|
Cost-effectiveness of a primary care intervention for depressed females |
Pyne J M, Smith J, Fortney J, Zhang M, Williams D K, Rost K |
|
|
Record Status 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. Health technology The use of enhanced care versus normal or usual care for the treatment of depression. Enhanced care provided additional training of office nurses, to supplement the primary care from physicians.
Study population The study population comprised male and female patients who met American Psychiatric Association criteria (DSM-IV) for current major depression. Patients were eligible if they reported at least five criteria for major depression weeks on the Inventory to Diagnose Depression in the past 2 or had screened negative for lifetime mania. They were also eligible if they had screened negative for alcohol dependence with current drinking, or did not meet DSM-IV criteria for bereavement-related depression.
Setting The setting was primary care clinics. The economic analysis was conducted in the USA.
Dates to which data relate The effectiveness data related to 1996 to 1997. The cost data were estimated using 1997 health care finance administrative data and 1999 average Medicare payment rates. Prices relating to the year 2000 were used.
Source of effectiveness data The effectiveness data were gathered from a single prospective study.
Link between effectiveness and cost data The costing appears to have been undertaken prospectively on the same group as that used in the effectiveness study.
Study sample No power calculations to determine the sample size were reported. A total of 479 patients were recruited for the study. Of these, 211 (44.1%) were beginning a new treatment episode at the start of the intervention, 178 females and 33 males. Males and females were not statistically different at baseline in terms of depression severity, SF-36 visual analogue scale (VAS) score and antidepressant acceptability.
Study design The study was a randomised controlled study that was conducted in 12 primary care practices. The 12 practices were matched into six pairs, on the basis of the participating physicians' baseline depression treatment patterns, using a stratified blocking procedure. One practice in each pair was randomised to the enhanced care condition, whilst the other practice delivered usual care to the study participants. Two physicians and one administrative staff member participated in each practice. One office nurse also participated in the intervention practice. None of the practices employed on-site mental health specialists. Nurse care managers completed an additional 8-hour training session compared with the physicians.
The participants were recruited to the intervention and control groups in stages during 1996 to 1997. This process was described in an earlier article (Rost et al., see Other Publications of Related Interest). The patients completed structured telephone interviews to collect the effectiveness data. The interviewer was blinded to the patients' treatment conditions, while physicians in usual care practices were not informed as to which patients were participating in the study. The time horizon was 12 months. No loss to follow-up was reported.
Analysis of effectiveness The basis of the analysis was intention to treat. Antidepressant acceptors beginning a new depression treatment episode were the focus of the analysis. The analysis used a random-effects model, in which the patients were nested within practices, to test for interclass correlation. The effectiveness measures were the VAS SF-36 scores and the depression severity measurement (mCES-D). The VAS score conversion formulae came from a published source, whereas the quadratic terms added for physical functioning and the mental health index did not. With the exception that females tended to be younger and have lower adjusted household incomes than males, the two groups (females and males) were shown to be clinically and sociodemographically comparable at baseline.
Effectiveness results The VAS SF-36 and mCES-D results were not reported.
The authors did not report the effectiveness results, even though the VAS SF-36 scores were used to derive the benefit measure used in the economic analysis.
Clinical conclusions No effectiveness results were reported and the clinical conclusions were drawn from the benefit measure (see 'Author's Conclusions' section).
Measure of benefits used in the economic analysis The measure of benefits in the economic analysis was the quality-adjusted life-years (QALYs) saved. SF-36 scores for the medical outcomes study were converted into QALYs to compare the one-year effectiveness of enhanced versus usual care, according to gender.
Direct costs The analysis was conducted from a societal perspective. The direct costs in the main analysis were for the intervention without training, health care expenditures, time and transportation. The intervention costs were for patient screening, preparation, post-session record keeping, communication and post-training supervision. Health care expenditures included the emergency room, outpatient primary care and speciality care, and psychotropic medication. Time costs considered travel to and from the medical services, waiting time and time with the clinician. The sensitivity analyses included the costs of the hospital and intervention training. Training costs included trainee time, training manual costs and trainer time.
The costs of inpatients were estimated using 1997 health care finance administrative data on the average per diem Medicare payment. The costs for outpatient clinician visits were also estimated using 1999 average Medicare payment rates. Prices relating to the year 2000 were used. Time costs were calculated using self-reported wage rates. For unemployed patients, the federal minimum wage was used as a proxy of their time costs. Discounting was unnecessary because of the 12-month time horizon. The costs and the quantities were not reported separately.
Statistical analysis of costs The authors analysed the natural log transformation of health care costs using the one-part smearing retransformation to calculate predicted costs in the main analysis. Sociodemographic and clinical covariates were collected and used in the cost models.
Indirect Costs The indirect costs were not included in the main analysis, but were included in the sensitivity analysis. These included productivity costs, which were estimated from the cost of lost wages. The authors used 0.4 as the fraction of a workday missed for health reasons and the same base wage rates described above. Discounting was unnecessary.
Sensitivity analysis Four separate univariate sensitivity analyses were carried out. The first analysis included the hospital costs, the second included the training costs for the intervention, the third included the productivity costs (indirect costs), and the fourth included the original SF-36 to QALY conversion formula.
Estimated benefits used in the economic analysis Among antidepressant nonacceptors, the interaction term between gender and enhanced care in predicting QALYs was non significant (interaction term beta=0.03, t=0.35, P=0.73). This supported the decision not to pursue further analyses with antidepressant nonacceptors. Among antidepressant acceptors, the interaction term between gender and enhanced care in predicting QALYs was significant (beta=-0.20, t=-2.86, P=0.005).
For female antidepressant acceptor patients, the mean QALY score was 0.595 among those who received enhanced care and 0.472 among those who received usual care. Enhanced care was therefore associated with a significant saving of 0.123 QALY in comparison with usual care.
For male antidepressant acceptor patients, the mean QALY score was 0.473 among those who received enhanced care and 0.546 among those who received usual care. Enhanced care was therefore associated with a non significant loss of 0.073 QALY in comparison with usual care.
The economic analysis also considered physical and psychological side effects, in order to explain the differential effect of the intervention according to gender.
Cost results The mean annualised cost for females antidepressant acceptors was $2,316 (95% confidence interval CI: 1,803 - 2,975) in enhanced care and $1,671 (95% CI: 1,269 - 2,199) in usual care. Therefore, the mean incremental cost of the intervention was $645 in comparison with usual care.
The mean annualised cost for males was $2,239 (95% CI: 1,354 - 3,700) in enhanced care and $2,249 (95% CI: 1,253 - 4,036) in usual care. Therefore, the mean incremental cost of the intervention was -$10 in comparison with usual care.
Synthesis of costs and benefits The mean incremental cost-effectiveness ratio (ICER) of the intervention for females was $5,244 per QALY saved.
No synthesis of the costs and benefits was performed for males, as the mean incremental cost was essentially zero and the mean incremental QALYs were neutral.
With additional hospital costs, the ICER of the intervention was $8,195 per QALY saved (1st sensitivity analysis) for females. For males, the mean incremental cost was more negative (-$506).
With additional intervention training costs, the ICER of the intervention was $9,740 per QALY saved for females and $4,644 per QALY lost for males (2nd sensitivity analysis).
With additional productivity costs, the ICER of the intervention was $5,171 per QALY saved for females, $15,068 per QALY lost for males (3rd sensitivity analysis).
Without the quadratic terms in the QALY measurement, the ICER of the intervention was $4,886 per QALY saved for females (4th sensitivity analysis).
Authors' conclusions For female patients, enhanced care was more expensive and more effective than usual care. For male patients, enhanced care was cost and outcome neutral in comparison with usual care. The estimated cost-effectiveness ratio of enhanced care was within the acceptable range for females, but not for males.
CRD COMMENTARY - Selection of comparators The choice of the comparator used was justified on the grounds that it represented the current practice in the authors' setting. You should consider whether this is a widely used health technology in your own setting.
Validity of estimate of measure of effectiveness The estimate of effectiveness is likely to have been internally valid given that a blind, randomised controlled trial was used. However, the sample size was very small, particularly in males (n=33), which suggested a lack of power. Power calculations to ascertain a sample size were not reported. Appropriate statistical analyses were undertaken to take account of potential biases and confounding factors. No effectiveness results were reported.
Validity of estimate of measure of benefit The estimation of benefits was directly obtained from the effectiveness analysis. This choice of effectiveness (QALY) was justified. Appropriate methods were used to measure the health-related quality of life and the QALY saved. A limitation, as highlighted by the authors, was that the QALY conversion formula was derived from a relatively small and unrepresentative sample.
Validity of estimate of costs A societal perspective was adopted, although the indirect costs were not introduced until the sensitivity analysis. The costs and the quantities were not reported separately and only limited details were given on the unit costs and quantities. These facts hinder the reproducibility of the results to other settings. However, the price year was reported, thus adding reflation exercises. A statistical analysis was performed. In addition, a sensitivity analysis that included both the hospital costs and indirect costs was conducted.
Other issues The generalisability of the results was not specifically discussed. Adequate comparisons were made with studies dealing with the same topic. The study enrolled patients with current major depression and this was reflected in the authors' conclusions. The authors highlighted the limitations of their study. They do not appear to have reported their results selectively.
Implications of the study The findings suggest that interventions to improve primary care treatment of depression may need to be modified to improve their effectiveness in males, while maintaining their effectiveness in females. The authors recommend further studies with larger sample sizes to help clarify the issue and to examine further the differential gender response to depression interventions.
Source of funding Supported by a VA Career Development Award and the VISN 16 MIRECC (Dr. Pyne), and grants MH63651 and MH54444 (Dr. Rost), MH54444 (Mr. Smith) and AA12085 (Dr. Fortney).
Bibliographic details Pyne J M, Smith J, Fortney J, Zhang M, Williams D K, Rost K. Cost-effectiveness of a primary care intervention for depressed females. Journal of Affective Disorders 2003; 74(1): 23-32 Other publications of related interest Brazier J, Usherwood T, Harper R, Thomas K. Deriving a preference-based single index from the UK SF-36 health survey. Journal of Clinical Epidemiology 1998;51:1115-28.
Rost K, Nutting P, Smith J, Werner J. Designing and implementing a primary care intervention trial to improve the quality and outcome of care for major depression. General Hospital Psychiatry 2000;22:66-77.
Indexing Status Subject indexing assigned by NLM MeSH Adult; Cost-Benefit Analysis; Depressive Disorder, Major /economics /prevention & Female; Humans; Male; Mental Health Services /economics; Middle Aged; Multivariate Analysis; Primary Health Care /economics; control AccessionNumber 22003009245 Date bibliographic record published 30/11/2003 Date abstract record published 30/11/2003 |
|
|
|