This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn.

Pharmaceutical: Tricyclic antidepressants (TCAs), selective serotonin reuptake inhibitors (SSRIs).

Treatment.

Cost-effectiveness analysis.

Patients overdosing on either fluoxetine (SSRI) or tricyclic antidepressants (TCAs).

Hospital (emergency departments and intensive care/medical units). The economic study was carried out at 9 medical centres in the USA.

Effectiveness data were collected over the period October 1990 to March 1993. 1993 prices were used.

Effectiveness data were derived from a single study.

Costing was undertaken prospectively on the same patient sample as that used in the effectiveness study.

140 patients (124 with a TCA overdose and 16 with a fluoxetine overdose) from 622 identified patients met the inclusion criteria for the study. However, a further 4 patients were excluded from the economic analysis because of missing or incomplete bills, resulting in 136 patients: 121 (89%) in the TCA group and 15 (11%) in the fluoxetine group. Patients were included if they had ingested a single antidepressant without clinically significant co-ingestants, and if the overdose had been confirmed by laboratory testing. Patients were excluded if they had co-ingested trazodone, amfebutamone (bupropion) or monoamine oxidase inhibitors. There were no significant differences between the TCA and fluoxetine groups with regards to age (mean age 29.8 years +/- 1.2 for TCA; 25.7 years +/- 3.0 for fluoxetine), transport by ambulance to the emergency department (ED), or time since ingestion of the drug to ED arrival. It was not stated whether power calculations determined the sample size.

This was a prospective multicentre (9 medical centres) cohort study.

The analysis of the study was based on treatment completers only. The health outcomes used in the analysis included level of consciousness, length of stay, cardiopulmonary complications, vital signs or ECG abnormalities, agitation, seizures, central nervous system depression and death.

On arrival in the ED, more patients with a fluoxetine overdose were alert than patients with a tricyclic antidepressant (TCA) overdose (93.8% versus 27.4%, respectively; p<0.001). No seizures, comas or deaths were observed in the fluoxetine overdose patients and no fluoxetine overdose patients were admitted to an intensive care unit (ICU). In contrast 82.4% of the TCA overdose patients utilised ICU services, (p<0.001). Also fluoxetine overdose patients were less likely to be intubated (0% versus 46.3%; p=0.001) or have tachycardia (13.3% versus 50.4%; p=0.007). Mean length of stay (LOS) was 0.73 days (+/- standard error of the mean (SEM) 0.33) for fluoxetine overdose patients and 3.59 days (+/- SEM 0.48) for TCA overdose patients, (p=0.038).

With respect to minimising the adverse events associated with overdose in the treatment of depression, fluoxetine is a safer medication than tricyclic antidepressants.

The outcomes assessed were level of consciousness, length of stay, cardiopulmonary complications, vital signs or ECG abnormalities, agitation, seizures, central nervous system depression and death.

Quantities and costs were not analysed separately. Only health service costs were considered which included total and departmental hospital charges, physician charges, and estimated hospital and physician service costs. The primary source for costs was the Universal Bill (UB-82) summary hospital bill, a standard billing form that summarises hospital charges by cost centre (e.g. room & board, pharmacy, supplies, laboratory). In patients for whom UB-82s were not retrievable, detailed hospital bills or total hospital charges were used. Hospital charges were adjusted to costs using Health Care Financing Administration (HCFA) cost:charge ratios for each participating hospital. Charges for physician services, which were not incorporated in the hospital bill, were included when available. Charges for physician services during the period of acute hospitalisation were transformed into estimated costs, assuming that actual payments from insurers represented 90% of charges. All costs were adjusted to 1993 US dollars using the medical sector of the consumer price index.

Statistical analysis was performed on log-transformed hospital, physician and total medical costs using ordinary least squares regression analysis and general linear models to evaluate differences after controlling for patient gender and age. Non-parametric techniques (e.g. Wilcoxon rank-sum test) were used to evaluate differences in untransformed medical costs between the groups.

Not stated.

US dollars ($).

No sensitivity analysis was performed.

On arrival in the ED, more patients with a fluoxetine overdose were alert than those with a tricyclic antidepressant (TCA) overdose (93.8% versus 27.4%, respectively; p<0.001). No seizures, comas or deaths were observed in the fluoxetine overdose patients and no fluoxetine overdose patients were admitted to an intensive care unit (ICU). In contrast 82.4% of the TCA overdose patients utilised ICU services , (p<0.001). Also fluoxetine overdose patients were less likely to be intubated (0% versus 46.3%; p=0.001) or have tachycardia (13.3% versus 50.4%; p=0.007). Mean length of stay (LOS) was 0.73 days (+/- SEM 0.33) for fluoxetine overdose patients and 3.59 days (+/- SEM 0.48) for TCA overdose patients, (p=0.038).

Mean hospital costs were $668 and $4,691 for the fluoxetine and tricyclic (TCA) groups respectively, (p=0.044). These differences were also statistically significant when log-transformed costs were analysed (p<0.001) and a Wilcoxon rank-sum test was performed (p<0.001). Fluoxetine overdoses were also associated with significantly lower median costs ($470) than TCA overdoses ($2,984), (p<0.001). There were no significant differences in physician costs between the two groups, although mean physician costs for fluoxetine overdoses ($520) were lower than TCA overdoses ($1,316). In terms of total medical costs, the mean was $1,269 for the fluoxetine group and $5,764 for the TCA group. Analysis of untransformed medical costs was not significant, (p=0.118), but there were significant differences in log-transformed medical costs (p<0.001) and when the Wilcoxon rank-sum test was performed (p<0.001). Median total medical costs were $805 and $3,524 for the fluoxetine and TCA groups respectively, (p=0.002).

Costs and benefits were not combined. Fluoxetine overdose is dominant over TCA overdose as it is associated with lower costs and lower adverse effects.

The authors concluded that fluoxetine overdoses resulted in lower hospital and total medical costs compared with TCA overdoses. Although the cost of physician services associated with fluoxetine overdoses were less than 50% of TCA overdoses, these differences did not reach statistical significance. The clinical toxicity of TCAs leads to increased use of hospital services and additional costs to the health care system and some evidence exists to suggest that, as physicians became more experienced with managing fluoxetine overdoses, the medical costs associated with these overdoses decreases.

ustification was given for the comparators used. The comparators chosen, fluoxetine(SSRI) and TCAs, are both commonly used treatments for major depression.

study was based on a cohort study. However, the fluoxetine group only contained a small number of patients (n=15) and no treatment protocol was specified, which could potentially have an effect on the costs.

quate details were given of the sources of estimates and prices and the price date. However, clinical management of an antidepressant overdose results in costs that constitute only a small fraction of the total economic perspective of depression. Significant medical costs may be associated with outpatient services, psychological counselling and medications.

cost data may not be generalisable to other settings or countries. The authors acknowledged that individuals who attempt suicide by overdosing using antidepressants also ingest other drugs and alcohol, which can have an effect on clinical management and costs.

If it is the priority of the physician, in treating major depression, to minimise the adverse events associated with possible overdose, then the recommended medication is fluoxetine vis a vis tricyclic antidepressants.

Grant from Lilly Research Laboratories, Indianapolis, Indiana, USA.