos alunos. A coleta de dados foi feita a partir dos questionários da Prova Brasil, respondidos por diretores, professores e alunos das escolas do grupo de tratamento e do grupo controle. Foram considerados na análise: a variação no ano de implementação dos programas (dose), seus diferentes desenhos (tipo) e a série/ano dos alunos (série) que responderam a Prova Brasil. A abordagem metodológica adotada para a análise proposta foi a de pareamento pelo escore de propensão, seguido do cálculo de diferenças em diferenças, entre os Estados que implementaram a política e os que não adotaram. Os resultados variam de acordo com a série e dose, havendo impacto positivo para algumas categorias de análise na dimensão 1, relacionado à frequência com que professores corrigem o dever de casa. Na dimensão 2, os resultados apontam para uma redução de preocupação dos diretores com a rotatividade dos professores e aumento da percepção de que o absenteísmo dos docentes é um problema na realidade escolar pública brasileira. Quanto à dimensão 3, os dados revelam, de modo geral, ausência de impacto. Permanece a dúvida sobre a eficiência e a eficácia do estímulo financeiro para produzir mudanças no comportamento dos professores e diretores que possam promover resultados intermediários na melhoria da aprendizagem dos alunos.

L’impatto degli incentivi sull’efficienza delle imprese manifatturiere

Produttività ed efficienza sono termini comunemente utilizzati per caratterizzare l’abilità di un’impresa nell’utilizzazione delle risorse, sia in ambito privato che pubblico. Entrambi i concetti sono legati da una teoria della produzione che diventa essenziale per la determinazione dei criteri base con i quali confrontare i risultati dell’attività produttiva e i fattori impiegati per ottenerli. D’altronde, le imprese scelgono di produrre e di investire sulla base delle proprie prospettive di mercato e di costi dei fattori. Quest’ultimi possono essere influenzati dalle politiche dello Stato che fornisce incentivi e sussidi allo scopo di modificare le decisioni riguardanti l’allocazione e la crescita delle imprese. In questo caso le stesse imprese possono preferire di non collocarsi nell’equilibrio produttivo ottimo, massimizzando produttività ed efficienza, per poter invece utilizzare tali incentivi. In questo caso gli stessi incentivi potrebbero distorcere quindi l’allocazione delle risorse delle imprese che sono agevolate. L’obiettivo di questo lavoro è quello di valutare attraverso metodologie parametriche e non parametriche se incentivi erogati dalla L. 488/92, la principale politica regionale in Italia nelle regioni meridionali del paese nel periodo 1995-2004, hanno avuto o meno effetti sulla produttività totale dei fattori delle imprese agevolate. Si è condotta una ricognizione rispetto ai principali lavori proposti in letteratura riguardanti la TFP e l’aiuto alle imprese attraverso incentivi al capitale e (in parte) dell’efficienza. La stima della produttività totale dei fattori richiede di specificare una funzione di produzione ponendo l’attenzione su modelli di tipo parametrico che prevedono, quindi, la specificazione di una determinata forma funzionale relativa a variabili concernenti i fattori di produzione. Da questa si è ricavata la Total Factor Productivity utilizzata nell’analisi empirica che è la misura su cui viene valutata l’efficienza produttiva delle imprese. Il campione di aziende è dato dal merge tra i dati della L.488 e i dati di bilancio della banca dati AIDA. Si è provveduto alla stima del modello e si sono approfonditi diversi modelli per la stima della TFP; infine vengono descritti metodi non parametrici (tecniche di matching basate sul propensity score) e metodi parametrici (Diff-In-Diffs) per la valutazione dell’impatto dei sussidi al capitale. Si descrive l’analisi empirica condotta. Nella prima parte sono stati illustrati i passaggi cruciali e i risultati ottenuti a partire dalla elaborazione del dataset. Nella seconda parte, invece, si è descritta la stima del modello per la TFP e confrontate metodologie parametriche e non parametriche per valutare se la politica ha influenzato o meno il livello di TFP delle imprese agevolate.

Direct Mechanical Intervention Versus Combined Intravenous and Mechanical Intervention in Large Artery Anterior Circulation Stroke: A Matched-Pairs Analysis.

BACKGROUND AND PURPOSE Five randomized controlled trials have consistently shown that mechanical thrombectomy (MT) in addition to best medical treatment (±intravenous tissue-type plasminogen activator) improves outcome after acute ischemic stroke in patients with large artery anterior circulation stroke. Whether direct MT is equally effective as combined intravenous thrombolysis with MT (ie, bridging thrombolysis) remains unclear. METHODS We retrospectively compared clinical and radiological outcomes in 167 bridging patients with 255 patients receiving direct MT because of large artery anterior circulation stroke. We matched all patients from the direct MT group who would have qualified for intravenous tissue-type plasminogen activator with controls from the bridging group, using multivariate and propensity score analyses. Functional independence was defined as modified Rankin Scale score of 0 to 2. RESULTS From February 2009 to August 2014, 40 patients from the direct MT group would have qualified for bridging thrombolysis but were treated with MT only. Clinical and radiological characteristics did not differ from the bridging cohort, except for higher rates of hypercholesterolemia (P=0.019), coronary heart disease (P=0.039), and shorter intervals from symptom onset to endovascular intervention (P=0.01) in the direct MT group. Functional independence, mortality, and intracerebral hemorrhage rates did not differ (P>0.1). After multivariate matching analysis outcome in both groups did not differ, except for lower rates of asymptomatic intracerebral hemorrhage (P=0.023) and lower mortality (P=0.007) in the direct MT group. CONCLUSIONS In patients with large anterior circulation stroke, direct mechanical intervention seems to be equally effective as bridging thrombolysis. A randomized trial comparing direct MT with bridging therapy is warranted.

Superior outcomes of decompression with an interlaminar dynamic device versus decompression alone in patients with lumbar spinal stenosis and back pain: a cross registry study.

INTRODUCTION Surgical decompression for lumbar spinal stenosis (LSS) has been associated with poorer outcomes in patients with pronounced low back pain (LBP) as compared to patients with predominant leg pain. This cross registry study assessed potential benefits of the interlaminar coflex® device as an add-on to bony decompression alone. METHODS Patients with lumbar decompression plus coflex® (SWISSspine registry) were compared with decompressed controls (Spine Tango registry). Inclusion criteria were LSS and a preoperative back pain level of ≥5 points. 1:1 propensity score-based matching was performed. Outcome measures were back and leg pain relief, COMI score improvement, patient satisfaction, complication, and revision rates. RESULTS 50 matched pairs without residual significant differences but age were created. At the 7-9 months follow-up interval the coflex® group had higher back (p=0.014) and leg pain relief (p<0.001) and COMI score improvement (p=0.029) than the decompression group. Patient satisfaction was 90% in both groups. No revision was documented in the coflex® and one in the decompression group (2.0%). DISCUSSION In the short-term, lumbar decompression with coflex® compared with decompression alone in patients with LSS and pronounced LBP at baseline is a safe and effective treatment option that appears beneficial regarding clinical and functional outcomes. However, residual confounding of non-measured covariables may have partially influenced our findings. Also, despite careful inclusion and exclusion of cases the cross registry approach introduces a potential for selection bias that we could not totally control for and that makes additional studies necessary.

What is the Impact of the Underserved Pathway program on Entering an Underserved Family Medicine Residency? A Comparison of Three Approaches for Estimating the Average Treatment Effect

Thesis (Master's)--University of Washington, 2016-06

Streptokinase and Enoxaparin as an Alternative to Fibrin-Specific Lytic-Based Regimens An ExTRACT-TIMI 25 Analysis

Background: Enoxaparin was superior to unfractionated heparin (UFH), regardless of fibrinolytic agent in ST-elevation myocardial infarction (STEMI) patients receiving fibrinolytic therapy in ExTRACT-TIMI 25 (Enoxaparin and Thrombolysis Reperfusion for Acute Myocardial Infarction Treatment Thrombolysis in Myocardial Infarction 25) trial. Objective: This post hoc analysis compared outcomes with streptokinase plus enoxaparin to the standard regimen of fibrin-specific lytic (FSL) plus UFH and to the newer combination of FSL plus enoxaparin. Methods: In ExTRACT-TIMI 25, STEMI patients received either streptokinase or a FSL (alteplase, reteplase or tenecteplase) at the physician`s discretion and were randomized to enoxaparin or UFH, stratified by fibrinolytic type. Thirty-day outcomes were adjusted for baseline characteristics, region, in-hospital percutaneous coronary intervention (PCI) and a propensity score for the choice of lytic. Results: The primary trial endpoint of 30-day death/myocardial infarction (MI) occurred in fewer patients in the streptokinase-enoxaparin cohort (n = 2083) compared with FSL-UFH (n = 8141) [10.2% vs 12.0%, adjusted odds ratio [OR(adj)] 0.76; 95% CI 0.62, 0.93; p = 0.008]. Major bleeding was significantly increased with streptokinase-enoxaparin compared with FSL-UFH (ORadj 2.74; 95% CI 1.81; 4.14; p < 0.001) but intracranial haemorrhage (ICH) was similar (OR(adj) 0.90; 95% CI 0.40, 2.01; p = 0.79). Net clinical outcomes, defined as either death/MI/major bleeding or as death/MI/ICH tended to favour streptokinase-enoxaparin compared with FSL-UFH (OR(adj) 0.88; 95% CI 0.73, 1.06; p = 0.17; and OR(adj) 0.77; 95% CI 0.63, 0.93; p = 0.008, respectively). Patients receiving FSL-enoxaparin (n = 8142) and streptokinase-enoxaparin therapies experienced similar adjusted rates of the primary endpoint (OR(adj) 1.08; 95% CI 0.87, 1.32; p = 0.49) and net clinical outcomes. Conclusions: Our results suggest that fibrinolytic therapy with the combination of streptokinase and the potent anticoagulant agent enoxaparin resulted in similar adjusted outcomes compared with more costly regimens utilizing a FSL.

Angioplastia Primária na Mulher: Realidade Nacional

Objetivos: A mortalidade na mulher após angioplastia primária (ICP-P) é superior à do homem. Contudo, permanece contraditório o papel do sexo poder ser fator de risco independente para mortalidade no contexto de enfarte agudo do miocárdio com supradesnivelamento de ST (EAMST). Com base no Registo Nacional de Cardiologia de Intervenção (RNCI),pretendemos avaliar como é que o género feminino influencia o prognóstico a curto prazo nos doentes com EAMST submetidos a ICP-P a nível nacional. Métodos: De 60 158 doentes incluídos prospetivamente no RNCI de 2002-2012, incluímos na análise 7544 doentes com EAMST tratados por ICP-P, dos quais 25% foram mulheres. Utilizámos modelos de regressão logística e ajustamento por propensity score para avaliar o impacto do sexo na mortalidade hospitalar. Resultados: As mulheres foram mais idosas (68 ± 14 versus 61 ± 13, p < 0,001), mais diabéticas(30 versus 21%, p < 0,001) e hipertensas (69 versus 55%, p < 0,001). Os homens foram revascularizados mais cedo (71 versus 63% nas primeiras 6 horas, p < 0,001). Choque cardiogénico foi mais frequente nas mulheres (7,1 versus 5,7%, p = 0,032). Estas apresentaram um pior prognóstico a curto prazo, com 1,7 x maior risco de morte intra-hospitalar (4,3 versus 2,5%; IC 95% 1,30-2,27; p < 0,001). Utilizando um modelo de regressão ajustado através de um propensity score, o sexo deixa de ser preditor de mortalidade hospitalar (OR 1,00; IC 95% 0,68-1,48; p = 1,00). Conclusões: No RNCI as mulheres com EAMST tratadas com ICP-P apresentaram maior risco cardiovascular, um acesso menos atempado a ICP-P e um pior prognóstico. Contudo, após ajustamento do risco, o género feminino deixa de ser preditor independente de mortalidade hospitalar.

The Swiss Systemic lupus erythematosus Cohort Study (SSCS) - cross-sectional analysis of clinical characteristics and treatments across different medical disciplines in Switzerland.

OBJECTIVES: To describe disease characteristics and treatment modalities in a multidisciplinary cohort of systemic lupus erythematosus (SLE) patients in Switzerland. METHODS: Cross-sectional analysis of 255 patients included in the Swiss SLE Cohort and coming from centres specialised in Clinical Immunology, Internal Medicine, Nephrology and Rheumatology. Clinical data were collected with a standardised form. Disease activity was assessed using the Safety of Estrogens in Lupus Erythematosus National Assessment-SLE Disease Activity Index (SELENA-SLEDAI), an integer physician's global assessment score (PGA) ranging from 0 (inactive) to 3 (very active disease) and the erythrocyte sedimentation rate (ESR). The relationship between SLE treatment and activity was assessed by propensity score methods using a mixed-effect logistic regression with a random effect on the contributing centre. RESULTS: Of the 255 patients, 82% were women and 82% were of European ancestry. The mean age at enrolment was 44.8 years and the median SLE duration was 5.2 years. Patients from Rheumatology had a significantly later disease onset. Renal disease was reported in 44% of patients. PGA showed active disease in 49% of patients, median SLEDAI was 4 and median ESR was 14 millimetre/first hour. Prescription rates of anti-malarial drugs ranged from 3% by nephrologists to 76% by rheumatologists. Patients regularly using anti-malarial drugs had significantly lower SELENA-SLEDAI scores and ESR values. CONCLUSION: In our cohort, patients in Rheumatology had a significantly later SLE onset than those in Nephrology. Anti-malarial drugs were mostly prescribed by rheumatologists and internists and less frequently by nephrologists, and appeared to be associated with less active SLE.