A simulation study on implementing marginal structural models in an observational study with switching medication based on a biomarker.

Assessing treatment effectiveness in longitudinal data can be complex when treatments are not randomly assigned and patients are allowed to switch treatment to other or no treatment, often in a manner that is driven by changes in one or more variables associated with patient or clinical characteristics. There can be confounding of the treatment effect from a time-varying variable, i.e., one which is affected by previous exposure and can in turn also influence subsequent treatment changes. Precision medicine relies on validated biomarkers to better classify patients by their probable response to treatment. However, biomarkers may be a source of time-varying confounding, which are affected by prior treatment in the evaluation and are also subject to measurement errors. The impact of switching medications based on a biomarker has received less attention. We conducted simulation studies to explore biased estimation under various scenarios when marginal structural model estimations are employed. Holding model misspecification issues constant, bias is severe in the presence of multiple switching, along with measurement error and missing data in the covariates.

Predictors of blood pressure response to angiotensin receptor blocker/diuretic combination therapy: a secondary analysis of the irbesartan/hydrochlorothiazide blood pressure reductions in diverse patient populations (INCLUSIVE) study.

The secondary analysis of the Irbesartan/Hydrochlorothiazide Blood Pressure Reductions in Diverse Patient Populations (INCLUSIVE) clinical trial investigated whether baseline demographic and clinical variables are predictive of different degrees of blood pressure reduction following an angiotensin II receptor blocker/diuretic treatment regimen. Irbesartan/hydrochlorothiazide and other angiotensin receptor blocker combinations with a diuretic have been shown to be effective in reducing systolic blood pressure in a diverse patient population previously uncontrolled on monotherapy. Ordinary least squares regression analysis was performed on the intent-to-treat population of the INCLUSIVE study to identify variables predictive of variations in blood pressure changes in response to irbesartan/hydrochlorothiazide combination therapy. Higher baseline systolic blood pressure, female sex, type 2 diabetes, and statin therapy were found to be predictive of additional blood pressure lowering with this combination. The impact of higher baseline systolic blood pressure and diabetic state on changes in systolic blood pressure were diminished in female patients compared with male patients. In conclusion, a significant correlation may exist between certain clinical/demographic characteristics and the extent of the therapeutic response with irbesartan/hydrochlorothiazide treatment.

The Frontlines of Medicine Project progress report: standardized communication of emergency department triage data for syndromic surveillance.

This article reports progress since the original publication of the Frontlines of Medicine Project. This project is a collaborative effort of emergency medicine (including emergency medical services and clinical toxicology), public health, other government agencies involved in health care and preparedness, law enforcement, and informatics to develop nonproprietary, standardized methods for reporting emergency department patient data. These data may be used for a variety of public health or clinical care initiatives, including syndromic surveillance for chemical and biological terrorism. This article reviews the outcome of the Project meeting in April 2002. Also, the article describes a Delphi Survey process to define the data elements in a triage surveillance report and to define a set of codified values for the chief complaint data element. An initial retrospective validation of the codified chief complaint values is provided, and prospective study of the proposed Frontlines' standards is encouraged.

Income, race, and preventable hospitalizations: a small area analysis in New Jersey.

This research examines the effect of income, race, and cultural factors on preventable hospitalizations, using age and sex-adjusted preventable admissions from 53 contiguous zip codes in New Jersey from 1993 to 1995. Low income was strongly associated with high rates of preventable hospitalization in the study zip codes. Income is likely a proxy for education level, barriers to accessing primary care, and health insurance. A floor effect of income levels was present that may reflect a natural level of preventable hospitalization not affected by income, education, or health insurance status. An independent relationship found between nonwhite race and high preventable hospitalization may be in part the result of delays in seeking care affected by antecedent cultural factors. Removing financial barriers is critical but may be insufficient for reducing preventable hospitalizations if other barriers are not also addressed.

Effect of lunar cycle on temporal variation in cardiopulmonary arrest in seven emergency departments during 11 years.

OBJECTIVE: To determine the effect of the phase of the full and new moon on the variation in the number of daily cardiopulmonary resuscitations. DESIGN: A retrospective analysis of a computerized billing database of emergency department visits in a cohort of seven northern New Jersey (USA) emergency departments. PARTICIPANTS: Consecutive patients seen by emergency department physicians over an 11-year period (1 January 1988 to 31 December 1998). We determined the timing of full and new moon days from the National Oceanographic and Aeronautic Administration website. INTERVENTIONS: Time series regression estimated the independent effect of full and new moon days on the daily variation in cardiopulmonary resuscitations. Tests of statistical significance were made at alpha=0.05. RESULTS: A total of 2 370 233 emergency department visits were made during the 4018-day period of study. A total of 6827 had an emergency department diagnosis of cardiopulmonary resuscitation. We found no significant difference in the occurrence of cardiopulmonary resuscitations during the full moon (P=0.97). On average there were 6.5% fewer cardiopulmonary resuscitations during new moon days (P=0.02; 95% confidence interval 1.3-11.7%). CONCLUSION: Contrary to the traditional belief that more cardiopulmonary resuscitations occur during the full moon, we were unable to identify a significant effect during full moon days. However, there were on average 6.5% fewer cardiopulmonary resuscitations during new moon days than other days.

Current use of aspirin and antithrombotic agents in the United States among outpatients with atherothrombotic disease (from the REduction of Atherothrombosis for Continued Health [REACH] Registry).

Despite its proven efficacy, low cost, and wide availability, aspirin remains underused. We examined current aspirin use and determined factors that influence its use among outpatients in the United States (US). The REduction of Atherothrombosis for Continued Health (REACH) Registry is an international, prospective, longitudinal study of >68,000 outpatients with established atherothrombosis or >or=3 atherothrombotic risk factors. The rates of aspirin use were compared in various patient subgroups. Multivariate logistic regression models were constructed to determine the factors influencing the baseline use of aspirin and other antithrombotic agents in the US population. Approximately 70% of 25,686 US outpatients were treated with aspirin, with greater use in the Midwest and among men, whites, and those aged <65 years. Among aspirin users, 18% took other antiplatelet agents and 6% took oral anticoagulants. Low-dose aspirin (

PROCLAIM: pilot study to examine the effects of clopidogrel on inflammatory markers in patients with metabolic syndrome receiving low-dose aspirin.

Metabolic syndrome is associated with intravascular inflammation, as determined by increased levels of inflammatory biomarkers and an increased risk of ischemic atherothrombotic events. Evidence suggests that atherothrombosis and intravascular inflammation share predictive biomarkers, including high-sensitivity C-reactive protein, CD40 ligand, P-selectin, and N-terminal pro-brain natriuretic peptide. Patients who had metabolic syndrome were randomized to receive clopidogrel 75 mg/day plus aspirin 81 mg/day (n = 89) or placebo plus aspirin 81 mg/day (n = 92) for 9 weeks to assess the efficacy of each treatment in suppression of inflammatory markers. Change from baseline in the levels of high-sensitivity C-reactive protein, CD40 ligand, P-selectin, and N-terminal pro-brain natriuretic peptide at 6 weeks was assessed to evaluate each treatment. There was a significant difference at Week 6 in model-adjusted CD40-ligand levels in favor of clopidogrel plus aspirin compared with placebo plus aspirin in both the intent-to-treat population (difference between least-squares means = -186.5; 95% confidence interval, -342.3 to -30.8; P = 0.02) and the per-protocol population (P = 0.05). No significant differences were observed between the treatment arms for high-sensitivity C-reactive protein, P-selectin, and N-terminal pro-brain natriuretic peptide. There were no deaths or serious adverse events in either treatment arm. Data from this study suggest that clopidogrel can decrease the expression of the CD40-ligand biomarker.

Efficacy and safety of fixed combinations of irbesartan/hydrochlorothiazide in hypertensive women: the inclusive trial.

OBJECTIVE: This post hoc analysis of the Irbesartan/Hydrochlorothiazide Blood Pressure Reductions in Diverse Patient Populations (INCLUSIVE) trial evaluated the efficacy and safety of irbesartan/hydrochlorothiazide (HCTZ) in a diverse population of hypertensive women. METHODS: INCLUSIVE was a multicenter, prospective, open-label, single-arm trial. Adult subjects had uncontrolled systolic blood pressure (SBP 140-159 mm Hg; 130-159 mm Hg for those with type 2 diabetes mellitus [T2DM]) after > or =4 weeks of antihypertensive monotherapy. Treatment was sequential: placebo (4-5 weeks), HCTZ 12.5 mg (2 weeks), irbesartan/HCTZ 150/12.5 mg (8 weeks), and irbesartan/HCTZ 300/25 mg (8 weeks). Mean changes from baseline to treatment end in SBP and diastolic blood pressure (DBP), BP goal attainment, and safety were assessed. RESULTS: Treatment with irbesartan/HCTZ was associated with significant mean reductions in BP (intent-to-treat population, n = 370; SBP/DBP: -22.9/-10.3 +/- 14.7/8.8 mm Hg). Improvements in SBP were observed in all subgroups (p < 0.001): Caucasian (n = 207) -23.5 +/- 13.5 mm Hg; African American (n = 93) -21.0 +/- 17.2 mm Hg; Hispanics/Latino (n = 66) -23.6 +/- 14.3 mm Hg; age <65 years (n = 281) -22.5 +/- 14.7 mm Hg; age > or =65 years (n = 89) -24.3 +/- 14.5 mm Hg; T2DM (n = 97) -19.0 +/- 15.1 mm Hg; and metabolic syndrome (n = 187) -22.1 +/- 14.6 mm Hg. Overall, 82% (95% confidence interval [CI] 78%-86%) of women achieved their SBP goal, 86% (95% CI 83%-90%) achieved their DBP goal, and 76% (95% CI 71%-80%) achieved their dual SBP/DBP goal. Treatments were well tolerated in all groups. CONCLUSIONS: Irbesartan/HCTZ treatment was effective and well tolerated in a diverse population of women whose BP was previously uncontrolled on monotherapy.

Meta-analysis of prostate-specific antigen and digital rectal examination as screening tests for prostate carcinoma.

BACKGROUND: Physicians commonly screen for prostate cancer by using prostate-specific antigen (PSA) and digital rectal examination (DRE). The usefulness of these screening mechanisms is not well established, however. A meta-analysis of PSA and DRE to detect prostate carcinoma was conducted with a focus on sensitivity, specificity, and positive predictive value. METHODS: A literature search of OVID database (1966 to November 1999) using the medical subject headings "prostate-specific antigen" and "mass screening," as well as "prostate carcinoma," was performed. Thirteen articles were selected for the meta-analysis in this study. Most studies included asymptomatic men older than 50 years from various countries. Pooled results were calculated from the individual reports for sensitivity, specificity, and positive predictive value for PSA and DRE based on biopsy result as the reference standard. RESULTS: The overall detection rate of prostate carcinoma was 1.8% based on a positive biopsy. Of the prostate carcinoma detected, 83.4% was localized. The pooled sensitivity, specificity, and positive predictive value for PSA were 72.1%, 93.2% and 25.1%, respectively; and for DRE were 53.2%, 83.6% and 17.8%, respectively. CONCLUSIONS: There were two major outcomes of this meta-analysis. One was the potential for detecting early-stage prostate cancer with these screening tests, because 83.4% of total cancers detected were localized. The second important outcome was that the overall sensitivity, specificity, and positive predictive value for PSA were higher than those for DRE when used as a screening tool to detect prostate cancer. When a patient has abnormal findings using PSA and DRE, the chance of cancer is 1 in 4 or 5. Conversely, when findings from PSA and DRE are normal, the chance of missing a cancer is about 10%.

Agreement between pediatric medication orders and medication cardex.

This study examined the individual and temporal factors that explain whether the medication cardex agrees with a physician's pediatric medication order. After controlling for several potentially confounding factors, it was found that, among other things, pediatric intensive care unit cardexes were 62% less likely to agree with the physician order than cardexes from other units. Cardexes for "stat" orders were twice as likely to agree with the physician order. These data support the possibility that "low-tech" changes in the process of providing care can improve the likelihood that the medication order will agree with the cardex and, as a result, reduce the likelihood of medication errors.

Calendar patterns in the occurrence of cardiac arrest.

Our objective was to determine whether monthly and daily patterns existed in cardiac arrests in a 7 emergency department (ED) cohorts in New Jersey. We conducted a retrospective analysis of a computerized database over an 11-year period containing 2,370,233 patient visits and 6,827 nontraumatic cardiac arrests. Time-series regression revealed colder months having more cardiac arrests, especially for patients >or=65 years of age. In that age group, December and March were highest with 17% more (P =.002) than the average of nonsignificant months. August was the lowest with 19% (P =.001) fewer cardiac arrests. Day-of-week variation was found only for patients <65 years of age. Saturdays and Mondays had more arrests than nonsignificant days, 15% (P <.0001) and 9% (P =.01), respectively. We conclude that there are monthly and daily patterns in cardiac arrests, with more arrests in the colder months for those >or=65 years of age, and more for ages <65 on Saturdays and Mondays.