Adverse outcome analyses of observational data: assessing cardiovascular risk in HIV disease.

Clinical decisions are ideally based on randomized trials but must often rely on observational data analyses, which are less straightforward and more influenced by methodology. The authors, from a series of expert roundtables convened by the Forum for Collaborative HIV Research on the use of observational studies to assess cardiovascular disease risk in human immunodeficiency virus infection, recommend that clinicians who review or interpret epidemiological publications consider 7 key statistical issues: (1) clear explanation of confounding and adjustment; (2) handling and impact of missing data; (3) consistency and clinical relevance of outcome measurements and covariate risk factors; (4) multivariate modeling techniques including time-dependent variables; (5) how multiple testing is addressed; (6) distinction between statistical and clinical significance; and (7) need for confirmation from independent databases. Recommendations to permit better understanding of potential methodological limitations include both responsible public access to de-identified source data, where permitted, and exploration of novel statistical methods.

Determinants of antidepressant treatment compliance: implications for policy.

Depression is among the most prevalent, devastating, and undertreated disorders in our society. Treatment with antidepressant medications is effective in controlling symptoms, but treatment beyond the point of symptom resolution is necessary to restore functional status and prevent recurrent episodes. An important step in improving compliance is to identify the determinants of antidepressant treatment compliance. A broader motivation for our study is to examine compliance by patients with a chronic but treatable disease. With claims data between 1990 and 1993, this study uses logistic regression analysis to examine the determinants of compliance among 2,012 antidepressant recipients. The results show that initiating treatment with a tricyclic antidepressant reduces the probability of antidepressant treatment compliance. Initiating treatment with a selective serotonin reuptake inhibitor and undergoing family, group, or individual psychotherapy treatments increase the probability of compliance. Case management does not meaningfully affect compliance. Implications for policy and clinical practice are discussed.

Abciximab provides cost-effective survival advantage in high-volume interventional practice.

BACKGROUND: Placebo-controlled randomized trials of platelet glycoprotein (GP) IIb/IIIa blockade during percutaneous coronary intervention have demonstrated efficacy of these agents for reducing the risk of periprocedural ischemic events. However, cost-effectiveness of this adjunctive pharmacotherapy has been scrutinized. Extrapolation of cost-efficacy observations from clinical trials to "real world" interventional practice is problematic. METHODS: Consecutive percutaneous coronary interventions (n = 1472) performed by Ohio Heart Health Center operators at The Christ Hospital, Cincinnati, Ohio, in 1997 were analyzed for procedural and long-term (6-month) outcomes and charges. Observations on cost and efficacy (survival) were adjusted for nonrandomized abciximab allocation by means of "propensity scoring" methods. RESULTS: Abciximab therapy was associated with a survival advantage to 6 months after percutaneous coronary intervention. The average reduction in mortality rate at 6 months was 3.4% (unadjusted) and 4.9% when adjusted for nonrandomization. The average charge increment to 6 months was $1512 (unadjusted) and $950 when adjusted for nonrandomization. Patients deriving the greatest reduction in mortality rates also had a reduction in total cardiovascular charges to 6 months. Distinguishing demographics of this population included multivessel coronary intervention, coronary stent deployment, intervention within 1 week of myocardial infarction, and lower left ventricular ejection fraction. The average cost per life-year gained in this study was $2875 for all patients (unadjusted) and $1243 when adjusted for nonrandomization. CONCLUSIONS: Abciximab provides a cost-effective survival advantage in high-volume interventional practice that compares favorably with currently accepted standards. Clinical and procedural demographics associated with increased cost-effectiveness included multivessel coronary intervention, stent deployment, recent (<1 week) myocardial infarction, and impaired left ventricular function.

Cost-effectiveness of fluoxetine plus pindolol in patients with major depressive disorder: results from a randomized, double-blind clinical trial.

Some preliminary studies have suggested that the beta-adrenoceptor 5-HT1A antagonist pindolol (PIN) could increase the effect of selective serotonin reuptake inhibitors (SSRIs). We prospectively estimated the cost-effectiveness of fluoxetine and pindolol versus fluoxetine plus placebo, using results from the first double-blind randomized clinical trial comparing both treatments. Efficacy and medical care resource utilization were collected prospectively in a parallel, randomized, double-blind clinical trial conducted in a single centre in Spain. Average cost-effectiveness (cost/% response and cost/% remission) as well as the incremental cost-effectiveness were calculated for both treatments. A 'bootstrap' method was used to calculate confidence limits around the incremental cost-effectiveness ratio. A significantly greater percentage of patients (one-tailed P < 0.05) in the fluoxetine FLX + PIN group than in the FLX + PLA group had experienced a therapeutic response (74.5% versus 58.97%) at 6 weeks. Direct medical costs were lower in the FLX + PIN group (mean 2508 pesetas per patient) than in the FLX + PLA group (mean 31870 pesetas per patient). Hospital admissions due to worsening of depressive symptoms were significantly lower (P < 0.05) in the FLX + PIN group (0/55) than in the FLX + PLA group (4/56). The observed differences in average costs and percentage response in the study were -29362 pesetas (< 0) and 15.6% (> 0), respectively, and the resulting cost-effectiveness ratio was negative. These outcomes indicate that the FLX + PIN option completely dominates FLX + PLA. These results suggest that, over a course of 6 weeks of treatment, the combination of fluoxetine and pindolol incurs lower direct medical costs than treatment with fluoxetine placebo. Despite their limitations, economic assessments in addition to clinical trials allow a 'dynamic assessment' on the potential success of the drug, both from a clinical and an economic point of view, allowing decisions on priorities to be made earlier.

Determinants of antidepressant treatment outcome.

OBJECTIVE: To understand the determinants of the outcome of an episode of major depression, including factors that affect receipt of guideline-consistent care and their subsequent effect on treatment outcomes, particularly relapse or recurrence. Results of previous studies are generalized to a population typical of depressed individuals in the United States, i.e., a cohort of antidepressant users with employer-provided health benefits. STUDY DESIGN: A quasi-experimental design was used to assess the determinants of the outcome of an episode of major depression. Healthcare utilization-based measures of treatment characteristics and outcomes were used. PATIENTS AND METHODS: The final analytical file for this study contained data on 2917 patients who had an antidepressant prescription associated with an indicator of a depressive disorder. We identified relapse or recurrence of depression by (1) a new episode of antidepressant therapy, (2) suicide attempt, (3) psychiatric hospitalization, (4) mental health-related emergency department visits, or (5) electroconvulsive therapy. Antidepressant use patterns were used to construct a measure for adherence to treatment guidelines. Multivariate Cox proportional hazard and logit regression models were used to predict relapse/recurrence and adherence with treatment guidelines, respectively, for each patient. RESULTS: Factors that affect relapse/recurrence include comorbidities, demographics, and adherence to treatment guidelines. Factors that affect adherence to treatment guidelines include choice of initial antidepressant drug, comorbidities, psychotherapy, and frequency of physician visits. CONCLUSIONS: Adherence to treatment guidelines was associated with a significant reduction in the likelihood of relapse or recurrence of depression. Choice of initial antidepressant drug affects adherence to treatment guidelines.

Resampling and multiplicity in cost-effectiveness inference.

We compare published methods for placing statistical confidence limits around incremental cost-effectiveness ratio statistics and show that only a nonparametric, bootstrap approach using polar angles gives completely reasonable results when neither treatment has significant advantages in cost or effectiveness. We also discuss alternative ways to report analytical results using plots, confidence or tolerance limits, and quadrant acceptability fractions. Finally, we use simulation to study the multiplicity bias that can be introduced into ICER confidence limits when only the most favorable results are reported over several possible choices of numerator cost measure and denominator effectiveness measure.

Reliability, validity, and application of the medical outcomes study 36-item short-form health survey (SF-36) in schizophrenic patients treated with olanzapine versus haloperidol.

UNLABELLED: Schizophrenia leads to impairments in mental, social, and physical functioning, which should be included in evaluations of treatment. OBJECTIVES: This study was designed to determine the reliability and validity of the Medical Outcomes Study Short Form Health Survey (SF-36) for schizophrenic patients, to characterize perceived functioning and well being and to compare short-term change in SF-36 scores for patients treated with olanzapine or haloperidol. RESEARCH DESIGN: Data were obtained from a randomized, double-blind trial comparing these agents for safety, efficacy, and cost effectiveness. A 6-week acute treatment portion preceded a 46-week "responder extension" phase. SUBJECTS: A subsample (n = 1,155) completing a pre-treatment SF-36 provided data for this study. MEASURES: Psychometric analyses were conducted, and perceived level of functioning was compared with that for the US adult population. Change from baseline to 6 weeks was examined by treatment group. RESULTS: Clear evidence was obtained for the instrument's reliability and validity for these patients. There were marked deficits in General health, Vitality, Mental health, Social functioning, and in Role limitations resulting from both physical and emotional problems. Olanzapine-treated patients improved in 5 of 8 domains to a significantly greater degree than did haloperidol patients. CONCLUSIONS: The SF-36 can be a reliable and valid measure of perceived functioning and well being for schizophrenic patients. The perceptions of functioning can be valuable indices of disease burden and can help to demonstrate the effectiveness of newer antipsychotic medications such as olanzapine.

The application of sample selection models to outcomes research: the case of evaluating the effects of antidepressant therapy on resource utilization.

Non-randomized studies of treatment effects have come under criticism because of their failure to control for potential biases introduced by unobserved variables correlated with treatment selection and outcomes. This paper describes the basic concepts of sample selection models--a technique used widely in the economics evaluation literature for nearly two decades--and discusses the potential role of these models in outcomes research. In addition, it presents a case study of the application of the sample selection modelling approach to evaluation of the effects of antidepressant therapies on medical expenditures for physician services. This case study presents empirical comparisons of alternative model specifications and discusses practical issues in evaluation of sample selection models. We demonstrate that, in this particular case, sample selection models yield very different conclusions regarding treatment effects than traditional ordinary least squares regression.

What does treatment of depression really cost?

This DataWatch presents estimates of the health care charges for adults who are diagnosed and treated for depression in primary care. More than nine out of ten of these adults sought care for at least one nondepressive illness during the year following treatment initiation. One average, these conditions accounted for more than 70 percent of the total charges. Attempts to manage the costs of caring for depressed persons must consider the impact of nondepressive illness.

Effect of antidepressant therapy on health care utilization and costs in primary care.

OBJECTIVE: Four groups of patients receiving different antidepressant drugs in a primary care setting were compared in terms of duration of antidepressant therapy and health and mental health care utilization and costs. METHODS: A retrospective analysis of the medical and pharmacy claims of an employed population and their families was conducted. A total of 1,242 patients with a diagnosis of depression were included in the analyses. The four antidepressant cohorts were fluoxetine (N = 799), trazodone (N = 89), the tricyclics amitriptyline and imipramine (N = 104), and the secondary amine tricyclics desipramine and nortriptyline (N = 250). The primary outcome measures were total health care charges, total charges for mental health services, and the pattern of antidepressant use. Secondary measures included charges for outpatient care and pharmacy and the number of outpatient visits. Data analysis involved use of two-stage multivariate regression modeling known as sample selection models. RESULTS: Patients taking fluoxetine achieved higher rates of continuous use for at least six months compared with those taking the other drugs. After selection bias due to observed and unobserved characteristics and other confounding variables was adjusted for, no significant differences were found between drug cohorts in total medical charges. CONCLUSIONS: Improvements in the process of care at no apparent increase in total charges appear possible through appropriate medication therapy.

Bootstrap analyses of cost effectiveness in antidepressant pharmacotherapy.

In this study, we describe 'bootstrap' methodology for placing statistical confidence limits around an incremental cost effectiveness ratio (ICER). This approach was applied to a retrospective study of annual charges for patients undergoing pharmacotherapy for depression. We used MarketScanSM (service mark) data from 1990 to 1992, which includes medical and pharmacy claims for a privately insured group of employed individuals and their families in the US. Our primary effectiveness measure was the proportion of patients who remained stable on their initial antidepressant medication for at least 6 consecutive months. Our primary cost measure was the total annual charge incurred by patients taking the selective serotonin reuptake inhibitor fluoxetine, a tricyclic antidepressant or a heterocyclic antidepressant. On average, fluoxetine pharmacotherapy tended to decrease annual charges by $US16.48 per patient for each percentage increase in depressed patients remaining stable on initial pharmacotherapy for 6 months, resulting in a negative ICER point-estimate. However, the upper ICER confidence limit is positive, which means that fluoxetine treatment may possibly increase annual per patient charges. With 95% confidence, any such increase was no more than $US130 per patient for each percentage increase in patients remaining stable on initial pharmacotherapy for at least 6 months. One advantage of using a bootstrap approach to ICER analysis is that it does not require restrictive distributional assumptions about cost and outcome measures. Bootstrapping also yields a dramatic graphical display of the variability in cost and effectiveness outcomes that result when a study is literally 'redone' hundreds of times. This graphic also displays the ICER confidence interval as a 'wedge-shaped' region on the cost-effectiveness plane. In fact, bootstrapping is easier to explain and appreciate than the elaborate calculations and approximations otherwise involved in ICER estimation. Our discussion addresses key technical questions, such as the role of logarithmic transformation in symmetrising highly skewed cost distributions. We hope that our discussion contributes to a dialogue, leading ultimately to a consensus on analysis of ICERs.

Methacholine-induced pulmonary gas trapping in guinea pigs, hamsters, mice, and rats.

Postmortem pulmonary gas trapping was investigated as an index of in vivo airway obstruction following methacholine inhalation in four different rodent species. Male guinea pigs (Hartley), hamsters (golden Syrian), mice (A/J, BALB/c, and ICR), and rats (Brown-Norway, Fischer 344, Lewis, and Sprague-Dawley) were exposed to aerosols of methacholine or sodium chloride. Maximum excised lung gas volumes (ELGV) of methacholine-exposed guinea pigs, hamsters, mice, and rats were 2.3-8.7 times those of sodium chloride-treated animals. Mean ELGV values of sodium chloride-exposed animals ranged from 1.50 +/- 0.20 ml/kg for guinea pigs to 2.75 +/- 0.20 ml/kg for Brown-Norway rats. Although all species responded to methacholine, guinea pigs were the most responsive, with approximately 1.6 microgram/kg of inhaled methacholine needed to increase ELGV to 200% of control. Compared with guinea pigs, hamsters, mice, and rats were 11- to 1,395-fold less responsive. Although hamsters, mice, and rats are less sensitive than guinea pigs to the airway-obstructive effects of methacholine, pulmonary gas trapping appears useful as a measure of airway responses in these species.

Statistical auditing of toxicology reports.

Statistical auditing is a new report review process used by the quality assurance unit at Eli Lilly and Co. Statistical auditing allows the auditor to review the process by which the report was generated, as opposed to the process by which the data was generated. We have the flexibility to use different sampling techniques and still obtain thorough coverage of the report data. By properly implementing our auditing process, we can work smarter rather than harder and continue to help our customers increase the quality of their products (reports). Statistical auditing is helping our quality assurance unit meet our customers' need, while maintaining or increasing the quality of our regulatory obligations.

Bronchopulmonary actions of 1-(1,2,3,4-tetrahydro-1-naphthylenyl)-1H-imidazole, nitric acid salt (LY150310), a substituted imidazole, in the guinea pig.

1-(1,2,3,4-tetrahydro-1-naphthylenyl)-1H-imidazole, nitric acid salt (LY150310), was examined for bronchodilator activity in the guinea pig. In guinea pig tracheal preparations, LY150310 competitively antagonized the contractile effects of exogenous histamine and blocked the histamine-mediated component of contractions produced by ovalbumin challenge. LY150310 had little effect on the nonhistamine component of ovalbumin-induced contractions of lung parenchymal strips, but it enhanced the production of prostaglandin (PG) E2 and PGF2 alpha although it partially inhibited thromboxane B2 formation. In other studies, in which postmortem pulmonary gas trapping was used as an index of in vivo airway obstruction, i.v. LY150310 dose-dependently inhibited the bronchospasm produced by aerosols of the divalent cationic ionophore A23187, histamine, 5-hydroxytryptamine, leukotriene D4, methacholine, ovalbumin or platelet activating factor. LY150310 was equal to or more potent than aminophylline in all test systems. Also, orally administered LY150310 inhibited the airway obstruction produced by selected challenge aerosols. In ex vivo studies, LY150310 elevated PGE2 and tended to decrease thromboxane B2 in sodium arachidonate-stimulated whole blood. However, PGE2 and other cyclooxygenase products did not appear to account for in vivo bronchodilation, because combining LY150310 and piroxicam did not alter inhibition of A23187-induced airway obstruction. Our results demonstrate that LY150310 reduces airway obstruction caused by a variety of bronchoconstrictive agents, including A23187 and ovalbumin. Although this substituted imidazole appears to have activity as a histamine H1-receptor antagonist and can alter prostanoid concentrations in vitro and in vivo, its mode of bronchodilation is unclear.