|
Cost-effectiveness of collaborative care for depression in a primary care veteran population |
Liu C-F, Hedrick S C, Chaney E F, Heagerty P, Felker B, Hasenberg N, Fihn S, Katon W |
|
|
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 A collaborative care intervention for depression was compared with consult-liaison care.
Economic study type Cost-effectiveness analysis.
Study population The study population comprised patients in a Department of Veterans Affairs (VA) primary care clinic. The VA system serves a primarily ageing male population in a lower socioeconomic bracket. To be eligible for the study, patients had to have a current major depressive episode and/or dysthymia. The authors stated that they limited the exclusion criteria to the extent necessary to maximise the generalisability of the results.
Setting The setting was primary, secondary and community care. The economic study was carried out in the general internal medicine clinic (GIMC) of Seattle VA Medical Center, USA.
Dates to which data relate The paper suggested that the patients were enrolled into the study between January 1998 and March 1999. The follow-up period was 9 months, giving a finish date of December 1999. The effectiveness and resource use data were collected during this period. The cost estimates were adjusted to year 2000 US dollars.
Source of effectiveness data The evidence for the final effectiveness outcomes was derived from a single study.
Link between effectiveness and cost data The costing was undertaken retrospectively on the same patient sample as that used in the effectiveness study.
Study sample The sample size was not determined in the planning phase of the study to assure a certain power. The authors recognised that the sample was not sufficiently large to provide statistical power to detect differences in the outcomes and costs. The authors reported that several methods were used to locate eligible patients. These included referral from two ongoing unrelated studies, a prevention survey undertaken at the clinic and referral by primary care providers. After the initial screening, each prospective participant was given a computer-assisted structured interview to assess the following:
severity of depression,
current and past use of medication or therapy,
health status,
current and past alcohol use,
symptoms of post-traumatic stress disorder (PTSD),
history of metal illness, and
barriers of care.
The assessment of depression and anxiety symptoms was based on the PRIME-MD, with additional questions taken from the Structured Clinical Interview for DSM-IV. The authors did not fully justify the choice of the patient sample in terms of the characteristics of the treatment under investigation. They reported that the exclusion criteria were limited in order to maximise the generalisability of the results. The study enrolled 354 patients in total. There were 168 patients in the collaborative care group and 186 in the consult-liaison care group.
Study design This appears to have been a non-randomised controlled trial as no details were given of any randomisation process. The study took place at a single centre (the GIMC of Seattle VA Medical Center, USA). The two groups were followed up for 9 months. Twenty-two patients in each of the two groups were lost to follow up. The authors did not report the reasons for loss to follow-up.
Analysis of effectiveness The basis of the analysis was treatment completers only. The primary health outcome used in the analysis was the number of depression-free days. Symptom Check-list-90 (SCL) depression scores from baseline and follow-up assessments were used to calculate the number of depression-free days. The authors reported that the number of depression-free days was calculated by taking depression severity data from two consecutive outcome assessments to estimate the severity of depression for each day during the interval by linear interpolation. Days for which the SCL depression score was 0.5 or less were considered depression free. Days for which the SCL depression score was 2.0 or greater were considered fully symptomatic. Days with intermediate severity scores were assigned a value between depression free and fully symptomatic by linear interpolation. The authors explained that a day with an SCL score of 1.25 would be considered to be 50% depression free. Confidence intervals (CIs) for depression-free days were estimated by bootstrapping with 1,000 replications.
The authors reported that there were no significant differences in the baseline demographic characteristics, SCL depression scores, or costs during the year prior to the enrolment of the two groups. However, it was noted that patients receiving collaborative care were more likely to have had depressive episodes than patients receiving consult-liaison care, 56% versus 48%, (p<0.05).
Effectiveness results The mean plus/minus standard deviation (SD) number of depression-free days was 112.7 (+/- 81.1) for the collaborative care group and 107 (+/- 75.6) for the consult-liaison care group.
After adjusting for baseline characteristics, the difference between the two groups was 14.5 days (95% CI: -0.5 - 29.6; p=0.59).
The proportion of depression-free days at baseline was 37% in the collaborative care group and 36% in the consult-liaison group. At 3 months the proportion was 43% (collaborative care) versus 36% (consult-liaison care), and at 9 months the proportion was 42% versus 41%.
The patients receiving collaborative care were significantly more likely to have a primary care depression visit than those receiving consult-liaison care (84% versus 56%). They were also significantly more likely to be given a prescription for an antidepressant (80% versus 61%).
It was found that the utilisation pattern reflected the study interventions. The patients in the collaborative care group were more likely to visit a primary care provider for depression treatment than those in the consult-liaison group (77% compared with 39%), with significantly more visits per patient (2.57 +/- 3.82 compared with 1.63 +/- 4.61).
Twenty-four per cent of the patients in the collaborative care group received cognitive-behavioural therapy.
By intervention design, the patients who received collaborative care were less likely to be referred to GIMC psychiatrists than those in consult-liaison care (3% compared with 35%), with significantly fewer mean visits per patient (0.05 +/-0.37 compared with 1.80 +/-2.5).
Clinical conclusions The study found that the collaborative care model was associated with a modest increase in time free of depression when compared with the consult-liaison model. However, the observed difference was not found to be statistically significant.
Modelling Adjusted differences between collaborative care and consult-liaison care were estimated using regression models with bootstrap interval estimates. All models were also adjusted for interclustering correlation at the provider level, using Huber's estimator from a robust regression.
Measure of benefits used in the economic analysis No summary measure of benefit was used. The number of depression-free days was used as a measure of benefit in the economic analysis (see 'Effectiveness Results' section).
Direct costs The resource quantities and the costs were not reported separately. The unadjusted cost of care per patient for 9 months after baseline was reported for total depression costs, total outpatient service costs and total costs. The total depression costs were shown by primary care visits, antidepressant medications, mental health specialty visits and intervention costs (shown by social work follow-up calls and team treatment meetings). Data on the use of VA care in the 9-month follow-up period were obtained from the local VA data warehouse. The costs of care were estimated using national average cost estimates from the Cost Distribution Report (CDR), the VA's cost accounting system. The CDR provides an average cost per visit in a specific clinic. Other costs (i.e. team meeting and follow-up patient telephone calls) were estimated by sampling staff activity records, then computing the average cost on the basis of event duration by using actual input costs (e.g. labour, fringe benefits and overheads). Discounting was not relevant since the study period was only 9 months. The study reported the average costs. Dates to which the costs referred were not stated explicitly within the paper.
Statistical analysis of costs CIs for cost measures were estimated by bootstrapping with 1,000 replications. No further statistical analyses of the costs were carried out.
Indirect Costs Although it was stated in the paper that the cost estimates included the indirect costs, none appear to have been included in the analysis.
Sensitivity analysis A sensitivity analysis was not carried out.
Estimated benefits used in the economic analysis No summary benefit measure was used. However, the mean (+/- SD) number of depression-free days was 112.7 (+/- 81.1) for the collaborative care group and 107 (+/- 75.6) for the consult-liaison care group.
Cost results Over the 9-month study period, the total costs were $7,946 (95% CI: 5,582 - 10,310) for collaborative care compared with $6,789 (95% CI: 4,720 - 8,585) for consult-liaison care.
The costs of adverse effects or knock-on costs were not dealt with in the costing.
The adjusted, incremental total costs of a collaborative care intervention for veterans with depression were $169 (95% CI: -1,851 - 2,453).
Synthesis of costs and benefits The estimated benefits and costs were combined as the cost per additional depression-free day. An incremental analysis was performed.
The adjusted, incremental total cost per additional depression-free day was $2 (95% CI: -254 - 398). The adjusted, incremental depression treatment cost per additional depression-free day was $24 (95% CI: -105 - 148). The adjusted, incremental total outpatient cost per additional depression-free day was $33 (95% CI: -106 - 232).
Authors' conclusions The collaborative care model was associated with modest increases in time free of depression and in treatment costs over the 9-month study period. However, the difference in the number of depression-free days between the groups was not statistically significant.
CRD COMMENTARY - Selection of comparators The authors did not provide an explicit justification for their choice of the comparator, although it would appear to represent traditional practice in their setting. You should decide if this is a widely used health technology in your own setting.
Validity of estimate of measure of effectiveness The analysis was based on results from a controlled trial. This was appropriate for the study question, although a randomised controlled trial may have been preferable. It was not explicitly stated whether the study sample was representative of the study population. The authors stated that the patient groups were comparable at baseline. All the comparisons of cost and utilisation took the full study sample into consideration, although the outcomes were analysed for treatment completers only.
Validity of estimate of measure of benefit The estimation of benefits was obtained directly from the effectiveness analysis. The choice of estimate was not justified.
Validity of estimate of costs The authors reported that the study did not include utilisation and cost data for services obtained outside the VA. They also reported that the VA cost accounting method provides an average cost per clinic visit, which does not capture variation in outpatient cost across visits with different intensities of care. The costs and the quantities were not reported separately. Although it was stated in the paper that the indirect costs were recorded, none appear to been reported in the paper. The resource use data were taken from the local VA data warehouse and the costs were taken from the authors' own setting. Statistical analyses of the quantities and prices were not performed. Charges were not used to proxy prices. The date to which the prices related was not reported explicitly.
Other issues The authors made appropriate comparisons of their findings with those from other studies. In addition, the issue of generalisability to other settings was addressed. Although only limited results were presented, the authors do not appear to have presented these selectively. The study considered a veteran population and this was reflected in the authors' conclusions. The authors reported a number of further limitations to their study. First, the results were obtained from a single site with existing mental health integration in primary care, and so may not be generalisable to other primary care settings. Second, depression-free days were calculated on the basis of the SCL depression scale rather than the Hamilton Depression Scale that had been used in prior studies.
Implications of the study The authors did not make any recommendations for changes in policy or practice, and/or the need for further research.
Source of funding Supported by grant II-R-95-097 from the Health Services Research and Development Service of the Department of Veterans Affairs.
Bibliographic details Liu C-F, Hedrick S C, Chaney E F, Heagerty P, Felker B, Hasenberg N, Fihn S, Katon W. Cost-effectiveness of collaborative care for depression in a primary care veteran population. Psychiatric Services 2003; 54(5): 698-704 Indexing Status Subject indexing assigned by NLM MeSH Cost-Benefit Analysis; Depressive Disorder /economics /therapy; Female; Health Care Costs /statistics & Hospitals, Veterans /economics /statistics & Humans; Male; Mental Health Services /organization & Middle Aged; Patient Care Team /organization & Primary Health Care /economics /statistics & Veterans /psychology; Washington; administration; administration; numerical data; numerical data; numerical data AccessionNumber 22003009609 Date bibliographic record published 30/11/2004 Date abstract record published 30/11/2004 |
|
|
|