Using natural language processing to study homelessness longitudinally with electronic health record data subject to irregular observations

The Electronic Health Record (EHR) contains information about social determinants of health (SDoH) such as homelessness. Much of this information is contained in clinical notes and can be extracted using natural language processing (NLP). This data can provide valuable information for researchers and policymakers studying long-term housing outcomes for individuals with a history of homelessness. However, studying homelessness longitudinally in the EHR is challenging due to irregular observation times. In this work, we applied an NLP system to extract housing status for a cohort of patients in the US Department of Veterans Affairs (VA) over a three-year period. We then applied inverse intensity weighting to adjust for the irregularity of observations, which was used generalized estimating equations to estimate the probability of unstable housing each day after entering a VA housing assistance program. Our methods generate unique insights into the long-term outcomes of individuals with a history of homelessness and demonstrate the potential for using EHR data for research and policymaking.

Do superficial infections increase the risk of deep infections in tibial plateau and plafond fractures?

PURPOSE

Open reduction internal fixation of tibial plateau and pilon fractures may be complicated by deep surgical site infection requiring operative debridement and antibiotic therapy. The management of superficial surgical site infection is controversial. We sought to determine whether superficial infection is associated with an increased risk of deep infection requiring surgical debridement after fixation of tibial plateau and pilon fractures.

METHODS

This is a secondary analysis of data from the VANCO trial, which included 980 adult patients with a tibial plateau or pilon fracture at elevated risk of infection who underwent open reduction internal fixation with plates and screws with or without intrawound vancomycin powder. An association of superficial surgical site infection with deep surgical site infection requiring debridement surgery and antibiotics was explored after matching on risk factors for deep surgical site infection.

RESULTS

Of the 980 patients, we observed 30 superficial infections (3.1%) and 76 deep infections (7.8%). Among patients who developed a superficial infection, the unadjusted incidence of developing a deep infection within 90 days was 12.8% (95% confidence interval [CI] 1.3-24.2%). However, after a 3:1 match on infection risk factors, the 90-day marginal probability of a deep surgical site infection after sustaining a superficial infection was 6.0% (95% CI  - 6.5-18.5%, p = 0.35).

CONCLUSION

Deep infection after superficial infection is uncommon following operative fixation of tibial plateau and pilon fractures. Increased risk of subsequent deep infection attributable to superficial infection was inconclusive in these data.

LEVEL OF EVIDENCE

Prognostic Level II.

New Users of Angiotensin II Receptor Blocker-Versus Angiotensin-Converting Enzyme Inhibitor-Based Antihypertensive Medication Regimens and Cardiovascular Disease Events: A Secondary Analysis of ACCORD-BP and SPRINT

BACKGROUND Angiotensin II receptor blockers (ARBs) and angiotensin-converting enzyme inhibitors (ACEIs) block distinct components of the renin-angiotensin system. Whether this translates into differential effects on cardiovascular disease events remains unclear. METHODS AND RESULTS We used the ACCORD-BP (Action to Control Cardiovascular Risk in Diabetes-Blood Pressure) trial and the SPRINT (Systolic Blood Pressure Intervention Trial) to emulate target trials of new users of ARBs versus ACEIs on cardiovascular disease events (primary outcome) and death (secondary outcome). We estimated marginal cause-specific hazard ratios (HRs) and treatment-specific cumulative incidence functions with inverse probability of treatment weights. We identified 3298 new users of ARBs or ACEIs (ACCORD-BP: 374 ARB versus 884 ACEI; SPRINT: 727 ARB versus 1313 ACEI). For participants initiating ARBs versus ACEIs, the inverse probability of treatment weight rate of the primary outcome was 3.2 versus 3.5 per 100 person-years in ACCORD-BP (HR, 0.91 [95% CI, 0.63-1.31]) and 1.8 versus 2.2 per 100 person-years in SPRINT (HR, 0.81 [95% CI, 0.56-1.18]). There were no appreciable differences in pooled analyses, except that ARBs versus ACEIs were associated with a lower death rate (HR, 0.56 [95% CI, 0.37-0.85]). ARBs were associated with a lower rate of the primary outcome among subgroups of male versus female participants, non-Hispanic Black versus non-Hispanic White participants, and those randomly assigned to standard versus intensive blood pressure (Pinteraction: <0.01, 0.05, and <0.01, respectively). CONCLUSIONS In this secondary analysis of ACCORD-BP and SPRINT, new users of ARB- versus ACEI-based antihypertensive medication regimens experienced similar cardiovascular disease events rates, with important subgroup differences and lower rates of death overall. REGISTRATION URL: https://www.clinicaltrials.gov; Unique identifiers: NCT01206062, NCT00000620.

Losartan in hospitalized patients with COVID-19 in North America: An individual participant data meta-analysis

BACKGROUND

Angiotensin-converting enzyme inhibitors and angiotensin II receptor blockers (ARBs) have been hypothesized to benefit patients with COVID-19 via the inhibition of viral entry and other mechanisms. We conducted an individual participant data (IPD) meta-analysis assessing the effect of starting the ARB losartan in recently hospitalized COVID-19 patients.

METHODS

We searched ClinicalTrials.gov in January 2021 for U.S./Canada-based trials where an angiotensin-converting enzyme inhibitors/ARB was a treatment arm, targeted outcomes could be extrapolated, and data sharing was allowed. Our primary outcome was a 7-point COVID-19 ordinal score measured 13 to 16 days post-enrollment. We analyzed data by fitting multilevel Bayesian ordinal regression models and standardizing the resulting predictions.

RESULTS

325 participants (156 losartan vs 169 control) from 4 studies contributed IPD. Three were randomized trials; one used non-randomized concurrent and historical controls. Baseline covariates were reasonably balanced for the randomized trials. All studies evaluated losartan. We found equivocal evidence of a difference in ordinal scores 13-16 days post-enrollment (model-standardized odds ratio [OR] 1.10, 95% credible interval [CrI] 0.76-1.71; adjusted OR 1.15, 95% CrI 0.15-3.59) and no compelling evidence of treatment effect heterogeneity among prespecified subgroups. Losartan had worse effects for those taking corticosteroids at baseline after adjusting for covariates (ratio of adjusted ORs 0.29, 95% CrI 0.08-0.99). Hypotension serious adverse event rates were numerically higher with losartan.

CONCLUSIONS

In this IPD meta-analysis of hospitalized COVID-19 patients, we found no convincing evidence for the benefit of losartan versus control treatment, but a higher rate of hypotension adverse events with losartan.

Using Trial and Observational Data to Assess Effectiveness: Trial Emulation, Transportability, Benchmarking, and Joint Analysis

Comparisons between randomized trial analyses and observational analyses that attempt to address similar research questions have generated many controversies in epidemiology and the social sciences. There has been little consensus on when such comparisons are reasonable, what their implications are for the validity of observational analyses, or whether trial and observational analyses can be integrated to address effectiveness questions. Here, we consider methods for using observational analyses to complement trial analyses when assessing treatment effectiveness. First, we review the framework for designing observational analyses that emulate target trials and present an evidence map of its recent applications. We then review approaches for estimating the average treatment effect in the target population underlying the emulation: using observational analyses of the emulation data alone; and using transportability analyses to extend inferences from a trial to the target population. We explain how comparing treatment effect estimates from the emulation against those from the trial can provide evidence on whether observational analyses can be trusted to deliver valid estimates of effectiveness - a process we refer to as benchmarking - and, in some cases, allow the joint analysis of the trial and observational data. We illustrate different approaches using a simplified example of a pragmatic trial and its emulation in registry data. We conclude that synthesizing trial and observational data - in transportability, benchmarking, or joint analyses - can leverage their complementary strengths to enhance learning about comparative effectiveness, through a process combining quantitative methods and epidemiological judgements.

Aspirin or Low-Molecular-Weight Heparin for Thromboprophylaxis after a Fracture

BACKGROUND

Clinical guidelines recommend low-molecular-weight heparin for thromboprophylaxis in patients with fractures, but trials of its effectiveness as compared with aspirin are lacking.

METHODS

In this pragmatic, multicenter, randomized, noninferiority trial, we enrolled patients 18 years of age or older who had a fracture of an extremity (anywhere from hip to midfoot or shoulder to wrist) that had been treated operatively or who had any pelvic or acetabular fracture. Patients were randomly assigned to receive low-molecular-weight heparin (enoxaparin) at a dose of 30 mg twice daily or aspirin at a dose of 81 mg twice daily while they were in the hospital. After hospital discharge, the patients continued to receive thromboprophylaxis according to the clinical protocols of each hospital. The primary outcome was death from any cause at 90 days. Secondary outcomes were nonfatal pulmonary embolism, deep-vein thrombosis, and bleeding complications.

RESULTS

A total of 12,211 patients were randomly assigned to receive aspirin (6101 patients) or low-molecular-weight heparin (6110 patients). Patients had a mean (±SD) age of 44.6±17.8 years, 0.7% had a history of venous thromboembolism, and 2.5% had a history of cancer. Patients received a mean of 8.8±10.6 in-hospital thromboprophylaxis doses and were prescribed a median 21-day supply of thromboprophylaxis at discharge. Death occurred in 47 patients (0.78%) in the aspirin group and in 45 patients (0.73%) in the low-molecular-weight-heparin group (difference, 0.05 percentage points; 96.2% confidence interval, -0.27 to 0.38; P<0.001 for a noninferiority margin of 0.75 percentage points). Deep-vein thrombosis occurred in 2.51% of patients in the aspirin group and 1.71% in the low-molecular-weight-heparin group (difference, 0.80 percentage points; 95% CI, 0.28 to 1.31). The incidence of pulmonary embolism (1.49% in each group), bleeding complications, and other serious adverse events were similar in the two groups.

CONCLUSIONS

In patients with extremity fractures that had been treated operatively or with any pelvic or acetabular fracture, thromboprophylaxis with aspirin was noninferior to low-molecular-weight heparin in preventing death and was associated with low incidences of deep-vein thrombosis and pulmonary embolism and low 90-day mortality. (Funded by the Patient-Centered Outcomes Research Institute; PREVENT CLOT ClinicalTrials.gov number, NCT02984384.).

Randomized Trials With Repeatedly Measured Outcomes: Handling Irregular and Potentially Informative Assessment Times

Randomized trials are often designed to collect outcomes at fixed points in time after randomization. In practice, the number and timing of outcome assessments can vary among participants (i.e., irregular assessment). In fact, the timing of assessments may be associated with the outcome of interest (i.e., informative assessment). For example, in a trial evaluating the effectiveness of treatments for major depressive disorder, not only did the timings of outcome assessments vary among participants but symptom scores were associated with assessment frequency. This type of informative observation requires appropriate statistical analysis. Although analytic methods have been developed, they are rarely used. In this article, we review the literature on irregular assessments with a view toward developing recommendations for analyzing trials with irregular and potentially informative assessment times. We show how the choice of analytic approach hinges on assumptions about the relationship between the assessment and outcome processes. We argue that irregular assessment should be treated with the same care as missing data, and we propose that trialists adopt strategies to minimize the extent of irregularity; describe the extent of irregularity in assessment times; make their assumptions about the relationships between assessment times and outcomes explicit; adopt analytic techniques that are appropriate to their assumptions; and assess the sensitivity of trial results to their assumptions.

Hydroxychloroquine/chloroquine for the treatment of hospitalized patients with COVID-19: An individual participant data meta-analysis

BACKGROUND

Results from observational studies and randomized clinical trials (RCTs) have led to the consensus that hydroxychloroquine (HCQ) and chloroquine (CQ) are not effective for COVID-19 prevention or treatment. Pooling individual participant data, including unanalyzed data from trials terminated early, enables more detailed investigation of the efficacy and safety of HCQ/CQ among subgroups of hospitalized patients.

METHODS

We searched ClinicalTrials.gov in May and June 2020 for US-based RCTs evaluating HCQ/CQ in hospitalized COVID-19 patients in which the outcomes defined in this study were recorded or could be extrapolated. The primary outcome was a 7-point ordinal scale measured between day 28 and 35 post enrollment; comparisons used proportional odds ratios. Harmonized de-identified data were collected via a common template spreadsheet sent to each principal investigator. The data were analyzed by fitting a prespecified Bayesian ordinal regression model and standardizing the resulting predictions.

RESULTS

Eight of 19 trials met eligibility criteria and agreed to participate. Patient-level data were available from 770 participants (412 HCQ/CQ vs 358 control). Baseline characteristics were similar between groups. We did not find evidence of a difference in COVID-19 ordinal scores between days 28 and 35 post-enrollment in the pooled patient population (odds ratio, 0.97; 95% credible interval, 0.76-1.24; higher favors HCQ/CQ), and found no convincing evidence of meaningful treatment effect heterogeneity among prespecified subgroups. Adverse event and serious adverse event rates were numerically higher with HCQ/CQ vs control (0.39 vs 0.29 and 0.13 vs 0.09 per patient, respectively).

CONCLUSIONS

The findings of this individual participant data meta-analysis reinforce those of individual RCTs that HCQ/CQ is not efficacious for treatment of COVID-19 in hospitalized patients.

Hydroxychloroquine/chloroquine for the treatment of hospitalized patients with COVID-19: An individual participant data meta-analysis

Background

Results from observational studies and randomized clinical trials (RCTs) have led to the consensus that hydroxychloroquine (HCQ) and chloroquine (CQ) are not effective for COVID-19 prevention or treatment. Pooling individual participant data, including unanalyzed data from trials terminated early, enables more detailed investigation of the efficacy and safety of HCQ/CQ among subgroups of hospitalized patients.

Methods

We searched ClinicalTrials.gov in May and June 2020 for US-based RCTs evaluating HCQ/CQ in hospitalized COVID-19 patients in which the outcomes defined in this study were recorded or could be extrapolated. The primary outcome was a 7-point ordinal scale measured between day 28 and 35 post enrollment; comparisons used proportional odds ratios. Harmonized de-identified data were collected via a common template spreadsheet sent to each principal investigator. The data were analyzed by fitting a prespecified Bayesian ordinal regression model and standardizing the resulting predictions.

Results

Eight of 19 trials met eligibility criteria and agreed to participate. Patient-level data were available from 770 participants (412 HCQ/CQ vs 358 control). Baseline characteristics were similar between groups. We did not find evidence of a difference in COVID-19 ordinal scores between days 28 and 35 post-enrollment in the pooled patient population (odds ratio, 0.97; 95% credible interval, 0.76-1.24; higher favors HCQ/CQ), and found no convincing evidence of meaningful treatment effect heterogeneity among prespecified subgroups. Adverse event and serious adverse event rates were numerically higher with HCQ/CQ vs control (0.39 vs 0.29 and 0.13 vs 0.09 per patient, respectively).

Conclusions

The findings of this individual participant data meta-analysis reinforce those of individual RCTs that HCQ/CQ is not efficacious for treatment of COVID-19 in hospitalized patients.

The 1-Year Economic Impact of Work Productivity Loss Following Severe Lower Extremity Trauma

BACKGROUND

Severe lower extremity trauma among working-age adults is highly consequential for returning to work; however, the economic impact attributed to injury has not been fully quantified. The purpose of this study was to examine work and productivity loss during the year following lower extremity trauma and to calculate the economic losses associated with lost employment, lost work time (absenteeism), and productivity loss while at work (presenteeism).

METHODS

This is an analysis of data collected prospectively across 3 multicenter studies of lower extremity trauma outcomes in the United States. Data were used to construct a Markov model that accumulated hours lost over time due to lost employment, absenteeism, and presenteeism among patients from 18 to 64 years old who were working prior to their injury. Average U.S. wages were used to calculate economic loss overall and by sociodemographic and injury subgroups.

RESULTS

Of 857 patients working prior to injury, 47.2% had returned to work at 1 year. The average number of productive hours of work lost was 1,758.8/person, representing 84.6% of expected annual productive hours. Of the hours lost, 1,542.3 (87.7%) were due to working no hours or lost employment, 71.1 (4.0%) were due to missed hours after having returned, and 145.4 (8.3%) were due to decreased productivity while working. The 1-year economic loss due to injury totaled $64,427/patient (95% confidence interval [CI], $63,183 to $65,680). Of the 1,758.8 lost hours, approximately 88% were due to not being employed (working zero hours), 4% were due to absenteeism, and 8% were due to presenteeism. Total productivity loss was higher among older adults (≥40 years), men, those with a physically demanding job, and the most severe injuries (i.e., those leading to amputation as well as Gustilo type-IIIB tibial fractures and type-III pilon/ankle fractures).

CONCLUSIONS

Patients with severe lower extremity trauma carry a substantial economic burden. The costs of lost productivity should be considered when evaluating outcomes.

Discussion on "Causal mediation of semicompeting risks" by Yen-Tsung Huang

Huang proposes a method for assessing the impact of a point treatment on mortality either directly or mediated by occurrence of a nonterminal health event, based on data from a prospective cohort study in which the occurrence of the nonterminal health event may be preemptied by death but not vice versa. The author uses a causal mediation framework to formally define causal quantities known as natural (in)direct effects. The novelty consists of adapting these concepts to a continuous-time modeling framework based on counting processes. In an effort to posit "scientifically interpretable estimands," statistical and causal assumptions are introduced for identification. In this commentary, we argue that these assumptions are not only difficult to interpret and justify, but are also likely violated in the hepatitis B motivating example and other survival/time to event settings as well.

Effect of Intrawound Vancomycin Powder in Operatively Treated High-risk Tibia Fractures: A Randomized Clinical Trial

Importance

Despite the widespread use of systemic antibiotics to prevent infections in surgically treated patients with fracture, high rates of surgical site infection persist.

Objective

To examine the effect of intrawound vancomycin powder in reducing deep surgical site infections.

Design, Setting, and Participants

This open-label randomized clinical trial enrolled adult patients with an operatively treated tibial plateau or pilon fracture who met the criteria for a high risk of infection from January 1, 2015, through June 30, 2017, with 12 months of follow-up (final follow-up assessments completed in April 2018) at 36 US trauma centers.

Interventions

A standard infection prevention protocol with (n = 481) or without (n = 499) 1000 mg of intrawound vancomycin powder.

Main Outcomes and Measures

The primary outcome was a deep surgical site infection within 182 days of definitive fracture fixation. A post hoc comparison assessed the treatment effect on gram-positive and gram-negative-only infections. Other secondary outcomes included superficial surgical site infection, nonunion, and wound dehiscence.

Results

The analysis included 980 patients (mean [SD] age, 45.7 [13.7] years; 617 [63.0%] male) with 91% of the expected person-time of follow-up for the primary outcome. Within 182 days, deep surgical site infection was observed in 29 of 481 patients in the treatment group and 46 of 499 patients in the control group. The time-to-event estimated probability of deep infection by 182 days was 6.4% in the treatment group and 9.8% in the control group (risk difference, -3.4%; 95% CI, -6.9% to 0.1%; P = .06). A post hoc analysis of the effect of treatment on gram-positive (risk difference, -3.7%; 95% CI, -6.7% to -0.8%; P = .02) and gram-negative-only (risk difference, 0.3%; 95% CI, -1.6% to 2.1%; P = .78) infections found that the effect of vancomycin powder was a result of its reduction in gram-positive infections.

Conclusions and Relevance

Among patients with operatively treated tibial articular fractures at a high risk of infection, intrawound vancomycin powder at the time of definitive fracture fixation reduced the risk of a gram-positive deep surgical site infection, consistent with the activity of vancomycin.

Trial Registration

ClinicalTrials.gov Identifier: NCT02227446.

A structured methodology review showed analyses of functional outcomes are frequently limited to "survivors only" in trials enrolling patients at high risk of death

OBJECTIVE

This structured methodology review evaluated statistical approaches used in randomized controlled trials (RCTs) enrolling patients at high risk of death and makes recommendations for reporting future RCTs.

STUDY DESIGN AND SETTING

Using PubMed, we searched for RCTs published in five general medicine journals from January 2014 to August 2019 wherein mortality was ≥10% in at least one randomized group. We abstracted primary and secondary outcomes, statistical analysis methods, and patient samples evaluated (all randomized patients vs. "survivors only").

RESULTS

Of 1947 RCTs identified, 434 met eligibility criteria. Of the eligible RCTs, 91 (21%) and 351 (81%) had a primary or secondary functional outcome, respectively, of which 36 (40%) and 263 (75%) evaluated treatment effects among "survivors only". In RCTs that analyzed all randomized patients, the most common methods included use of ordinal outcomes (e.g., modified Rankin Scale) or creating composite outcomes (primary: 41 of 91 [45%]; secondary: 57 of 351 [16%]).

CONCLUSION

In RCTs enrolling patients at high risk of death, statistical analyses of functional outcomes are frequently conducted among "survivors only," for which conclusions might be misleading. Given the growing number of RCTs conducted among patients hospitalized with COVID-19 and other critical illnesses, standards for reporting should be created.

PREVENTion of CLots in Orthopaedic Trauma (PREVENT CLOT): a randomised pragmatic trial protocol comparing aspirin versus low-molecular-weight heparin for blood clot prevention in orthopaedic trauma patients

INTRODUCTION

Patients who sustain orthopaedic trauma are at an increased risk of venous thromboembolism (VTE), including fatal pulmonary embolism (PE). Current guidelines recommend low-molecular-weight heparin (LMWH) for VTE prophylaxis in orthopaedic trauma patients. However, emerging literature in total joint arthroplasty patients suggests the potential clinical benefits of VTE prophylaxis with aspirin. The primary aim of this trial is to compare aspirin with LMWH as a thromboprophylaxis in fracture patients.

METHODS AND ANALYSIS

PREVENT CLOT is a multicentre, randomised, pragmatic trial that aims to enrol 12 200 adult patients admitted to 1 of 21 participating centres with an operative extremity fracture, or any pelvis or acetabular fracture. The primary outcome is all-cause mortality. We will evaluate non-inferiority by testing whether the intention-to-treat difference in the probability of dying within 90 days of randomisation between aspirin and LMWH is less than our non-inferiority margin of 0.75%. Secondary efficacy outcomes include cause-specific mortality, non-fatal PE and deep vein thrombosis. Safety outcomes include bleeding complications, wound complications and deep surgical site infections.

ETHICS AND DISSEMINATION

The PREVENT CLOT trial has been approved by the ethics board at the coordinating centre (Johns Hopkins Bloomberg School of Public Health) and all participating sites. Recruitment began in April 2017 and will continue through 2021. As both study medications are currently in clinical use for VTE prophylaxis for orthopaedic trauma patients, the findings of this trial can be easily adopted into clinical practice. The results of this large, patient-centred pragmatic trial will help guide treatment choices to prevent VTE in fracture patients.

TRIAL REGISTRATION NUMBER

NCT02984384.

Global sensitivity analysis of randomized trials with nonmonotone missing binary outcomes: Application to studies of substance use disorders

In this paper, we present a method for conducting global sensitivity analysis of randomized trials in which binary outcomes are scheduled to be collected on participants at prespecified points in time after randomization and these outcomes may be missing in a nonmonotone fashion. We introduce a class of missing data assumptions, indexed by sensitivity parameters, which are anchored around the missing not at random assumption introduced by Robins (Statistics in Medicine, 1997). For each assumption in the class, we establish that the joint distribution of the outcomes is identifiable from the distribution of the observed data. Our estimation procedure uses the plug-in principle, where the distribution of the observed data is estimated using random forests. We establish n asymptotic properties for our estimation procedure. We illustrate our methodology in the context of a randomized trial designed to evaluate a new approach to reducing substance use, assessed by testing urine samples twice weekly, among patients entering outpatient addiction treatment. We evaluate the finite sample properties of our method in a realistic simulation study. Our methods have been implemented in an R package entitled slabm.

Methods for Step Count Data: Determining "Valid" Days and Quantifying Fragmentation of Walking Bouts

BACKGROUND

Step count monitors are frequently used in clinical research to measure walking activity. Systematically determining valid days and extracting informative measures of walking beyond total daily step count are among major analytical challenges.

RESEARCH QUESTION

We introduce a novel data-driven anomaly detection algorithm to determine days representing typical walking activity (valid days) and examine the value of measures of walking fragmentation beyond total daily step count.

METHODS

StepWatch data were collected on 230 adults with severe foot or ankle fractures. Average steps per minute (SC), average steps per active minute (SCA), active to sedentary transition probability (ASTP) and sedentary to active transition probability (SATP) were computed for each participant. The joint distribution of these measures was used to identify and eliminate invalid days through a multi-step process based on the support vector machine. The value of SCA, ASTP and SATP beyond SC were assessed by regressing Short Musculoskeletal Functional Assessment (SMFA), a measure of self-reported function, on these measures and quantifying the increase in the adjusted R-squared. In an unsupervised comparison, the total joint variability of SCA, ASTP and SATP was decomposed into the variability explained by SC and the unique variability of these three measures.

RESULTS

Of the 4,448 days in the original data set, 39% were determined invalid. Individuals with higher average SC had higher SCA, lower ASTP and higher SATP. Measures of fragmentation (SCA, ASTP and SATP) explained 25% more of the variability in SMFA compared with SC alone. Approximately 41% of the variability in SCA, ASTP and SATP could not be explained by SC suggesting that these three measures provide unique information about walking patterns.

SIGNIFICANCE

Applying SVM and quantifying fragmentation in walking bouts for step count data can help to more precisely assess activity in clinical studies employing this modality.

idem: An R Package for Inferences in Clinical Trials with Death and Missingness

In randomized controlled trials of seriously ill patients, death is common and often defined as the primary endpoint. Increasingly, non-mortality outcomes such as functional outcomes are co-primary or secondary endpoints. Functional outcomes are not defined for patients who die, referred to as "truncation due to death", and among survivors, functional outcomes are often unobserved due to missed clinic visits or loss to follow-up. It is well known that if the functional outcomes "truncated due to death" or missing are handled inappropriately, treatment effect estimation can be biased. In this paper, we describe the package idem that implements a procedure for comparing treatments that is based on a composite endpoint of mortality and the functional outcome among survivors. Among survivors, the procedure incorporates a missing data imputation procedure with a sensitivity analysis strategy. A web-based graphical user interface is provided in the idem package to facilitate users conducting the proposed analysis in an interactive and user-friendly manner. We demonstrate idem using data from a recent trial of sedation interruption among mechanically ventilated patients.

Global identifiability of latent class models with applications to diagnostic test accuracy studies: A Gröbner basis approach

Identifiability of statistical models is a fundamental regularity condition that is required for valid statistical inference. Investigation of model identifiability is mathematically challenging for complex models such as latent class models. Jones et al. used Goodman's technique to investigate the identifiability of latent class models with applications to diagnostic tests in the absence of a gold standard test. The tool they used was based on examining the singularity of the Jacobian or the Fisher information matrix, in order to obtain insights into local identifiability (ie, there exists a neighborhood of a parameter such that no other parameter in the neighborhood leads to the same probability distribution as the parameter). In this paper, we investigate a stronger condition: global identifiability (ie, no two parameters in the parameter space give rise to the same probability distribution), by introducing a powerful mathematical tool from computational algebra: the Gröbner basis. With several existing well-known examples, we argue that the Gröbner basis method is easy to implement and powerful to study global identifiability of latent class models, and is an attractive alternative to the information matrix analysis by Rothenberg and the Jacobian analysis by Goodman and Jones et al.

A constructive critique of the draft ICH E9 Addendum

In this article, I review the key elements of the proposed International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use E9 Addendum, present a constructive critique, and provide recommendations of how it can be improved. To highlight ideas, I present a case study involving a confirmatory trial for a chronic pain medication.

Multisite Evaluation of a Custom Energy-Storing Carbon Fiber Orthosis for Patients with Residual Disability After Lower-Limb Trauma

BACKGROUND

The Intrepid Dynamic Exoskeletal Orthosis (IDEO) is a custom energy-storing carbon fiber ankle-foot orthosis developed for lower-extremity trauma patients. Studies conducted at the military treatment facility where the IDEO was developed demonstrated benefits of the IDEO when used with the Return to Run Physical Therapy (RTR PT) program. The current study was designed to determine if results could be replicated at other military treatment facilities and to examine whether early performance gains in patient-reported functional outcomes remained at 12 months.

METHODS

Study participants included service members who had functional deficits that interfered with daily activities at least 1 year after a traumatic unilateral lower-extremity injury at or below the knee. Participants were evaluated before receiving the IDEO, immediately following completion of RTR PT, and at 6 and 12 months. Agility, strength/power, and speed were assessed using well-established performance tests. Self-reported function was measured using the Short Musculoskeletal Function Assessment (SMFA). The Orthotics and Prosthetics Users' Survey was administered to assess satisfaction with the IDEO. Of 87 participants with complete baseline data, 6 did not complete any physical therapy and were excluded from the analysis. Follow-up rates immediately following completion of the RTR PT and at 6 and 12 months were 88%, 75%, and 79%, respectively.

RESULTS

Compared with baseline, improvement at completion of RTR PT was observed in all but 1 performance test. SMFA scores for all domains except hand and arm function were lower (improved function) at 6 and 12 months. Satisfaction with the IDEO was high following completion of RTR PT, with some attenuation at the time of follow-up.

CONCLUSIONS

This study adds to the evidence supporting the efficacy of the IDEO coupled with RTR PT. However, despite improvement in both performance and self-reported functioning, deficits persist compared with population norms.

LEVEL OF EVIDENCE

Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.

Global sensitivity analysis of clinical trials with missing patient-reported outcomes

Randomized trials with patient-reported outcomes are commonly plagued by missing data. The analysis of such trials relies on untestable assumptions about the missing data mechanism. To address this issue, it has been recommended that the sensitivity of the trial results to assumptions should be a mandatory reporting requirement. In this paper, we discuss a recently developed methodology (Scharfstein et al., Biometrics, 2018) for conducting sensitivity analysis of randomized trials in which outcomes are scheduled to be measured at fixed points in time after randomization and some subjects prematurely withdraw from study participation. The methodology is explicated in the context of a placebo-controlled randomized trial designed to evaluate a treatment for bipolar disorder. We present a comprehensive data analysis and a simulation study to evaluate the performance of the method. A software package entitled SAMON (R and SAS versions) that implements our methods is available at www.missingdatamatters.org .

Military and Civilian Collaboration: The Power of Numbers

The purpose of this study was to compare the number and types of extremity injuries treated at civilian trauma centers (CIV CENs) versus military treatment facilities (MTFs) participating in the Major Extremity Trauma Research Consortium (METRC) and to investigate the potential benefits of a clinical research network that includes both civilian trauma centers and MTFs. Two analyses were performed. First, registry data collected on all surgically treated fractures at four core MTFs and 21 CIV CENs over one year were compared. Second, actual numbers and distribution of patients by type of injury enrolled in three METRC studies were compared. While MTFs demonstrated higher percentages of severe injuries including open fractures, traumatic amputations, vascular injuries, contamination, and injuries with bone, muscle, and skin loss when compared to CIV CENS, the CIV CENs treated a substantially higher number and, more importantly, enrolled patients in almost all categories. Comparison of service members to civilians was challenged by several differences between the two patient populations including mechanism of injury, the medical care environment, and confounding factors such as age, social setting and co-morbidities. Despite these limitations, in times without active military conflict, clinical trials will likely rely on civilian trauma centers for patient enrollment; only when numbers are pooled across a large number of centers can requisite sample sizes be met. These data demonstrate the benefits of maintaining a military-civilian partnership to address the major gaps in research defined by the Military.

Statistical methods to compare functional outcomes in randomized controlled trials with high mortality

Mortality is a common primary endpoint in randomized controlled trials of patients with a high severity of illness, such as critically ill patients. However, researchers are increasingly evaluating functional outcomes, such as quality of life. Importantly, in such trials some patients may die before the assessment of a functional outcome, resulting in the functional outcome being “truncated due to death.” As described in this paper, defining and testing treatment effects on functional outcomes in this setting requires careful consideration. Data from a completed trial of critically ill patients are used to highlight key differences among three statistical approaches used when analyzing such trials.

Predicting Acute Compartment Syndrome (PACS): The Role of Continuous Monitoring

The diagnosis of acute compartment syndrome (ACS) is a common clinical challenge among patients who sustain high-energy orthopaedic trauma, largely because no validated criteria exist to reliably define the presence of the condition. In the absence of validated diagnostic standards, concern for the potential clinical and medicolegal impact of a missed compartment syndrome may result in the potential overuse of fasciotomy in "at-risk" patients. The goal of the Predicting Acute Compartment Syndrome Study was to develop a decision rule for predicting the likelihood of ACS that would reduce unnecessary fasciotomies while guarding against potentially missed ACS. Of particular interest was the utility of early and continuous monitoring of intramuscular pressure and muscle oxygenation using near-infrared spectroscopy in the timely diagnosis of ACS. In this observational study, 191 participants aged 18-60 with high-energy tibia fractures were prospectively enrolled and monitored for up to 72 hours after admission, then followed for 6 months. Treating physicians were blinded to continuous pressure and oxygenation data. An expert panel of 9 orthopaedic surgeons retrospectively assessed the likelihood that each patient developed ACS based on data collected on initial presentation, clinical course, and known functional outcome at 6 months. This retrospectively assigned likelihood is modeled as a function of clinical data typically available within 72 hours of admission together with continuous pressure and oxygenation data. This study will improve our understanding of the natural history of compartment syndrome and examine the utility of early and continuous monitoring of the physiologic status of the injured extremity in the timely diagnosis of ACS.

Inference in randomized trials with death and missingness

In randomized studies involving severely ill patients, functional outcomes are often unobserved due to missed clinic visits, premature withdrawal, or death. It is well known that if these unobserved functional outcomes are not handled properly, biased treatment comparisons can be produced. In this article, we propose a procedure for comparing treatments that is based on a composite endpoint that combines information on both the functional outcome and survival. We further propose a missing data imputation scheme and sensitivity analysis strategy to handle the unobserved functional outcomes not due to death. Illustrations of the proposed method are given by analyzing data from a recent non-small cell lung cancer clinical trial and a recent trial of sedation interruption among mechanically ventilated patients.

Estimation of treatment effects in matched-pair cluster randomized trials by calibrating covariate imbalance between clusters

We address estimation of intervention effects in experimental designs in which (a) interventions are assigned at the cluster level; (b) clusters are selected to form pairs, matched on observed characteristics; and (c) intervention is assigned to one cluster at random within each pair. One goal of policy interest is to estimate the average outcome if all clusters in all pairs are assigned control versus if all clusters in all pairs are assigned to intervention. In such designs, inference that ignores individual level covariates can be imprecise because cluster-level assignment can leave substantial imbalance in the covariate distribution between experimental arms within each pair. However, most existing methods that adjust for covariates have estimands that are not of policy interest. We propose a methodology that explicitly balances the observed covariates among clusters in a pair, and retains the original estimand of interest. We demonstrate our approach through the evaluation of the Guided Care program.

A matched-pair cluster-randomized trial of guided care for high-risk older patients

BACKGROUND

Patients at risk for generating high health care expenditures often receive fragmented, low-quality, inefficient health care. Guided Care is designed to provide proactive, coordinated, comprehensive care for such patients.

OBJECTIVE

We hypothesized that Guided Care, compared to usual care, produces better functional health and quality of care, while reducing the use of expensive health services.

DESIGN

32-month, single-blind, matched-pair, cluster-randomized controlled trial of Guided Care, conducted in eight community-based primary care practices.

PATIENTS

The "Hierarchical Condition Category" (HCC) predictive model was used to identify high-risk older patients who were insured by fee-for-service Medicare, a Medicare Advantage plan or Tricare. Patients with HCC scores in the highest quartile (at risk for generating high health care expenditures during the coming year) were eligible to participate.

INTERVENTION

A registered nurse collaborated with two to five primary care physicians in providing eight services to participants: comprehensive assessment, evidence-based care planning, proactive monitoring, care coordination, transitional care, coaching for self-management, caregiver support, and access to community-based services.

MAIN MEASURES

Functional health was measured using the Short Form-36. Quality of care and health services utilization were measured using the Patient Assessment of Chronic Illness Care and health insurance claims, respectively.

KEY RESULTS

Of the eligible patients, 904 (37.8 %) gave written consent to participate; of these, 477 (52.8 %) completed the final interview, and 848 (93.8 %) provided complete claims data. In intention-to-treat analyses, Guided Care did not significantly improve participants' functional health, but it was associated with significantly higher participant ratings of the quality of care (difference = 0.27, 95 % CI = 0.08-0.45) and 29 % lower use of home care (95 % CI = 3-48 %).

CONCLUSIONS

Guided Care improves high-risk older patients' ratings of the quality of their care, and it reduces their use of home care, but it does not appear to improve their functional health.

Observational studies in the era of randomized trials: finding the balance

Randomized controlled trials (RCTs) constitute the gold standard for the generation of evidence-based medicine, but may not always be feasible. Furthermore, randomization alone does not guarantee the utility of the research, as evidenced by thousands of uninformative RCTs documented in the literature. Observational studies, including longitudinal, retrospective, and case-control designs, can contribute to the body of evidence in meaningful ways, provide useful information when an RCT is unethical or not feasible, generate hypotheses for RCTs, or provide preliminary work to better inform design of future RCTs. They can also be used to study rare outcomes, risk factors, and side effects, and to examine whether results from RCTs translate into effective treatment in routine practice. Use of modern statistical techniques, both in the study design and in the analysis stage, can improve the usefulness of the evidence obtained from observational studies.

On the prevention and analysis of missing data in randomized clinical trials: the state of the art

We summarize and elaborate on the recently published National Research Council report entitled "The Prevention and Treatment of Missing Data in Clinical Trials." We tailor our discussion to orthopaedic trials. In particular, we discuss the intent-to-treat principle, review study design and prevention ideas to minimize missing data, and present state-of-the-art sensitivity analysis methods for analyzing and reporting the results of studies with missing data.

A Bayesian Shrinkage Model for Incomplete Longitudinal Binary Data with Application to the Breast Cancer Prevention Trial

We consider inference in randomized longitudinal studies with missing data that is generated by skipped clinic visits and loss to follow-up. In this setting, it is well known that full data estimands are not identified unless unverified assumptions are imposed. We assume a non-future dependence model for the drop-out mechanism and partial ignorability for the intermittent missingness. We posit an exponential tilt model that links non-identifiable distributions and distributions identified under partial ignorability. This exponential tilt model is indexed by non-identified parameters, which are assumed to have an informative prior distribution, elicited from subject-matter experts. Under this model, full data estimands are shown to be expressed as functionals of the distribution of the observed data. To avoid the curse of dimensionality, we model the distribution of the observed data using a Bayesian shrinkage model. In a simulation study, we compare our approach to a fully parametric and a fully saturated model for the distribution of the observed data. Our methodology is motivated by, and applied to, data from the Breast Cancer Prevention Trial.

The effect of guided care teams on the use of health services: results from a cluster-randomized controlled trial

BACKGROUND

The effect of interdisciplinary primary care teams on the use of health services by patients with multiple chronic conditions is uncertain. This study aimed to measure the effect of guided care teams on multimorbid older patients' use of health services.

METHODS

Eligible patients from 3 health care systems in the Baltimore, Maryland-Washington, DC, area were cluster-randomized to receive guided care or usual care for 20 months between November 1, 2006, and June 30, 2008. Eight services of a guided care nurse working in partnership with patients' primary care physicians were provided: comprehensive assessment, evidence-based care planning, monthly monitoring of symptoms and adherence, transitional care, coordination of health care professionals, support for self-management, support for family caregivers, and enhanced access to community services. Outcome measures were frequency of use of emergency departments, hospitals, skilled nursing facilities, home health agencies, primary care physician services, and specialty physician services.

RESULTS

The study included 850 older patients at high risk for using health care heavily in the future. The only statistically significant overall effect of guided care in the whole sample was a reduction in episodes of home health care (odds ratio, 0.70; 95% confidence interval, 0.53-0.93). In a preplanned analysis, guided care also reduced skilled nursing facility admissions (odds ratio, 0.53; 95% confidence interval, 0.31-0.89) and days (0.48; 0.28-0.84) among Kaiser-Permanente patients.

CONCLUSIONS

Guided care reduces the use of home health care but has little effect on the use of other health services in the short run. Its positive effect on Kaiser-Permanente patients' use of skilled nursing facilities and other health services is intriguing. Trial Registration clinicaltrials.gov Identifier: NCT00121940.

On estimation of the survivor average causal effect in observational studies when important confounders are missing due to death

SUMMARY

We focus on estimation of the causal effect of treatment on the functional status of individuals at a fixed point in time t* after they have experienced a catastrophic event, from observational data with the following features: (i) treatment is imposed shortly after the event and is nonrandomized, (ii) individuals who survive to t* are scheduled to be interviewed, (iii) there is interview nonresponse, (iv) individuals who die prior to t* are missing information on preevent confounders, and (v) medical records are abstracted on all individuals to obtain information on postevent, pretreatment confounding factors. To address the issue of survivor bias, we seek to estimate the survivor average causal effect (SACE), the effect of treatment on functional status among the cohort of individuals who would survive to t* regardless of whether or not assigned to treatment. To estimate this effect from observational data, we need to impose untestable assumptions, which depend on the collection of all confounding factors. Because preevent information is missing on those who die prior to t*, it is unlikely that these data are missing at random. We introduce a sensitivity analysis methodology to evaluate the robustness of SACE inferences to deviations from the missing at random assumption. We apply our methodology to the evaluation of the effect of trauma center care on vitality outcomes using data from the National Study on Costs and Outcomes of Trauma Care.

Guided care and the cost of complex healthcare: a preliminary report

OBJECTIVE

Guided Care (GC) is a model of proactive, evidence-based comprehensive healthcare provided by physician-nurse teams for people with several chronic health conditions. Our objective was to evaluate the preliminary effects of GC on health service utilization and costs.

STUDY DESIGN

Cluster-randomized controlled trial of GC involving 14 primary care teams (49 physicians) and 904 of their chronically ill patients age 65 years or older.

METHODS

Using insurance claims, we compared the health services used by patients who received GC with the health services used by patients who received usual care during the first 8 months of the study.

RESULTS

After adjustment for baseline characteristics, GC patients experienced, on average, 24% fewer hospital days (95% confidence interval [CI]: 49% fewer, 13% more), 37% fewer skilled nursing facility days (95% CI: 65% fewer, 5% more), 15% fewer emergency department visits (95% CI: 38% fewer, 18% more), and 29% fewer home healthcare episodes (95% CI: 53% fewer, 8% more), as well as 9% more specialist visits (95% CI: 8% fewer, 29% more). Based on current Medicare payment rates and GC costs, these differences in utilization represent an annual net savings of $75,000 (95% CI: -$244,000, $150,900) per nurse, or $1364 per patient.

CONCLUSIONS

Initial introduction of GC into primary care practices may be associated with less use of expensive health services and a net savings in healthcare costs among older patients with several chronic health conditions. Final results from the remaining 2 years of this ongoing study will be published in 2011.

Inference for cumulative incidence functions with informatively coarsened discrete event-time data

We consider the problem of comparing cumulative incidence functions of non-mortality events in the presence of informative coarsening and the competing risk of death. We extend frequentist-based hypothesis tests previously developed for non-informative coarsening and propose a novel Bayesian method based on comparing a posterior parameter transformation with its expected distribution under the null hypothesis of equal cumulative incidence functions. Both methods use estimates derived by extending previously published estimation procedures to accommodate censoring by death. The data structure and analysis goal are exemplified by the AIDS Link to the Intravenous Experience (ALIVE) study, where researchers are interested in comparing incidence of human immunodeficiency virus seroconversion by risk behavior categories. Coarsening in the forms of interval and right censoring and censoring by death in ALIVE is thought to be informative; thus, we perform a sensitivity analysis by incorporating elicited expert information about the relationship between seroconversion and censoring into the model.

Sensitivity analysis using elicited expert information for inference with coarsened data: illustration of censored discrete event times in the AIDS Link to Intravenous Experience (ALIVE) Study

In this paper, the authors use the rubric of "coarsened data," of which missing and censored data are special cases, to motivate the elicitation and use of expert information for performing sensitivity analyses of censored event-time data. Elicited information is important because observed data are insufficient to estimate how study participants with coarsened data compare with participants with uncoarsened data, and misspecifying this comparison may produce biased analysis results. In the presence of coarsening, performing a sensitivity analysis over a range of plausible assumptions is the best one can do. Here the authors illustrate an approach for eliciting expert information for use in sensitivity analyses to compare cumulative incidence functions of censored nonmortality outcomes. An example of such data is the AIDS Link to Intravenous Experience (ALIVE) Study, where the authors aim to estimate and compare cumulative incidence functions for human immunodeficiency virus between risk factor categories. The interval and right-censoring and censoring due to death found in the ALIVE data (1988-1998) are thought to be informative; thus, a sensitivity analysis is performed using information elicited from 2 ALIVE scientists and an expert in acquired immunodeficiency syndrome epidemiology about the relation between seroconversion and censoring.

Accounting for within-patient correlation in assessing relative sensitivity of an adjunctive diagnostic test: application to lung cancer

We address the comparison of results between two diagnostic tests applied multiple times to the same subjects. The estimand of interest is the sensitivity of the combined test (primary and adjunct) relative to a primary test. Analytical methods are first described that assume independence between the multiple observations within a subject. In order to account for the within-subject correlation introduced by the multiple measurements, analytical approaches for correlated, categorical response data are described. In the discussion of these methods, we pay particular attention to the presence of a structural zero which results from the decision rule for the combination of diagnostic tests. In a simulation study, we compare the finite sample performances of all analytical approaches in terms of confidence interval coverage rates and median lengths. Our methods are cast in the context of a diagnostic bronchoscopy technology for the detection of lung cancer.

A longitudinal study of vaginal douching and bacterial vaginosis--a marginal structural modeling analysis

The etiology of bacterial vaginosis is unknown, and there are no long-term therapies for preventing this frequently recurring condition. Vaginal douching has been reported to be associated with bacterial vaginosis in observational studies. However, this association may be due to confounding by indication--that is, confounding by women douching in response to vaginal symptoms associated with bacterial vaginosis. The authors used marginal structural modeling to estimate the causal effect of douching on bacterial vaginosis risk while controlling for this confounding effect. In 1999-2002, nonpregnant women (n = 3,620) were recruited into a prospective study when they visited one of 12 public health clinics in Birmingham, Alabama, for routine care. Participants were assessed quarterly for 1 year. Bacterial vaginosis was based on a Nugent's Gram stain score of 7 or higher. Thirty-two percent of participants douched in every study interval, and 43.0% never douched. Of the 12,349 study visits, 40.2% were classified as involving bacterial vaginosis. The relative risk for regular douching as compared with no douching was 1.21 (95% confidence interval: 1.08, 1.38). These findings indicate that douching confers increased risk of disruption of vaginal flora. In the absence of a large randomized trial, these findings provide the best evidence to date for a risk of bacterial vaginosis associated with douching.

The effect of vaginal douching cessation on bacterial vaginosis: a pilot study

OBJECTIVE

The objective of the study was to evaluate the risk for bacterial vaginosis (BV) in a douching cessation trial.

STUDY DESIGN

Thirty-nine reproductive-age women who reported use of douche products were enrolled into a 20-week study consisting of a 4 week douching observation (phase I) followed by 12-weeks of douching cessation (phase II). In phase III, participants then chose to resume douching or continue cessation for the remaining 4 weeks. Self-collected vaginal samples were obtained twice weekly in the first 16 weeks, and 1 sample was collected during week 20 (1107 samples total). BV was diagnosed by Nugent score of 7 or greater. Conditional logistic regression was used to evaluate douching cessation on the risk of BV.

RESULTS

The adjusted odds ratio (aOR) for BV in the douching cessation phase, as compared with the douching-observation phase was 0.76 (95% confidence interval [CI], 0.33 to 1.76). Among women who reported their primary reason for douching was to cleanse after menstruation, BV was significantly reduced in douching cessation (aOR:0.23; 95% CI, 0.12 to 0.44).

CONCLUSION

Vaginal douching cessation may reduce the risk for BV in a subset of women.

The impact of trauma-center care on functional outcomes following major lower-limb trauma

BACKGROUND

Although studies have shown that treatment at a trauma center reduces a patient's risk of dying following major trauma, important questions remain as to the effect of trauma centers on functional outcomes, especially among patients who have sustained major lower-limb trauma.

METHODS

Domain-specific scores on the Medical Outcomes Study Short Form Health Survey (SF-36) supplemented by scores on the mobility subscale of the Musculoskeletal Function Assessment (MFA) and the Revised Center for Epidemiologic Studies Depression Scale (CESD-R) were compared among patients treated in eighteen hospitals with a level-I trauma center and fifty-one hospitals without a trauma center. Included in the study were 1389 adults, eighteen to eighty-four years of age, with at least one lower-limb injury with a score of >/=3 points according to the Abbreviated Injury Scale (AIS). To account for the competing risk of death, we estimated the survivors' average causal effect. Estimates were derived for all patients with a lower-limb injury and separately for a subset of patients without associated injuries of the head or spinal cord.

RESULTS

For patients with a lower-limb injury resulting from a high-energy force, care at a trauma center yielded modest but clinically meaningful improvements in physical functioning and overall vitality at one year after the injury. After adjustment for differences in case mix and the competing risk of death, the average differences in the SF-36 physical functioning and vitality scores and the MFA mobility score were 7.82 points (95% confidence interval: 2.65, 12.98), 6.80 points (95% confidence interval: 2.53, 11.07), and 6.31 points (95% confidence interval: 0.25, 12.36), respectively. These results were similar when the analysis was restricted to patients without associated injuries to the head or spine. Treatment at a trauma center resulted in negligible differences in outcome for the subset of patients with injuries resulting from low-energy forces.

CONCLUSIONS

This study provides evidence that patients who sustain high-energy lower-limb trauma benefit from treatment at a level-I trauma center.

Prevalence and incidence of depressive disorder: the Baltimore ECA follow-up, 1981-2004

OBJECTIVE

To describe trends in prevalence and incidence of depressive disorder in a cohort from Eastern Baltimore.

METHOD

Twenty-three-year-old longitudinal cohort, the Baltimore Epidemiologic Catchment Area Follow-up. Participants were selected probabilistically from the household population in 1981, and interviewed in 1981, 1993, and 2004. Diagnoses were made via the Diagnostic Interview Schedule according to successive editions of the American Psychiatric Association Diagnostic and Statistical Manual.

RESULTS

Older age, lower education, non-White race, and cognitive impairment are independent predictors of attrition due to death and loss of contact, but depressive disorder is not related to attrition. Prevalence rates rise for females between 1981, 1993, and 2004. Incidence rates in the period 1993-2004 are lower than the period 1981-1993, suggesting the rise in prevalence is due to increasing chronicity.

CONCLUSION

There has been a rise in the prevalence of depression in the prior quarter century among middle-aged females.

Healthy Steps for Young Children: sustained results at 5.5 years

OBJECTIVE

We sought to determine whether Healthy Steps for Young Children has sustained treatment effects at 5.5 years, given early findings demonstrating enhanced quality of care and improvements in selected parenting practices.

METHODS

Healthy Steps was a clinical trial that incorporated developmental specialists and enhanced developmental services into pediatric care in the first 3 years of life. A total of 5565 children were enrolled at birth and followed through 5.5 years. Healthy Steps was evaluated at 6 randomization and 9 quasi-experimental sites. Computer-assisted telephone interviews were conducted with mothers when Healthy Steps children were 5.5 years of age. Outcomes included experiences seeking care, parent response to child misbehavior, perception of child's behavior, and parenting practices to promote development and safety. Logistic regression was used to estimate overall effects of Healthy Steps, adjusting for site and baseline demographic characteristics.

RESULTS

A total of 3165 (56.9%) families responded to interviews (usual care: n = 1441; Healthy Steps: n = 1724). Families that had received Healthy Steps services were more satisfied with care (agreed that pediatrician/nurse practitioner provided support, 82.0% vs 79.0%; odds ratio: 1.25 [95% confidence interval: 1.02-1.53]) and more likely to receive needed anticipatory guidance (54.9% vs 49.2%; odds ratio: 1.33 [95% confidence interval: 1.13-1.57]) (all P < .05). They also had increased odds of remaining at the original practice (65.1% vs 61.4%; odds ratio: 1.19 [95% confidence interval: 1.01-1.39]). Healthy Steps families reported reduced odds of using severe discipline (slap in face/spank with object, 10.1% vs 14.1%; odds ratio: 0.68 [95% confidence interval: 0.54-0.86]) and increased odds of often/almost always negotiating with their child (59.8% vs 56.3%; odds ratio: 1.20 [95% confidence interval: 1.03-1.39]). They had greater odds of reporting a clinical or borderline concern regarding their child's behavior (18.1% vs 14.8%; odds ratio: 1.35 [95% confidence interval: 1.10-1.64]) and their child reading books (59.4% vs 53.6%; odds ratio: 1.16 [95% confidence interval: 1.00-1.35]). There were no effects on safety practices.

CONCLUSIONS

Sustained treatment effects, albeit modest, are consistent with early findings. Universal, practice-based interventions can enhance quality of care for families with young children and can improve selected parenting practices beyond the duration of the intervention.

Survival curve estimation for informatively coarsened discrete event-time data

Interval-censored, or more generally, coarsened event-time data arise when study participants are observed at irregular time periods and experience the event of interest in between study observations. Such data are often analysed assuming non-informative censoring, which can produce biased results if the assumption is wrong. This paper extends the standard approach for estimating survivor functions to allow informatively interval-censored data by incorporating various assumptions about the censoring mechanism into the model. We include a Bayesian extension in which final estimates are produced by mixing over a distribution of assumed censoring mechanisms. We illustrate these methods with a natural history study of HIV-infected individuals using assumptions elicited from an AIDS expert.

The impact of an intensivist-model ICU on trauma-related mortality

OBJECTIVE

To evaluate the effect of an intensivist-model of critical care delivery on the risk of death following injury.

SUMMARY BACKGROUND DATA

An intensivist-model of ICU care is associated with improved outcomes and less resource utilization in mixed medical and surgical ICUs. The process of trauma center verification assures a relatively high standard of care and quality assurance; thus, it is unclear what the effect of a specific model of ICU care delivery might have on trauma-related mortality.

METHODS

Using data from a large multicenter (68 centers) prospective cohort study, we evaluated the relationship between the model of ICU care (open vs. intensivist-model) and in-hospital mortality following severe injury. An intensivist-model was defined as an ICU where critically ill trauma patients were either on a distinct ICU service (led by an intensivist) or were comanaged with an intensivist (a physician board-certified in critical care).

RESULTS

After adjusting for differences in baseline characteristics, the relative risk of death in intensivist-model ICUs was 0.78 (0.58-1.04) compared with an open ICU model. The effect was greatest in the elderly [RR, 0.55 (0.39-0.77)], in units led by surgical intensivists [RR, 0.67 (0.50-0.90)], and in designated trauma centers 0.64 (0.46-0.88).

CONCLUSIONS

Care in an intensivist-model ICU is associated with a large reduction in in-hospital mortality following trauma, particularly in elderly patients who might have limited physiologic reserve and extensive comorbidity. That the effect is greatest in trauma centers and in units led by surgical intensivists suggests the importance of content expertise in the care of the critically injured. Injured patients are best cared for using an intensivist-model of dedicated critical care delivery, a criterion that should be considered in the verification of trauma centers.

Causal inference for non-mortality outcomes in the presence of death

Evaluation of the causal effect of a baseline exposure on a morbidity outcome at a fixed time point is often complicated when study participants die before morbidity outcomes are measured. In this setting, the causal effect is only well defined for the principal stratum of subjects who would live regardless of the exposure. Motivated by gerontologic researchers interested in understanding the causal effect of vision loss on emotional distress in a population with a high mortality rate, we investigate the effect among those who would live both with and without vision loss. Since this subpopulation is not readily identifiable from the data and vision loss is not randomized, we introduce a set of scientifically driven assumptions to identify the causal effect. Since these assumptions are not empirically verifiable, we embed our methodology within a sensitivity analysis framework. We apply our method using the first three rounds of survey data from the Salisbury Eye Evaluation, a population-based cohort study of older adults. We also present a simulation study that validates our method.

Estimating causal effects from multiple cycle data in studies of in vitro fertilization

Prospective studies of reproductive outcomes frequently record data at multiple cycles. For example, studies of in vitro fertilization and embryo transfer (IVF-ET) follow women or couples for possibly several IVF cycles and record outcomes such as pregnancy status and embryo implantation. Several time-varying covariates, such as age and diagnostic markers, typically are available as well. When attention is focused on measurement of exposure effects, the use of multiple cycle data poses several complications. If the study is observational, the exposure probability may depend on subject characteristics. Moreover, attrition rates in IVF-ET can be substantial, and the attrition process can be expected to depend heavily on prior outcome. In fact, both success (pregnancy) and failure (lack of embryo implantations) can be prognostic of dropout. In this paper, we illustrate the use of causal modeling for multiple cycle data. Key assumptions are reviewed, and inference based on weighted estimating equations is described in detail. The methods are applied to a study of the effects of hydrosalpinx among women with tubal disease undergoing IVF-ET.

On estimation of vaccine efficacy using validation samples with selection bias

Using validation sets for outcomes can greatly improve the estimation of vaccine efficacy (VE) in the field (Halloran and Longini, 2001; Halloran and others, 2003). Most statistical methods for using validation sets rely on the assumption that outcomes on those with no cultures are missing at random (MAR). However, often the validation sets will not be chosen at random. For example, confirmational cultures are often done on people with influenza-like illness as part of routine influenza surveillance. VE estimates based on such non-MAR validation sets could be biased. Here we propose frequentist and Bayesian approaches for estimating VE in the presence of validation bias. Our work builds on the ideas of Rotnitzky and others (1998, 2001), Scharfstein and others (1999, 2003), and Robins and others (2000). Our methods require expert opinion about the nature of the validation selection bias. In a re-analysis of an influenza vaccine study, we found, using the beliefs of a flu expert, that within any plausible range of selection bias the VE estimate based on the validation sets is much higher than the point estimate using just the non-specific case definition. Our approach is generally applicable to studies with missing binary outcomes with categorical covariates.

A national evaluation of the effect of trauma-center care on mortality

BACKGROUND

Hospitals have difficulty justifying the expense of maintaining trauma centers without strong evidence of their effectiveness. To address this gap, we examined differences in mortality between level 1 trauma centers and hospitals without a trauma center (non-trauma centers).

METHODS

Mortality outcomes were compared among patients treated in 18 hospitals with a level 1 trauma center and 51 hospitals non-trauma centers located in 14 states. Patients 18 to 84 years old with a moderate-to-severe injury were eligible. Complete data were obtained for 1104 patients who died in the hospital and 4087 patients who were discharged alive. We used propensity-score weighting to adjust for observable differences between patients treated at trauma centers and those treated at non-trauma centers.

RESULTS

After adjustment for differences in the case mix, the in-hospital mortality rate was significantly lower at trauma centers than at non-trauma centers (7.6 percent vs. 9.5 percent; relative risk, 0.80; 95 percent confidence interval, 0.66 to 0.98), as was the one-year mortality rate (10.4 percent vs. 13.8 percent; relative risk, 0.75; 95 percent confidence interval, 0.60 to 0.95). The effects of treatment at a trauma center varied according to the severity of injury, with evidence to suggest that differences in mortality rates were primarily confined to patients with more severe injuries.

CONCLUSIONS

Our findings show that the risk of death is significantly lower when care is provided in a trauma center than in a non-trauma center and argue for continued efforts at regionalization.

Utility of saturation biopsy to predict insignificant cancer at radical prostatectomy

OBJECTIVES

To determine whether potential candidates for watchful waiting have undersampling of more substantial cancer.

METHODS

A total of 103 men were studied, who were predicted to have insignificant cancer in their radical prostatectomy (RP) specimen. All had limited cancer on routine needle biopsy (no core with more than 50% involvement; Gleason score less than 7, and fewer than 3 cores involved) with a serum prostate-specific antigen density of 0.15 or less. Insignificant tumor at RP was considered organ-confined tumor, no Gleason pattern 4 or 5, and a tumor volume of less than 0.5 cm3. Saturation biopsy (average 44 cores) and an alternate biopsy saturation scheme with one half the number of cores using an 18-gauge Biopty gun was performed in the pathology laboratory on totally embedded and serially sectioned RP specimens.

RESULTS

Of the tumors, 97% were organ confined. The RP Gleason score was less than 7 in 84% of the cases. The RP tumor volume was 0.01 to 2.39 cm3 (median 0.14). Of the cancer specimens, 71% were insignificant and 29% had been incorrectly classified before surgery using standard biopsy schemes. Using the full saturation biopsy scheme, if we predicted significant cancer, the probability of having insignificant cancer was only 11.5% (false-positive rate). If the model predicted insignificant cancer, the probability of significant cancer was also only 11.5% (false-negative rate; sensitivity 71.9% and specificity 95.8%). Using the alternate biopsy sampling scheme, the false-positive rate was 8% and the false-negative rate was 11.4% (sensitivity 71.9% and specificity 97.1%).

CONCLUSIONS

Saturation biopsy provides accurate predictability of prostate tumor volume and grade to select suitable candidates for watchful waiting therapy.