Marginal structural models might overcome confounding when analyzing multiple treatment effects in observational studies

One-Hour Bundle Protocols for Surgical Sepsis and Septic Shock in Surgical Intensive Care Unit: Clinical Outcome Aspects in the Thai Context

BACKGROUND

Surgical sepsis is a syndrome occurring during the perioperative period with a high mortality rate. Since the one-hour bundle protocol was recommended to decrease sepsis-related morbidity and mortality in clinical practice, the protocol has been applied to surgical patients with sepsis and septic shock. However, clinical outcomes in these surgical patients remain unknown. Thus, this study aimed to compare survival outcomes in patients before and after the implementation of one-hour bundle care in clinical practice.

METHODS

In this prospective cohort study, 401 surgical patients with sepsis were divided into two groups, with 195 patients undergoing the one-hour bundle from December 25, 2021, to March 31, 2024, and 206 patients undergoing usual care from January 1, 2018, to December 24, 2021, before the one-hour bundle protocol was implemented by the Surviving Sepsis Campaign (SSC). Demographic data, treatment processes, and clinical outcomes were recorded.

RESULTS

After the one-hour bundle protocol was applied in surgical practice, the median survival time was significantly increased in surgical patients who underwent one-hour bundle care (95% confidence interval (CI): 12.32-19.68) (p= 0.016). Factors influencing the increase in the mortality rate were delays in fluid resuscitation of >2 hours, vasopressor initiation of >2 hours, and empirical antibiotics of >5 hours (p= 0.017, 0.028, and 0.008, respectively).

CONCLUSION

One-hour bundle care for surgical patients with sepsis resulted in an increased median survival time. Delays in fluid resuscitation (>2 hours), vasopressor initiation (>2 hours), and empirical antibiotics (>5 hours) were factors associated with mortality.

The Preserving Kidney Function in Children With CKD (PRESERVE) Study: Rationale, Design, and Methods

Rationale & Objective

PRESERVE seeks to provide new knowledge to inform shared decision-making regarding blood pressure (BP) management for pediatric chronic kidney disease (CKD). PRESERVE will compare the effectiveness of alternative strategies for monitoring and treating hypertension on preserving kidney function; expand the National Patient-Centered Clinical Research Network (PCORnet) common data model by adding pediatric- and kidney-specific variables and linking electronic health record data to other kidney disease databases; and assess the lived experiences of patients related to BP management.

Study Design

Multicenter retrospective cohort study (clinical outcomes) and cross-sectional study (patient-reported outcomes [PROs]).

Setting & Participants

PRESERVE will include approximately 20,000 children between January 2009-December 2022 with mild-moderate CKD from 15 health care institutions that participate in 6 PCORnet Clinical Research Networks (PEDSnet, STAR, GPC, PaTH, CAPRiCORN, and OneFlorida+). The inclusion criteria were ≥1 nephrologist visit and ≥2 estimated glomerular filtration rate (eGFR) values in the range of 30 to <90 mL/min/1.73 m2 separated by ≥90 days without an intervening value ≥90 mL/min/1.73 m2 and no prior dialysis or kidney transplant.

Exposures

BP measurements (clinic-based and 24-hour ambulatory BP); urine protein; and antihypertensive treatment by therapeutic class.

Outcomes

The primary outcome is a composite event of a 50% reduction in eGFR, eGFR of <15 mL/min/1.73 m2, long-term dialysis or kidney transplant. Secondary outcomes include change in eGFR, adverse events, and PROs.

Analytical Approach

Longitudinal models for dichotomous (proportional hazards or accelerated failure time) and continuous (generalized linear mixed models) clinical outcomes; multivariable linear regression for PROs. We will evaluate heterogeneity of treatment effect by CKD etiology and degree of proteinuria and will examine variation in hypertension management and outcomes based on socio-demographics.

Limitations

Causal inference limited by observational analyses.

Conclusions

PRESERVE will leverage the PCORnet infrastructure to conduct large-scale observational studies that address BP management knowledge gaps for pediatric CKD, focusing on outcomes that are meaningful to patients.

Plain-Language Summary

Hypertension is a major modifiable contributor to loss of kidney function in chronic kidney disease (CKD). The purpose of PRESERVE is to provide evidence to inform shared decision-making regarding blood pressure management for children with CKD. PRESERVE is a consortium of 16 health care institutions in PCORnet, the National Patient-Centered Clinical Research Network, and includes electronic health record data for >19,000 children with CKD. PRESERVE will (1) expand the PCORnet infrastructure for research in pediatric CKD by adding kidney-specific variables and linking electronic health record data to other kidney disease databases; (2) compare the effectiveness of alternative strategies for monitoring and treating hypertension on preserving kidney function; and (3) assess the lived experiences of patients and caregivers related to blood pressure management.

Longitudinal associations between insomnia symptoms and all-cause mortality among middle-aged and older adults: a population-based cohort study

To date, there is no scientific consensus on whether insomnia symptoms increase mortality risk. We investigated longitudinal associations between time-varying insomnia symptoms (difficulty initiating sleep, difficulty maintaining sleep, early-morning awakening, and nonrestorative sleep) and all-cause mortality among middle-aged and older adults during 14 years of follow-up. Data were obtained from 2004 through 2018 survey waves of the Health and Retirement Study in the United States for a population-representative sample of 15 511 respondents who were ≥50 years old in 2004. Respondents were interviewed biennially and followed through the end of the 2018 survey wave for the outcome. Marginal structural discrete-time survival analyses were employed to account for time-varying confounding and selection bias. Of the 15 511 cohort respondents (mean [±SD] age at baseline, 63.7 [±10.2] years; 56.0% females), 5878 (31.9%) died during follow-up. At baseline (2004), 41.6% reported experiencing at least one insomnia symptom. Respondents who experienced one (HR = 1.11; 95% CI: 1.03-1.20), two (HR = 1.12; 95% CI: 1.01-1.23), three (HR = 1.15; 95% CI: 1.05-1.27), or four (HR = 1.32; 95% CI: 1.12-1.56) insomnia symptoms had on average a higher hazard of all-cause mortality, compared to those who were symptom-free. For each insomnia symptom, respondents who experienced difficulty initiating sleep (HR = 1.12; 95% CI: 1.02-1.22), early-morning awakening (HR = 1.09; 95% CI: 1.01-1.18), and nonrestorative sleep (HR = 1.17; 95% CI: 1.09-1.26), had a higher hazard of all-cause mortality compared to those not experiencing the symptom. The findings demonstrate significant associations between insomnia symptoms and all-cause mortality, both on a cumulative scale and independently, except for difficulty maintaining sleep. Further research should investigate the underlying mechanisms linking insomnia symptoms and mortality.

Post-discharge healthcare utilization in infants with neonatal opioid withdrawal syndrome

The opioid epidemic in the United States has led to a significant increase in the incidence of neonatal opioid withdrawal syndrome (NOWS); however, the understanding of long-term consequences of NOWS is limited. The objective of this study was to evaluate post-discharge healthcare utilization in infants with NOWS and examine the association between NOWS severity and healthcare utilization. A retrospective cohort design was used to ascertain healthcare utilization in the first year after birth-related discharge using the CERNER Health Facts® database. ICD-9/ICD-10 diagnostic codes were used to identify live births and to classify infants into two study groups: NOWS and uncomplicated births (a 25% random sample). Evaluated outcomes included rehospitalization, emergency department (ED) visits within 30-days and one-year after discharge, and a composite one-year utilization event (either hospitalization or emergency department visit during that year). NOWS severity was operationalized as pharmacologic treatment, length of hospitalization, and medical conditions often associated with NOWS. In 3,526 infants with NOWS (restricted to gestational age ≥ 33 weeks), NOWS severity was associated with an increase in composite one-year utilization (OR: 1.1; 95% CI: 1.04-1.2) after adjusting for prematurity, sepsis, jaundice, use of antibiotics, infant sex, insurance status, race, hospital bed size, year of birth, and census division. In a subset of full-term infants (3008 with NOWS and 88,452 uncomplicated births), having a NOWS diagnosis was associated with higher odds of a 30-day (OR: 1.6; 95% CI: 1.03-2.4) and one-year rehospitalization (OR: 1.6; 95% CI: 1.1-2.4) after adjusting for infant sex, race, type of medical insurance, hospital location, census division, year of primary encounter, hospital bed size, and medical conditions. This study found higher healthcare utilization during the first year of life in infants diagnosed with NOWS, especially those with severe NOWS. Findings suggest a need for closer post-discharge follow-up and management of infants with NOWS.

Demographics of Vaccine Hesitancy in Chandigarh, India

The impact of vaccine hesitancy on childhood immunization in low- and middle-income countries remains largely uncharacterized. This study describes the sociodemographic patterns of vaccine hesitancy in Chandigarh, India. Mothers of children <5 years old were sampled from a two-stage cluster, systematic sample based on Anganwadi child care centers in Chandigarh. Vaccine hesitancy was measured using a 10-item Vaccine Hesitancy Scale, which was dichotomized. A multivariable logistic regression assessed the association between socioeconomic factors and vaccine hesitancy score. Among 305 mothers, >97% of mothers thought childhood vaccines were important, effective, and were a good way to protect against disease. However, many preferred their child to receive fewer co-administered vaccines (69%), and were concerned about side effects (39%). Compared to the “other caste” group, scheduled castes or scheduled tribes had 3.48 times greater odds of vaccine hesitancy (95% CI: 1.52, 7.99). Those with a high school education had 0.10 times the odds of vaccine hesitancy compared to those with less education (95% CI: 0.02, 0.61). Finally, those having more antenatal care visits were less vaccine hesitant (≥4 vs. <4 visits OR: 0.028, 95% CI: 0.1, 0.76). As India adds more vaccines to its Universal Immunization Program, consideration should be given to addressing maternal concerns about vaccination, in particular about adverse events and co-administration of multiple vaccines.

A machine learning compatible method for ordinal propensity score stratification and matching

Although machine learning techniques that estimate propensity scores for observational studies with multivalued treatments have advanced rapidly in recent years, the development of propensity score adjustment techniques has not kept pace. While machine learning propensity models provide numerous benefits, they do not produce a single variable balancing score that can be used for propensity score stratification and matching. This issue motivates the development of a flexible ordinal propensity scoring methodology that does not require parametric assumptions for the propensity model. The proposed method fits a one-parameter power function to the cumulative distribution function (CDF) of the generalized propensity score (GPS) vector resulting from any machine learning propensity model, and is henceforth called the GPS-CDF method. The estimated parameter from the GPS-CDF method, ã , is a scalar balancing score that can be used to group similar subjects in outcome analyses. Specifically, subjects who received different levels of the treatment are stratified or matched based on their ã value to produce unbiased estimates of the average treatment effect (ATE). Simulation studies presented show remediation of covariate balance, minimal bias in ATEs, and maintain coverage probability. The proposed method is applied to the Mexican-American Tobacco use in Children (MATCh) study to determine whether an ordinal treatment of exposure to smoking imagery in movies causes cigarette experimentation in Mexican-American adolescents.

In search of optimal lithium levels and olanzapine doses in the long-term treatment of bipolar I disorder. A post-hoc analysis of the maintenance study by Tohen et al. 2005

Purpose

The aim of this study was to investigate whether lower lithium levels (LoLi) or olanzapine doses (LoOL) are risk factors for future mood episodes in patients with bipolar I disorder.

Methods

A post-hoc analysis of the olanzapine-lithium-maintenance study [31] was performed using proportional hazards Cox regression models and marginal structural models (MSMs), adjusting for non-random assignments of dose during treatment.

Results

The LoLi group (< 0.6 mmol/L) had a significantly increased risk of manic/mixed (hazard ratio [HR] = 1.96, p = 0.042), but not depressive (HR = 2.11, p = 0.272) episodes, compared to the combined medium (0.6–0.79 mmol/L) and high lithium level (≥ 0.8 mmol/L) groups. There was no significant difference in risk between the two higher lithium level groups (0.6-0.79 mmol/L; ≥ 0.8 mmol/L) for new manic/mixed (HR = 0.96, p = 0.893) or depressive (HR = 0.95, p = 0.922) episodes. The LoOL group (< 10 mg/day) showed a significantly increased risk of depressive (HR = 2.24, p = 0.025) episodes compared to the higher olanzapine (HiOL) dose group (HiOL: 10–20 mg/day), while there was no statistically significant difference in risk for manic/mixed episodes between the two groups (HR = 0.94, p = 0.895).

Conclusion

Lithium levels ≥ 0.6 mmol/L and olanzapine doses ≥ 10 mg/day may be necessary for optimal protection against manic/mixed or depressive episodes, respectively in patients with bipolar I disorder.

Time to Antibiotics and the Outcome of Patients with Septic Shock: A Propensity Score Analysis

BACKGROUND

Current sepsis guidelines recommend administration of antibiotics within 1 hour of emergency department (ED) triage. However, the quality of the supporting evidence is moderate, and studies have shown mixed results regarding the association between antibiotic administration timing and outcomes in septic shock. We investigated to evaluate the association between antibiotic administration timing and in-hospital mortality in septic shock patients in the ED, using propensity score analysis.

METHODS

An observational study using a prospective, multicenter registry of septic shock, comprising data collected from 10 EDs, was conducted. Septic shock patients were included, and patients were divided into 4 groups by the interval from triage to first antibiotic administration: group 1 (≤1 hour; reference), 2 (1-2 hours), 3 (2-3 hours), and 4 (>3 hours). The primary endpoint was in-hospital mortality. After inverse probability of treatment weighting, the outcomes of the groups were compared.

RESULTS

A total of 2250 septic shock patients were included, and the median time to first antibiotic administration was 2.29 hours. The in-hospital mortality of groups 2 and 4 were significantly higher than those of group 1 (odds ratio [OR] 1.248; 95% confidence interval [CI], 1.053-1.478; P = .011; OR 1.419; 95% CI, 1.203-1.675; P < .001, respectively), but those of group 3 was not (OR 1.186; 95% CI, 0.999-1.408; P = .052). Subgroup analyses of patients (n = 2043) with appropriate antibiotics presented similar results.

CONCLUSIONS

In patients with septic shock, rapid administration of antibiotics was generally associated with a decrease in in-hospital mortality, but no "every hour delay" was seen.

Using patient self-reports to study heterogeneity of treatment effects in major depressive disorder

BACKGROUNDS

Clinicians need guidance to address the heterogeneity of treatment responses of patients with major depressive disorder (MDD). While prediction schemes based on symptom clustering and biomarkers have so far not yielded results of sufficient strength to inform clinical decision-making, prediction schemes based on big data predictive analytic models might be more practically useful.

METHOD

We review evidence suggesting that prediction equations based on symptoms and other easily-assessed clinical features found in previous research to predict MDD treatment outcomes might provide a foundation for developing predictive analytic clinical decision support models that could help clinicians select optimal (personalised) MDD treatments. These methods could also be useful in targeting patient subsamples for more expensive biomarker assessments.

RESULTS

Approximately two dozen baseline variables obtained from medical records or patient reports have been found repeatedly in MDD treatment trials to predict overall treatment outcomes (i.e., intervention v. control) or differential treatment outcomes (i.e., intervention A v. intervention B). Similar evidence has been found in observational studies of MDD persistence-severity. However, no treatment studies have yet attempted to develop treatment outcome equations using the full set of these predictors. Promising preliminary empirical results coupled with recent developments in statistical methodology suggest that models could be developed to provide useful clinical decision support in personalised treatment selection. These tools could also provide a strong foundation to increase statistical power in focused studies of biomarkers and MDD heterogeneity of treatment response in subsequent controlled trials.

CONCLUSIONS

Coordinated efforts are needed to develop a protocol for systematically collecting information about established predictors of heterogeneity of MDD treatment response in large observational treatment studies, applying and refining these models in subsequent pragmatic trials, carrying out pooled secondary analyses to extract the maximum amount of information from these coordinated studies, and using this information to focus future discovery efforts in the segment of the patient population in which continued uncertainty about treatment response exists.

Severe COPD Exacerbation Risk and Long-Acting Bronchodilator Treatments: Comparison of Three Observational Data Analysis Methods

OBJECTIVE

Results from three observational methods for assessing effectiveness of long-acting bronchodilator therapies for reducing severe exacerbations of chronic obstructive pulmonary disease (COPD) were compared: intent-to-treat (ITT), as protocol (AP), and an as-treated analysis that utilized a marginal structural model (MSM) incorporating time-varying covariates related to treatment adherence and moderate exacerbations.

STUDY DESIGN AND SETTING

Severe exacerbation risk was assessed over a 2-year period using claims data for patients aged ≥40 years who initiated long-acting muscarinic antagonist (LAMA), inhaled corticosteroid/long-acting beta-agonist (ICS/LABA), or triple therapy (LAMA + ICS/LABA).

RESULTS

A total of 5475 COPD patients met inclusion criteria. Six months post-initiation, 53.5 % of patients discontinued using any therapy. The ITT analysis found an increased severe exacerbation risk for triple therapy treatment (hazard ratio [HR] 1.24; 95 % confidence interval [CI] 1.00-1.53). No increased risk was found in the AP (HR 1.00; 95 % CI 0.73-1.36), or MSM analyses (HR 1.11; 95 % CI 0.68-1.81). The MSM highlighted important associations among post-index events.

CONCLUSION

Neglecting to adjust for treatment discontinuation may produce biased risk estimates. The MSM approach is a promising tool to compare chronic disease management by illuminating relationships between treatment decisions, adherence, patient choices, and outcomes.

Comparative effectiveness of colony-stimulating factors in febrile neutropenia prophylaxis: how results are affected by research design

Aims: To examine the impact of research design on results in two published comparative effectiveness studies. Methods: Guidelines for comparative effectiveness research have recommended incorporating disease process in study design. Based on the recommendations, we develop a checklist of considerations and apply the checklist in review of two published studies on comparative effectiveness of colony-stimulating factors. Both studies used similar administrative claims data, but different methods, which resulted in directionally different estimates. Results: Major design differences between the two studies include: whether the timing of intervention in disease process was identified and whether study cohort and outcome assessment period were defined based on this temporal relationship. Conclusion: Disease process and timing of intervention should be incorporated into the design of comparative effectiveness studies.

Validation of predictors of adverse outcomes in hospital-acquired pneumonia in the ICU

OBJECTIVE

To validate a set of predictors of adverse outcomes in patients with ICU-acquired pneumonia in relation to clinically relevant assessment at 28 days.

DESIGN

Prospective, observational study.

SETTING

Six medical and surgical ICUs of a university hospital.

PATIENTS

Three hundred thirty-five patients with ICU-acquired pneumonia.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

Development of predictors of adverse outcomes was defined when at least one of the following criteria was present at an evaluation made 72-96 hours after starting treatment: no improvement of PaO2/FIO2, need for intubation due to pneumonia, persistence of fever or hypothermia with purulent respiratory secretions, greater than or equal to 50% increase in radiographic infiltrates, or occurrence of septic shock or multiple organ dysfunction syndrome. We also assessed the inflammatory response by different serum biomarkers. The presence of predictors of adverse outcomes was related to mortality and ventilator-free days at day 28. Sequential Organ Failure Assessment score was evaluated and related to mortality at day 28.One hundred eighty-four (55%) patients had at least one predictor of adverse outcomes. The 28-day mortality was higher for those with versus those without predictors of adverse outcomes (45% vs 19%, p<0.001), and ventilator-free days were lower (median [interquartile range], 0 [0-17] vs 22 [0-28]) for patients with versus patients without predictors of adverse outcomes (p<0.001). The lack of improvement of PaO2/FIO2 and lack of improvement in Sequential Organ Failure Assessment score from day 1 to day 5 were independently associated with 28-day mortality and fewer ventilator-free days. The marginal structural analysis showed an odds ratio of death 2.042 (95% CI, 1.01-4.13; p=0.047) in patients with predictors of adverse outcomes. Patients with predictors of adverse outcomes had higher serum inflammatory response accordingly to biomarkers evaluated.

CONCLUSIONS

The presence of any predictors of adverse outcomes was associated with mortality and decreased ventilator-free days at day 28. The lack of improvement in the PaO2/FIO2 and Sequential Organ Failure Assessment score was independently associated with mortality in the multivariate analysis.

Making valid causal inferences from observational data

The ability to make strong causal inferences, based on data derived from outside of the laboratory, is largely restricted to data arising from well-designed randomized control trials. Nonetheless, a number of methods have been developed to improve our ability to make valid causal inferences from data arising from observational studies. In this paper, I review concepts of causation as a background to counterfactual causal ideas; the latter ideas are central to much of current causal theory. Confounding greatly constrains causal inferences in all observational studies. Confounding is a biased measure of effect that results when one or more variables, that are both antecedent to the exposure and associated with the outcome, are differentially distributed between the exposed and non-exposed groups. Historically, the most common approach to control confounding has been multivariable modeling; however, the limitations of this approach are discussed. My suggestions for improving causal inferences include asking better questions (relates to counterfactual ideas and "thought" trials); improving study design through the use of forward projection; and using propensity scores to identify potential confounders and enhance exchangeability, prior to seeing the outcome data. If time-dependent confounders are present (as they are in many longitudinal studies), more-advanced methods such as marginal structural models need to be implemented. Tutorials and examples are cited where possible.

Effects of etomidate on complications related to intubation and on mortality in septic shock patients treated with hydrocortisone: a propensity score analysis

Introduction

Endotracheal intubation in the ICU is associated with a high incidence of complications. Etomidate use is debated in septic shock because it increases the risk of critical illness-related corticosteroid insufficiency, which may impact outcome. We hypothesized that hydrocortisone, administered in all septic shock cases in our ICU, may counteract some negative effects of etomidate.

The aim of our study was to compare septic shock patients who received etomidate versus another induction drug both for short-term safety and for long-term outcomes.

Methods

A single-center observational study was carried out in septic shock patients, treated with hydrocortisone and intubated within the first 48 hours of septic shock. Co-primary end points were life-threatening complications incidence occurring within the first hour after intubation and mortality during the ICU stay. Statistical analyses included unmatched and matched cohorts using a propensity score analysis. P < 0.05 was considered significant.

Results

Sixty patients in the etomidate cohort and 42 patients in the non-etomidate cohort were included. Critical illness-related corticosteroid insufficiency was 79% in the etomidate cohort and 52% in the non-etomidate cohort (P = 0.01). After intubation, life-threatening complications occurred in 36% of the patients whatever the cohort. After adjustment with propensity score analysis, etomidate was a protective factor for death in the ICU both in unmatched (hazard ratio, 0.33 (0.15 to 0.75); P < 0.01)) and matched cohorts (hazard ratio, 0.33 (0.112 to 0.988); P = 0.04).

Conclusion

In septic shock patients treated with hydrocortisone, etomidate did not decrease life-threatening complications following intubation, but when associated with hydrocortisone it also did not impair outcome.

Impact of adequate empirical combination therapy on mortality from bacteremic Pseudomonas aeruginosa pneumonia

BACKGROUND

Pseudomonas aeruginosa has gained an increasing amount of attention in the treatment of patients with pneumonia. However, the benefit of empirical combination therapy for pneumonia remains unclear. We evaluated the effects of adequate empirical combination therapy and multidrug-resistance in bacteremic Pseudomonas pneumonia on the mortality.

METHODS

A retrospective cohort study was performed at the 2,700-bed tertiary care university hospital. We reviewed the medical records of patients with bacteremic pneumonia between January 1997 and February 2011. Patients who received either inappropriate or appropriate empirical therapy were compared by using marginal structural model. Furthermore, we investigated the direct impact of combination therapy on clinical outcomes in patients with monomicrobial bacteremic pneumonia.

RESULTS

Among 100 consecutive patients with bacteremic Pseudomonas pneumonia, 65 patients were classified in the adequate empirical therapy group, 32 of whom received monotherapy and 33 combination therapy. In the marginal structural model, only inadequate therapy was significantly associated with 28-day mortality (p = 0.02), and multidrug-resistance was not a significant risk factor.To examine further the direct impact of combination therapy, we performed a subgroup analysis of the 65 patients who received adequate therapy. Multivariate logistic regression analysis identified absence of septic shock at the time of bacteremia (OR, 0.07; 95% CI, 0.01-0.49; p = 0.008), and adequate combination therapy (OR, 0.05; 95% CI, 0.01-0.34; p = 0.002) as variables independently associated with decreased all-cause 28-day mortality.

CONCLUSIONS

Our study suggests that adequate empirical combination therapy can decrease mortality in patients with bacteremic Pseudomonas pneumonia.

Intensive care unit discharge to the ward with a tracheostomy cannula as a risk factor for mortality: a prospective, multicenter propensity analysis

OBJECTIVE

To analyze the impact of decannulation before intensive care unit discharge on ward survival in nonexperimental conditions.

DESIGN

Prospective, observational survey.

SETTING

Thirty-one intensive care units throughout Spain.

PATIENTS

All patients admitted from March 1, 2008 to May 31, 2008.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

At intensive care unit discharge, we recorded demographic variables, severity score, and intensive care unit treatments, with special attention to tracheostomy. After intensive care unit discharge, we recorded intensive care unit readmission and hospital survival.

STATISTICS

Multivariate analyses for ward mortality, with Cox proportional hazard ratio adjusted for propensity score for intensive care unit decannulation. We included 4,132 patients, 1,996 of whom needed mechanical ventilation. Of these, 260 (13%) were tracheostomized and 59 (23%) died in the intensive care unit. Of the 201 intensive care unit tracheostomized survivors, 60 were decannulated in the intensive care unit and 141 were discharged to the ward with cannulae in place. Variables associated with intensive care unit decannulation (non-neurologic disease [85% vs. 64%], vasoactive drugs [90% vs. 76%], parenteral nutrition [55% vs. 33%], acute renal failure [37% vs. 23%], and good prognosis at intensive care unit discharge [40% vs. 18%]) were included in a propensity score model for decannulation. Crude ward mortality was similar in decannulated and nondecannulated patients (22% vs. 23%); however, after adjustment for the propensity score and Sabadell Score, the presence of a tracheostomy cannula was not associated with any survival disadvantage with an odds ratio of 0.6 [0.3-1.2] (p=.1).

CONCLUSION

In our multicenter setting, intensive care unit discharge before decannulation is not a risk factor.

Key points for designing and reporting Helicobacter pylori therapeutic trials: a personal view

BACKGROUND & OBJECTIVES

The aim of this document is to provide a methodological framework and to review key aspects for adequately designing trials to evaluate new treatments for Helicobacter pylori infection.

METHODS

Non-systematic literature review.

RESULTS & CONCLUSIONS

Regarding the design of the article, we suggest selecting for future trials drugs to which H. pylori has no significant primary resistances and evaluating therapies with pilot studies before engaging in randomized trials. The manuscript defines how the number and type of H. pylori diagnostic tests necessary before and after the trial depend on the setting and reliability of the tests. It recommends the best methods and timing for H. pylori testing before and after therapy. Other recommendations are using current standard treatments as comparators of new therapies, determining antibiotic sensitivity - whenever useful and possible - using adequate randomization and allocation concealment but not necessarily blinding, and performing an intention-to-treat and a per-protocol analysis. In addition, we give basic tips for reporting and discussing study results.

Cost-effectiveness of the Surviving Sepsis Campaign protocol for severe sepsis: a prospective nation-wide study in Spain

CONTEXT

Severe sepsis is associated with high mortality and increased costs. The 'Surviving Sepsis Campaign' (SSC) protocol was developed as an international initiative to reduce mortality. However, its cost-effectiveness is unknown.

OBJECTIVE

To determine the cost-effectiveness of the SSC protocol for the treatment of severe sepsis in Spain after the implementation of an educational program compared with the conventional care of severe sepsis.

DESIGN

Observational prospective before-and-after study.

SETTING

59 medical-surgical intensive care units located throughout Spain.

PATIENTS

A total of 854 patients were enrolled in the pre-educational program cohort (usual or standard care of severe sepsis) and 1,465 patients in the post-educational program cohort (SSC protocol care of severe sepsis).

INTERVENTIONS

The educational program aimed to increase adherence to the SSC protocol. The SSC protocol included pharmacological and medical interventions.

MAIN OUTCOME MEASURES

Clinical (hospital mortality) and economic (health-care resource and treatment costs) outcomes were recorded. A health-care system perspective was used for costs. The primary outcome was incremental cost-effectiveness ratio (ICER).

RESULTS

Patients in the SSC protocol care cohort had a lower risk of hospital mortality (44.0% vs. 39.7%, P = 0.04). However, mean costs per patient were 1,736 euros higher in the SSC protocol care cohort (95% CI 114-3,358 euros), largely as a result of increased length of stay. Mean life years gained (LYG) were higher in the SSC protocol care cohort: 0.54 years (95% CI 0.02-1.05 years). The adjusted ICER of the SSC protocol was 4,435 euros per LYG. Nearly all (96.5%) the bootstrap replications were below the threshold of 30,000 euros per LYG.

CONCLUSION

The SSC protocol seems to be a cost-effective option for treating severe sepsis in Spain.

Predictors and clinical consequences of non-adherence with antipsychotic medication in the outpatient treatment of schizophrenia

To assess baseline predictors and consequences of antipsychotic adherence during the long-term treatment of schizophrenia outpatients, data were taken from the 3-year, prospective, observational, European Schizophrenia Outpatients Health Outcomes (SOHO) study, in which outpatients starting or changing antipsychotics were assessed every 6 months. Physician-rated adherence was dichotomized as adherence/non-adherence. Regression models tested for predictors of adherence during follow-up, and associations between adherence and outcome measures. Of the 6731 patients analysed, 71.2% were adherent and 28.8% were non-adherent over 3 years. The strongest predictor of adherence was adherence in the month before baseline assessment. Other baseline predictors of adherence included initial treatment for schizophrenia and greater social activities. Baseline predictors of non-adherence were alcohol dependence and substance abuse in the previous month, hospitalization in the previous 6 months, independent housing and the presence of hostility. Non-adherence was significantly associated with an increased risk of relapse, hospitalization and suicide attempts. In conclusion, non-adherence is common but can partly be predicted. This may allow strategies to improve adherence to be targeted to high-risk patients. Also, reversal of some risk factors may improve adherence. Non-adherence is associated with a range of poorer long-term outcomes, with clinical and economic implications.

Effectiveness of treatments for severe sepsis: a prospective, multicenter, observational study

RATIONALE

Several Surviving Sepsis Campaign Guidelines recommendations are reevaluated.

OBJECTIVES

To analyze the effectiveness of treatments recommended in the sepsis guidelines.

METHODS

In a prospective observational study, we studied all adult patients with severe sepsis from 77 intensive care units. We recorded compliance with four therapeutic goals (central venous pressure 8 mm Hg or greater for persistent hypotension despite fluid resuscitation and/or lactate greater than 36 mg/dl, central venous oxygen saturation 70% or greater for persistent hypotension despite fluid resuscitation and/or lactate greater than 36 mg/dl, blood glucose greater than or equal to the lower limit of normal but less than 150 mg/dl, and inspiratory plateau pressure less than 30 cm H(2)O for mechanically ventilated patients) and four treatments (early broad-spectrum antibiotics, fluid challenge in the event of hypotension and/or lactate greater than 36 mg/dl, low-dose steroids for septic shock, drotrecogin alfa [activated] for multiorgan failure). The primary outcome measure was hospital mortality. The effectiveness of each treatment was estimated using propensity scores.

MEASUREMENTS AND MAIN RESULTS

Of 2,796 patients, 41.6% died before hospital discharge. Treatments associated with lower hospital mortality were early broad-spectrum antibiotic treatment (treatment within 1 hour vs. no treatment within first 6 hours of diagnosis; odds ratio, 0.67; 95% confidence interval, 0.50-0.90; P = 0.008) and drotrecogin alfa (activated) (odds ratio, 0.59; 95% confidence interval, 0.41-0.84; P = 0.004). Fluid challenge and low-dose steroids showed no benefits.

CONCLUSIONS

In severe sepsis, early administration of broad-spectrum antibiotics in all patients and administration of drotrecogin alfa (activated) in the most severe patients reduce mortality.

Recovery in the outpatient setting: 36-month results from the Schizophrenia Outpatients Health Outcomes (SOHO) study

Recovery is an important outcome of schizophrenia that has not been well defined or researched. Using a stringent definition of recovery that included long-lasting symptomatic and functional remission as well as an adequate quality of life for a minimum of 24 months and until the 36-month visit, we determined the frequency and predictors of recovery in patients with schizophrenia during 3 years of antipsychotic treatment in the prospective, observational Schizophrenia Outpatients Health Outcomes (SOHO) study. Of the 6642 patients analysed, 33% achieved long-lasting symptomatic remission, 13% long-lasting functional remission, 27% long-lasting adequate quality of life, and 4% achieved recovery during the 3 year follow-up period. Logistic regression analysis revealed that social functioning at study entry (having good occupational/vocational status, living independently and being socially active) and adherence with medication were factors significantly associated with achieving recovery. Higher negative symptom severity, higher BMI and lack of effectiveness as the reason for change of medication at baseline were baseline factors associated with a lower likelihood of achieving recovery. Treatment with olanzapine was also associated with a higher frequency of recovery compared with risperidone, quetiapine, typical antipsychotics (oral, depot) and patients taking two or more antipsychotic medications. There were no differences among the patients taking olanzapine, clozapine and amisulpride. Predictors of long-lasting symptomatic remission, functional remission and adequate quality of life were also independently analysed. Although the results should be interpreted conservatively due to the observational, non-randomised study design, they indicate that only a small proportion of patients with schizophrenia achieve recovery and suggest that social functioning, medication adherence and type of antipsychotic are important predictors of recovery.