Comparing SSRI Treatment Costs for Depression Using Retrospective Claims Data: The Role of Nonrandom Selection and Skewed Data

Background and Objectives: Since conventional randomized clinical trials often do not reflect the real world circumstances of prescribing behavior and patient outcomes, the use of retrospective administrative claims databases (RACD) has become more common in treatment cost comparisons among alternative pharmaceutical compounds. Several recent RACD studies have compared treatment costs for depressed patients prescribed SSRIs such as fluoxetine, sertraline and paroxetine. These cost comparisons have reached mixed conclusions. To begin to explain and reconcile the mixed SSRI cost comparison evidence, we undertake a variety of alternative multivariate analyses using a publicly available RACD.
Methods and Data: The 1995 to 1996 data encompasses a time period when all three SSRIs had become well-established agents. We report and compare results from multivariate linear regressions, logistic regressions, ordered probits and sample selectivity models, and examine robustness when adjustments are made for outlier observations and skewed distributions.
Results and Conclusions: While choice of initial SSRI is nonrandom, the effect of sample selectivity on total depression-related and total health care expenditure is neutral across SSRIs. Although most cost measures are numerically greatest for fluoxetine, depression-related outpatient and hospitalization costs do not significantly differ by choice of initial SSRI. These findings are robust to alternative assumptions, specifications, and procedures. Antidepressant medication costs, however, are significantly higher when fluoxetine is the initial SSRI rather than sertraline or paroxetine, reflecting the larger proportion of fluoxetine patients prescribed a daily dosage of two or more capsules. Both total depression-related and total health care log–transformed costs are significantly lower for sertraline than fluoxetine.     

Fluoxetine: a randomized clinical trial in the maintenance of weight loss

Because current weight-reduction treatments have considerable recidivism, a therapy that could help patients maintain weight loss would be of benefit. A six-center, randomized, double-blind trial compared the effects of the specific serotonin uptake inhibitor, fluoxetine hydrochloride, and placebo on maintenance of weight loss. Obese outpatients who had lost > or = 3.6 kg after 8 weeks of single-blind fluoxetine 60 mg/day in the qualification phase (N=317 [70.4% of patients entered]; mean +/- standard deviation [SD] weight loss, 6.8 +/- 2.8 kg) were randomly assigned to fluoxetine 20 mg/day (N=104), fluoxetine 60 mg/day (N=106), or placebo (N=107) for 40 weeks (maintenance phase). Patients received minimal nutrition/dietary counseling. Qualification phase clinic visits were biweekly; maintenance phase visits were monthly for 4 months, then bimonthly for 6 months. Patients treated with fluoxetine 60 mg/day continued to lose weight for 8 additional weeks (16 weeks total; maximum mean +/- SD weight loss, 7.2 +/- 4.6 kg); those treated with fluoxetine 20 mg/day or placebo began to regain weight. Mean weights remained below baseline values at week 48 (all groups); treatment differences were not statistically significant. Study completion rates were comparable (fluoxetine 20 mg/day, 67.3%; fluoxetine 60 mg/day, 56.6%; placebo, 67.3%; p = 0.175). Among commonly reported adverse events (> 10% incidence), only asthenia was reported statistically significantly (p < 0.050) more frequently with fluoxetine than with placebo. Few patients discontinued for any single adverse event. Fluoxetine 60 mg/day was effective for a longer period than fluoxetine 20 mg/day or placebo in maintaining weight loss. Overall, fluoxetine was safe and well tolerated.     

Cost-effectiveness analysis of treatments for premenstrual dysphoric disorder

Background

Premenstrual syndrome (PMS) is reported to affect between 13% and 31% of women. Between 3% and 8% of women are reported to meet criteria for the more severe form of PMS, premenstrual dysphoric disorder (PMDD). Although PMDD has received increased attention in recent years, the cost effectiveness of treatments for PMDD remains unknown.

Objective

To evaluate the cost effectiveness of the four medications with a US FDA-approved indication for PMDD: fluoxetine, sertraline, paroxetine and drospirenone plus ethinyl estradiol (DRSP/EE).

Methods

A decision-analytic model was used to evaluate both direct costs (medication and physician visits) and clinical outcomes (treatment success, failure and discontinuation). Medication costs were based on average wholesale prices of branded products; physician visit costs were obtained from a claims database study of PMDD patients and the Agency for Healthcare Research and Quality. Clinical outcome probabilities were derived from published clinical trials in PMDD. The incremental cost-effectiveness ratio (ICER) was calculated using the difference in costs and percentage of successfully treated patients at 6 months. Deterministic and probabilistic sensitivity analyses were used to assess the impact of uncertainty in parameter estimates. Threshold values where a change in the cost-effective strategy occurred were identified using a net benefit framework.

Results

Starting therapy with DRSP/EE dominated both sertraline and paroxetine, but not fluoxetine. The estimated ICER of initiating treatment with fluoxetine relative to DRSP/EE was $US4385 per treatment success (year 2007 values). Cost-effectiveness acceptability curves revealed that for ceiling ratios ≥$US3450 per treatment success, fluoxetine had the highest probability (≥0.37) of being the most cost-effective treatment, relative to the other options. The cost-effectiveness acceptability frontier further indicated that DRSP/EE remained the option with the highest expected net monetary benefit for ceiling values ≤$US3900 per treatment success.

Conclusion

These analyses suggest that initiating therapy with DRSP/EE may be a cost-effective option in the treatment of PMDD.     

All-Terrain Vehicle Safety and Use Patterns in Central Illinois Youth

Context: All-terrain vehicles' (ATVs) popularity and associated injuries among children are increasing in the United States. Currently, most known ATV use pattern data are obtained from injured youth and little documented data exist characterizing the typical ATV use patterns and safety practices among American children in general.
Purpose: To describe the typical ATV safety and use patterns of rural youth.
Methods: A cross-sectional anonymous mail survey was conducted of youth participants (ages 8-18) in the 4-H Club of America in four Central Illinois counties. Questions examined ATV use patterns, safety knowledge, safety equipment usage, crashes, and injuries.
Findings: Of 1,850 mailed surveys, 634 were returned (34% response rate) with 280 surveys (44% of respondents) eligible for analysis. Respondents were principally adolescent males from farms or rural locations. Most drove ≤1 day per week (60.2%) and used ATVs for recreation (36%) or work (22.6%) on farms and/or private property (53.4%). Most never used safety gear, including helmets (61.4%), and few (14.6%) had received safety education. Of the 67% who experienced an ATV crash, almost half (44%) were injured. Children with safety training had fewer crashes ( P = .01), and those riding after dark ( P = .13) or without adult supervision ( P = .042) were more likely injured.
Conclusions: ATV use is common in a rural 4-H population. Most child ATV users were adolescent boys, had little safety training and did not use safety equipment or helmets. ATV injury prevention efforts should focus on these areas.     

Mixture and single-substance acute toxicity of selective serotonin reuptake inhibitors in Ceriodaphnia dubia

Selective serotonin reuptake inhibitors (SSRIs) are neurologically active drugs that can contaminate surface waters and have the potential to negatively affect aquatic organisms. In this investigation, the 48-h acute toxicity of mixtures (binary and quaternary) of four common SSRIs (fluoxetine [Prozac], sertraline [Zoloft], paroxetine [Paxil], and citalopram [Celexa]) were determined in the daphnid Ceriodaphnia dubia. Logistic regression was used to model mortality data and to investigate the applicability of concentration addition and independent action models to explain observed mortality. The concentrations estimated to induce 50% mortality in 48 h for the individual SSRIs sertraline, fluoxetine, paroxetine, and citalopram were 0.48 to 0.66, 1.23 to 1.84, 2.23 to 3.57, and 10.47 to 14.53 microM, respectively. Concentration addition was a better predictor of mixture effects than independent action and suggested that the tested SSRIs have a similar mechanism of action. Results indicate that environmental hazard assessments should be conservative and consider that acutely toxic effects in aquatic organisms can be additive for each SSRI in a mixture.     

Depression Treatment in Rural California:Preliminary Survey of Nonpsychiatric Physicians

Depressive disorders have been recognized as disabling conditions of public health proportions. However, in areas underserved by mental health professionals, the treatment of depressed patients becomes challenging. Furthermore, patients living in rural areas and communities underserved by health professionals are at risk for high levels of depressive symptoms and low access to care. Physicians (N = 58) of multiple nonpsychiatric specialties in Imperial County, a rural underserved area in California, were surveyed to ascertain their preferred strategies in the management of depressed patients. More than half (57%) of the respondents preferred to either refer patients to a mental health specialist (p < .01) as the only strategy, or in combination with counseling, prescribing medication, or both. The most commonly reported form of counseling was of a supportive nature. The most commonly prescribed drugs were selective serotonin reuptake inhibitors (in order of frequency:fluoxetine, sertraline, and paroxetine). Tricyclic antidepressants and benzodiazepines were identified as first-line drugs by some pediatricians and surgeons. The results of this study support the need for enhanced postgraduate training in the treatment of depression for nonpsychiatric physicians, and greater exposure of psychiatric residents to rural psychiatry.     

Evaluating changes in sexual functioning in depressed patients: sensitivity to change of the CSFQ

Accurately evaluating alterations in sexual functioning requires a validated instrument that measures clinically relevant change over time. One-hundred one depressed patients from 15 Spanish out-patient clinics completed the Changes in Sexual Functioning Questionnaire (CFSQ; Clayton, McGarvey, & Clavet, 1997) at baseline and after 6 months of treatment with fluoxetine, nefazodone, paroxetine, or venlafaxine. Sexual desire/interest showed a nearly substantial floor effect (30% of patients indicated the maximum score) for women in the nefazodone group at baseline and in the paroxetine group at final visit. The percentage of dimensions recording change was greater for women (80%) than for men (20%) in the nefazodone group (improving changes) and greater for men (40%) than for women (20%) in the paroxetine group (worsening changes). Highest effect sizes were found on sexual desire/frequency with improvement in women in the nefazodone group (SES = 0.49), and on orgasm/ejaculation with worsening in men in the paroxetine group (SES = -1.45). In conclusion, the CSFQ is sensitive to bidirectional changes and is appropriate for measuring sexual dysfunction.     

PIC10 The Influence of Case Mix Bias on Costs of Hospitalisation for Lower Respiratory Tract Infection

OBJECTIVE: To compare costs of hospitalisation for lower respiratory tract infection (LRTI) in patients who received antibiotics before admission versus those who did not and in patients with and without underlying chronic obstructive airways disease (COAD).
METHODS: All hospitalisations were analysed in a population of 350,000 resident in Tayside during 1993–94. Three groups of patients were identified by primary discharge diagnosis in 1993–94 and previous admissions from 1980 to 1992: (1) acute exacerbation of COAD, (2) LRTI plus a secondary diagnosis of COAD or previous admission with COAD, and (3) LRTI but no secondary COAD or previous admission with COAD. Setting-specific costs were applied (e.g., general medicine, intensive care, geriatrics). Dispensed antibiotic prescribing in the 28 days before admission was identified from all community pharmacies. Non-parametric statistical tests were used.
RESULTS: Patients with COAD were more likely to have received antibiotics before admission: COAD (n = 893) 49%; COAD+LRTI (n = 316) 43%; LRTI only (n = 822) 33%. Odds ratio for COAD vs LRTI only 1.90 (95% CI 1.56 to 2.31); COAD+LRTI vs LRTI only 1.50 (95% CI 1.15 to 1.96). Patients who received antibiotics before admission had lower hospital costs than patients who did not. Median total costs per admission: COAD £1050 vs £1164 (p = 0.5); COAD+LRTI £1067 vs £1354 (p = 0.5); LRTI only £1220 vs £1500; (p = 0.009). Adjusted for community antibiotic prescribing, the hospital costs of patients with LRTI were significantly higher than those of patients with COAD (p = 0.001) but not those of patients with COAD+LRTI (p = 0.096).
CONCLUSION: Economic models of the potential impact of different community antibiotics on hospital LRTI costs will be subject to case mix bias unless they adjust for community antibiotic use and co-morbidity with COAD.     

A randomized controlled trial of information-giving to patients referred for coronary angiography: effects on outcomes of care

Objective To assess the impact of providing an educational videotape, ` Treatment Choices for Ischaemic Heart Disease: a Shared Decision-Making Program Videotape ,' to patients referred for coronary angiography compared with standard patient-physician decision making (usual care).
Study design Randomized controlled clinical trial.
Setting University Hospital and Veterans Affairs Hospital.
Patients A consecutive sample of 217 patients referred for coronary angiography were randomized to receive `usual care' or to receive the videotape in addition to standard patient physician decision making (videotape): 109 completed the study (50% completion rate).
Main outcome measures Knowledge of coronary artery disease, satisfaction, self-reported physical and mental health functioning, and the proportion of patients who were referred for coronary revascularization.
Results Compared with patients who received `usual care,' those who received the videotape were more knowledgeable (mean score 83 vs. 58%; P  < 0.0001) but less satisfied with their treatment (79 vs. 88%; P  = 0.038). There were no significant differences between the videotape and `usual care' groups with respect to satisfaction with the decision making process (mean score 73 vs. 77%; P  = 0.37), satisfaction with the decision made (mean score 73 vs. 78%; P  = 0.28), physical functioning (38 vs. 38%; P  = 0.76), mental health functioning (49 vs. 49%; P  = 0.94), or in referral for coronary revascularization (OR 0.60; 95% CI 0.22–1.65; P  = 0.33).
Conclusion Although the educational videotape increased patients' knowledge level, it was associated with a decrease in their level of satisfaction with treatment. Before there is wide-spread dissemination of this technology, advocates should demonstrate its effectiveness in everyday practice.     

PCV13 Prevalence and Cost of Hospitalizations Due to Angina in the United States using a National Database

OBJECTIVE: To determine clinical and economic outcomes of primary PTCA and assess the effects of comorbidities (congestive heart failure, diabetes, hypercholesterolemia, hypertension) on outcomes using a national database.
METHODS: Data were obtained from MarketScanâ Medstatâ, which contains claims data for 4–5 million people. Patients with a hospital admission for PTCA (CPT codes 92982, 92984) in 1992 and one year of follow-up charge data were identified. A total of 2,663 patients with a single vessel and 331 with a multiple vessel procedure were included. Patients who died or were missing prior to one-year follow-up were excluded (n = 653). Total charges were calculated by summing charges for the index procedure plus all inpatient and outpatient charges incurred during the year following PTCA. Multivariate regression was used to assess effects of comorbidities on outcomes (adjusted for age and other factors).
RESULTS: Patients with single vessel versus multiple vessel PTCA did not differ according to comorbid and other characteristics, or most clinical or economic outcomes; therefore combined data are presented. The frequency of clinical outcomes in the year following PTCA was: rehospitalization for angina-41%, myocardial infarction (MI)-33%, repeat PTCA-15%, coronary artery bypass grafting-7%, and stroke-3%. Comorbidities related to clinical outcomes were: congestive heart failure (CHF) for MI (p = 0.0001), diabetes for angina rehospitalization (p = 0.06), and diabetes (p = 0.05) and hypertension (p = 0.004) for repeat PTCA. Total charges per patient at one year after primary PTCA were $35,257, and total charges were strongly related to the presence of diabetes (p = 0.001) and CHF (p = 0.001). There was no association with hypertension or hypercholesterolemia.
CONCLUSION: These findings show that the presence of certain comorbid factors does influence clinical and economic outcomes following PTCA.     

Sertraline for the prevention of relapse in detoxicated alcohol dependent patients with a comorbid depressive disorder: a randomized controlled trial

AIMS: We performed a double-blind, placebo-controlled randomized trial of sertraline in recently detoxified alcohol-dependent patients with current depressive symptoms. The objectives of the study were to evaluate the efficacy of sertraline at achieving stable abstinence, at ameliorating depressive symptoms and at improving quality of life in these patients. METHODS: The study included 83 patients, who received either sertraline (50-150 mg/day) or placebo for 24 weeks. The primary outcome criteria were the rate of relapse into alcohol consumption and the rate of response on the Montgomery and Asberg Depression Rating Scale (MADRS). RESULTS: At the end of the treatment period, relapse rates were 23.1% in the placebo group and 31.8% in the sertraline group. Responder rates for depression were 38.5% for the placebo group and 44.2% for the sertraline group. There was no significant difference between treatment groups with either variable. However, when patients were stratified into severe (MADRS score >or=26) and moderate (MADRS score <26) depression at inclusion, a significant treatment benefit with sertraline was observed in the former group. Quality of life, determined by the SF-36, improved in both groups, with more benefit observed for the sertraline group on mental health items. Sertraline was well tolerated, and the incidence of adverse events was similar in the two treatment groups. CONCLUSIONS: The explanation for the overall good outcome in both treatment groups and for the inability to demonstrate a clear treatment effect may reside in the clinical features of the patients included.     

Mixture and single-substance toxicity of selective serotonin reuptake inhibitors toward algae and crustaceans

Selective serotonin reuptake inhibitors (SSRIs) are used as antidepressant medications, primarily in the treatment of clinical depression. They are among the pharmaceuticals most often prescribed in the industrialized countries. Selective serotonin reuptake inhibitors are compounds with an identical mechanism of action in mammals (inhibit reuptake of serotonin), and they have been found in different aqueous as well as biological samples collected in the environment. In the present study, we tested the toxicities of five SSRIs (citalopram, fluoxetine, fluvoxamine, paroxetine, and sertraline) as single substances and of citalopram, fluoxetine, and sertraline in binary mixtures in two standardized bioassays. Test organisms were the freshwater algae Pseudokirchneriella subcapitata and the freshwater crustacean Daphnia magna. In algae, test median effect concentrations (EC50s) ranged from 0.027 to 1.6 mg/L, and in daphnids, test EC50s ranged from 0.92 to 20 mg/L, with sertraline being one of the most toxic compounds. The test design and statistical analysis of results from mixture tests were based on isobole analysis. It was demonstrated that the mixture toxicity of the SSRIs in the two bioassays is predictable by the model of concentration addition. Therefore, in risk assessment based on chemical analysis of environmental samples, it is important to include the effect of all SSRIs that are present at low concentrations, and the model of concentration addition may be used to predict the combined effect of the mixture of SSRIs.     

Acute and chronic toxicity of five selective serotonin reuptake inhibitors in Ceriodaphnia dubia

Contamination of surface waters by pharmaceutical chemicals has raised concern among environmental scientists because of the potential for negative effects on aquatic organisms. Of particular importance are pharmaceutical compounds that affect the nervous or endocrine systems because effects on aquatic organisms are possible at low environmental concentrations. Selective serotonin reuptake inhibitors (SSRIs) are drugs used to treat clinical depression in humans, and have been detected in low concentrations in surface waters. In this investigation, the acute and chronic toxicity of five SSRIs (fluoxetine, Prozac; fluvoxamine, Luvox; paroxetine, Paxil; citalopram, Celexa; and sertraline, Zoloft) were evaluated in the daphnid Ceriodaphnia dubia. For each SSRI, the 48-h median lethal concentration (LC50) was determined in three static tests with neonate C. dubia, and chronic (8-d) tests were conducted to determine no-observable-effect concentrations (NOEC) and lowest-observable-effect concentrations (LOEC) for reproduction endpoints. The 48-h LC50 for the SSRIs ranged from 0.12 to 3.90 mg/L and the order of toxicity of the compounds was (lowest to highest): Citalopram, fluvoxamine, paroxetine, fluoxetine, sertraline. Mortality data for the 8-d chronic tests were similar to the 48-h acute data. The SSRIs negatively affected C. dubia reproduction by reducing the number of neonates per female, and for some SSRIs, by reducing the number of broods per female. For sertraline, the most toxic SSRI, the LOEC for the number of neonates per female was 0.045 mg/L and the NOEC was 0.009 mg/L. Results indicate that SSRIs can impact survival and reproduction of C. dubia; however, only at concentrations that are considerably higher than those expected in the environment.     

Absenteeism among employees treated for depression.

Depression-related costs include a relatively large share of indirect costs. We describe the impact of antidepressant treatment on absenteeism among workers diagnosed and treated for depression. Monthly absenteeism counts from employers were summed in the 6 months before and after the initiation of antidepressant therapy in 630 workers treated for depression with a tricyclic antidepressant or a selective serotonin reuptake inhibitor (fluoxetine, sertraline, paroxetine). Monthly mean absenteeism was compared using pairwise t tests. Absenteeism increased before antidepressant initiation and decreased after the treatment began for all antidepressant cohorts. Absenteeism in the selective serotonin reuptake inhibitor cohorts decreased at similar rates for 4 months but was higher in the paroxetine cohort in months five and six after the treatment initiation. Our data suggest that alternative treatments for depression may have differential impact on indirect costs, but further research is warranted.