|
Cost-effectiveness of a disease management program for major depression in elderly primary care patients |
Bosmans J, de Bruijne M, van Hout H, van Marwijk H, Beekman A, Bouter L, Stalman W, van Tulder M |
|
|
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 study investigated a disease management programme for major depression in elderly primary care patients. The programme comprised general practitioner (GP) training in the provision of an integrated programme of screening, diagnosis, drug treatment (20 mg paroxetine once daily), supportive contacts and patient education. Type of intervention Other: Health professional training in screening, diagnosis and treatment. Study population The study population comprised primary care patients, aged 55 years or older. Patients were screened using the Geriatric Depression Scale (GDS-15), using a score of 5 or more to identify those patients needing further evaluation. The mood module of the Primary Care Evaluation of Mental Disorders (PRIME-MD) was used for diagnosis; patients with major depression were deemed eligible for the study. Exclusion criteria were applied. These were current use of antidepressants, current psychosis, bipolar disorder, alcohol or drug abuse, severe social dysfunction, an inability to communicate in Dutch and impaired cognitive functioning. GPs in the intervention group had to agree with the diagnosis of major depression and be willing to prescribe an antidepressant. Setting The setting was primary care. The economic study was carried out in the Netherlands. Dates to which data relate The dates to which the data relate were not reported. However, screening for eligible patients took place between June 2000 and October 2002. The price year was 2002. Link between effectiveness and cost data The economic evaluation was conducted prospectively alongside the clinical trial, in the same sample of patients. Study sample There were no significant differences among GPs in the intervention and control groups with respect to age, gender, experience, or type and size of practice. In power calculations, it was estimated that 68 patients would be needed in each group to detect moderate clinical effects. A total of 3,937 consecutive patients were screened, of which 579 patients were further evaluated and 178 were eligible for the study (i.e. had major depression). One hundred and forty-five patients gave informed consent, 70 forming the intervention group and 75 the usual care group. Sixteen eligible patients were not included since the GP disagreed with the diagnosis or refused to prescribe antidepressants. Full follow-up at 12 months was completed for 125 (86%) patients, 58 in the intervention group and 67 in the usual care group. Patients lost to follow-up were older and more severely depressed than those who completed the study; the most common reason for withdrawal was unwillingness to participate. At baseline, patients did not differ across groups except in marital status. Study design A cluster randomised controlled trial (RCT) design was used. Patients from 34 general practices participated. Randomisation took place at practice level (18 allocated to the intervention, 16 to usual care). Twelve months of follow-up were gathered. More details of the study design were presented in an earlier paper (Bijl et al. 2003, see 'Other Publications of Related Interest' below for bibliographic details). Analysis of effectiveness Clinical measures of effectiveness were severity of depression, recovery from depression and quality of life. In the usual care group, GPs remained blind to the results of the PRIME-MD. Blinded researchers measured outcomes and resource use via home interviews. Recovery from depression was defined as the absence of a PRIME-MD diagnosis at 12 months. The Montgomery Asberg Depression Rating Scale (MADRS) was used to assess changing depression severity during the 12 months of the study. Quality of life was measured using the EuroQol (EQ-5D). Clinical outcomes were measured shortly after screening (T0) and at 2 (T1), 6 (T2) and 12 (T3) months of follow-up, except for the PRIME-MD, which was not measured at T1. Analyses were limited to patients completing all follow-up visits. Effectiveness results Forty-three per cent in the intervention group and 48% in the control group had recovered, based on the PRIME-MD, after 12 months; the mean difference was -4.7 (95% confidence interval, CI: -22.5 to 13.1; p=0.60).
The mean improvement in the MADRS score was -7.8 (standard deviation, SD=9.0) in the intervention group and -7.2 (SD=9.0) in the control group; the mean difference was -0.6 (95% CI: -3.8 to 2.6; p=0.70).
The mean quality-adjusted life-years (QALYs) gained were 0.65 (SD=0.19) in the intervention group and 0.70 (SD=0.21) in the control group; the mean difference was -0.04 (95% CI: -0.11 to 0.03; p=0.20). Clinical conclusions Differences in clinical outcomes between the intervention and control groups were small and not statistically significant. Measure of benefits used in the economic analysis The benefits used in the economic analysis were rates of recovery, improvement in depression severity and QALYs. QALYs were calculated from the EQ-5D scores and the time spent in each state, linearly interpolated. Direct costs Health service and patient health care and non-medical care direct costs were evaluated. Resource use was measured over a 12-month period using patient home interviews, with a recall period of 6 months, and valued using standard national costs. All health care costs were included, with no attempt made to separate depression-related costs. The exception was medication costs, where only psychotropic medications were included. The total cost of the GP training sessions in the intervention group was allocated equally among participating intervention patients. Complementary medicine visits were costed on the basis of patient estimates. The costs were inflated to 2002 levels using consumer price indices. Statistical analysis of costs Mean and SDs were calculated for resource use and cost categories. Indirect Costs Productivity costs were not included in the analysis because it was assumed that most patients were not of working age. Currency US dollars ($). A conversion rate to euros (EUR) was given: $1 = EUR 0.80. Sensitivity analysis Uncertainty around mean differences in costs and cost-effectiveness were analysed using bias-corrected bootstrapping methods. Estimated benefits used in the economic analysis See the 'Effectiveness Results' section. Cost results The total costs were $2,123 in the intervention group and $2,259 in the control group.
The mean difference was -$136 (95% CI: -1,194 to 1,110; p=0.82). Synthesis of costs and benefits Cost-effectiveness planes were presented for all three comparisons (recovery, improvement in severity and QALYs gained at 12 months). These indicated no statistically significant difference in cost-effectiveness between the two groups. Authors' conclusions The authors concluded that the disease management programme had no significant relationship with clinical outcomes, costs or cost-effectiveness. CRD COMMENTARY - Selection of comparators Unrestricted usual care was the selected comparator. The reader could consult Dutch guidelines for depression care as an indicator of the components of usual care in the Netherlands, and should decide whether this represents a useful comparator in their own setting. Validity of estimate of measure of effectiveness The analysis was based on a clustered RCT design, which was appropriate for the intervention. Most patients detected by the screening method as having "major depression" were only "mildly" or "moderately" depressed, according to their MADRS scores at T0, which raises questions about the representativeness of the sample or the definition of the population studied. The methodology reported suggested that the internal validity of the study was reasonable for a design of this type. The authors noted that patients in the usual care group were not blinded, which may have led to them seeking care for depression from their GPs. They also acknowledged a potential source of selection bias in the intervention group (i.e. that GPs had to agree with the diagnosis and prescribe antidepressants), but demonstrated that differences in severity at baseline were not significant. The authors appropriately estimated sample size to ensure a specified level of power in the outcome analysis. However, the final sample size was less than the estimated requirement and drop-out was not random. The impact on the validity of the study results was discussed, and the authors conducted a missing value analysis in which the results did not differ from the base-case. Validity of estimate of measure of benefit The authors used three measures of benefit. Recovery rate and improvement in severity were derived directly from the effectiveness study. QALYs were calculated from EQ-5D measurements performed during the effectiveness study. The methods of calculating the utility weights of the EQ-5D were not adequately reported. The authors provided measures of benefit that were both disease-specific and generic, which may be relevant to different audiences. Validity of estimate of costs From the health care system perspective, all the relevant costs appear to have been included. The authors stated that all direct health care and non-health care costs were included, but this wording is somewhat misleading as it can be seen from the table of resource use that "non-healthcare" referred to complementary medicine only, rather than all non-health care costs incurred by the patient during the study. The quantities and unit costs were presented in full, which greatly aids the generalisability of the results. It was not clear why the authors had converted their costs into US$, nor was the source of the exchange rate given. Uncertainty was appropriately investigated using bootstrap methods to produce confidence intervals. The study was acknowledged to be underpowered to detect relevant differences in costs, but the authors stated that this is common because it is unethical to increase study sample size beyond that needed to demonstrate clinical effectiveness. The authors discussed other costing limitations: the method of cost data collection and potential recall bias; and the exclusion of productivity and informal care costs, the second of which could be high in this population. Other issues The authors compared their findings with those from other studies and presented potential explanations for the differences seen. Generalisability was not addressed, but sufficient detail was available to judge whether the study or its results could be transferred to another setting. The authors do not appear to have presented their results selectively. Given the uncertainty in the disease severity of the patient population examined, it is possible that the sample did not reflect the population to be studied, or, alternatively, that the description of the population studied should have been refined - a review might affect the scope of the conclusions. Implications of the study The authors recommend that GPs continue to provide usual care (watchful waiting), and that future research should focus on improvement in the detection of clinically important and treatable depression and the cost-effectiveness of other interventions for major depression. Source of funding Supported by the Dutch Health Care Insurance Board. Bibliographic details Bosmans J, de Bruijne M, van Hout H, van Marwijk H, Beekman A, Bouter L, Stalman W, van Tulder M. Cost-effectiveness of a disease management program for major depression in elderly primary care patients. Journal of General Internal Medicine 2006; 21(10): 1020-1026 Other publications of related interest Because readers are likely to encounter and assess individual publications, NHS EED abstracts reflect the original publication as it is written, as a stand-alone paper. Where NHS EED abstractors are able to identify positively that a publication is significantly linked to or informed by other publications, these will be referenced in the text of the abstract and their bibliographic details recorded here for information.
Bijl D, van Marwijk HW, Beekman AT, et al. A randomized controlled trial to improve the recognition, diagnosis and treatment of major depression in elderly people in general practice: design, first results and feasibility of the West Friesland Study. Primary Care Psychia 2003;135-40. Indexing Status Subject indexing assigned by NLM MeSH Aged; Cost-Benefit Analysis /economics; Depressive Disorder, Major /economics /therapy; Disease Management; Female; Humans; Male; Middle Aged; Primary Health Care /economics /methods; Treatment Outcome AccessionNumber 22006001826 Date bibliographic record published 26/09/2006 Date abstract record published 09/08/2008 |
|
|
|