|
Patient self-management of anticoagulation therapy: a trial-based cost-effectiveness analysis |
Jowett S, Bryan S, Murray E, McCahon D, Raftery J, Richard Hobbs F D, Fitzmaurice D |
|
|
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 compared two management strategies, patient self-management (PSM) and usual care, for patients on long-term anticoagulation therapy. Under the PSM strategy, patients carried out international normalised ratio (INR) tests themselves using reliable portable machines and adjusted their dose using a dose algorithm. Under the usual practice strategy, patients attended a hospital outpatient clinic or primary care clinic for evaluation of the INR.
Study population The study population comprised patients with a long-term indication for oral anticoagulation. No further inclusion or exclusion criteria were reported.
Setting The settings were primary care, secondary care and the community. The economic study was carried out in the UK.
Dates to which data relate The dates to which the effectiveness data referred were not reported in the current study. Details are given in another study (Fitzmaurice et al. 2005, see 'Other Publications of Related Interest' below for bibliographic details). The cost data were derived from sources published between 2002 and 2004. The price year was 2003.
Source of effectiveness data The effectiveness data were derived from a single study.
Link between effectiveness and cost data It appears that the costing has been carried out prospectively on the same sample of patients as that used in the economic analysis.
Study sample Although full details of the trial were reported in another published study (Fitzmaurice et al. 2005), it appears that the sample size was not determined in the planning phase of the study. It was reported that the study sample referred to non-selected patients. The initial sample comprised 617 patients on long-term oral anticoagulation. These were randomly assigned to either the PSM group (n=337) or the usual care group (n=280). In the PSM group, 242 patients received the intervention and 95 did not. Of those who received the intervention 193 completed the study. Overall, 11 patients in the PSM group were excluded from the analysis. In the usual care group, 275 patients received routine care and 5 did not. Of those who received routine care, 250 completed the study while 25 either interrupted routine care or were lost to follow-up. Overall, 15 patients in the usual care group were excluded. Reasons for exclusion in the two groups were drop out before initiation of data collection, or interruption of warfarin where successive treatment was unknown. It was reported that complete resource data were obtainable for 326 (96.7%) patients in the PSM group and for 265 (94.6%) in the usual care group.
Study design The analysis was based on a UK-based randomised controlled trial. The number of centres included and method of randomisation were not reported in the current study (details given in Fitzmaurice et al. 2005). Patients in the PSM group were followed up for progress evaluation every 3 months by attending a practice-based study clinic. All patients were followed up for one year. It was reported that 49 (14.5%) patients in the PSM group and 25 (9%) in the usual care group were lost to follow-up.
Analysis of effectiveness The analysis was conducted on an intention to treat basis. The patient groups were only compared in terms of their age (see Fitzmaurice et al. 2005 for the comparability of the two groups). The primary health outcome used in the analysis was INR (time in range). Data on INR were collected using a case report form
Effectiveness results It was reported that there were no statistical differences in INR percentage time in range between the two groups.
Based on percentage time in range within the study, INR control was 70% in the PSM group and 68% in the usual care group (difference not statistically significant).
Clinical conclusions INR control (in terms of time in range) did not differ significantly between the two groups.
Measure of benefits used in the economic analysis The authors used health utility (quality-adjusted life-years, QALYs) as the measure of benefit in economic analysis. The QALYs were measured at baseline, and at 6 and 12 months using the EQ-5D questionnaire. QALYs were assigned by estimating the area under the health utility curve using linear interpolation for each patient. It was reported that missing EQ-5D data were handled using multiple imputation, with regression-based adjustments undertaken to account for biases due to differences in baseline EQ-5D scores. The multiple imputation analysis was conducted using NORM version 2.03 (Schafer, 2000).
Direct costs The health service costs included in the analysis were as follows.
For the PSM group: machine, test strip, internal quality control test, lancet, sharps bin, tissues (box of 150), Alcowipes, laminated dosing card, practice nurse phone call (10 minute), practice nurse salary and general practitioner salary when excluding and including training costs, specialist registrar, consultant, administration and clinic overheads, training sessions 1 and 2 (per person), practice test strips or lancets, and assessment clinic (per session).
Usual care costs per visit included: visit to hospital clinic; general practice blood sample, hospital analysis and dosing; general practice blood sample and dosing and subsequent hospital analysis; visit to practice near-patient testing clinic; visit to pharmacist-led practice clinic; and visit to Medical Laboratory Scientific Officer clinic.
It was reported that costs due to minor or serious adverse events were included. The unit costs were reported for all categories of costs. In addition, information on patient transportation to visit a primary, secondary or PSM practice clinic was collected. The costs and the quantities regarding transportation costs were not reported separately. The quantities were based on actual data derived from the single study. Salary costs, costs of medical procedures and adverse events were derived from the literature and from official published sources (e.g. NHS Reference Costs). Capital costs (equipment) were based on purchase prices and were depreciated over a 3- or 5-year period accordingly using an appropriate discount rate. All costs were reported for the price year 2003.
Statistical analysis of costs The costs were treated stochastically. As the cost data were skewed, the authors used the arithmetic mean and the non-parametric 95% confidence interval (CI). Cost-differences in mean values were analysed using independent sample t-tests. The analysis was carried out using the Statistical Package for the Social Science for Windows version 13.0.1 (SPSS Inc., Chicago). Bootstrapping was conducted using STATA version 8.2 (StataCorp LP, College Station) and Microsoft Excel 2002. For all statistical analyses, p-values of less than 0.05 were considered statistically significant.
Indirect Costs The analysis included costs on activities forgone due to clinic visits, such as leisure activities, care of relatives and employment. Productivity losses were estimated using the mean gross weekly wage. For patients not in paid employment, activity forgone was characterised as leisure time and assigned a value of 40% of the mean average wage. The resources of quantities were derived from a random sub-sample analysis of the original study sample, while the costs were derived from official published sources. All costs were reported for the price year 2003.
Sensitivity analysis The authors carried out a complete case analysis (using only patient data with full information on cost and quality of life) in order to assess the impact of multiple imputation. In addition, a one-way sensitivity analysis was carried out to investigate the robustness of the results to variability in the data. The parameters investigated in the sensitivity analysis were an extended machine lifetime of 5 or 10 years, the use of alternative health care professionals for training, and the exclusion of training costs from the analysis. In addition, uncertainty due to sampling variation in cost-effectiveness was addressed by plotting cost-effectiveness acceptability curves (CEACs), which demonstrated the probability of PSM being more cost-effective than usual care for different cost per QALY threshold values.
Estimated benefits used in the economic analysis The multiple imputation analysis demonstrated that the QALYs were 0.712 in the control group (n=265) and 0.721 in the PSM group (n=326). The difference in mean values (PSM - control group) was 0.009 (bootstrapped 95% CI: -0.012 to 0.030). The difference between the two groups was not statistically significant.
The complete case analysis demonstrated that the QALYs were 0.738 in the control group (n=165) and 0.739 in the PSM group (n=204). The difference in mean values (PSM - control group) was 0.001 (bootstrapped 95% CI: -0.027 to 0.032). The difference between the two groups was not statistically significant.
In addition, it was reported that for the anxiety/depression dimension of the EQ-5D there were no statistically significant differences between the two groups regarding changes from baseline, to 6 or 12 months.
Cost results The total per patient NHS costs were 122.32 (95% CI: 103.48 to 143.90) in the control group and 416.76 (95% CI: 393.95 to 441.81) in the PSM group. The difference was statistically significant, (p<0.001).
Accounting for societal costs, the mean societal costs were 179.80 (95% CI: 160.09 to 202.58) in the control group and 462.73 (95% CI: 439.28 to 489.15) in the PSM group. The difference was statistically significant, (p<0.001).
Synthesis of costs and benefits An incremental cost-effectiveness analysis was conducted. When adopting the health care perspective, the multiple imputation analysis demonstrated that the incremental cost-effectiveness ratio (ICER) for PSM was 32,716 per QALY gained. When adopting the societal perspective, the ICER for PSM was 31,437 per QALY gained.
The CEAC analysis demonstrated that when adopting a threshold of 20,000 per QALY gained, the probability that PSM was cost-effective was 30% when using multiple imputation analysis and 16% when using complete case analysis.
When adopting a threshold willingness-to-pay of 30,000 per QALY, the respective probabilities were 46% (imputed analysis) and 26% (complete case analysis). From a societal perspective, the probability was 32% at a threshold of 20,000 per QALY gained and 49% at a threshold of 30,000 per QALY gained.
The sensitivity analyses demonstrated the robustness of the results to the estimates used.
Authors' conclusions Using conventional willingness-to-pay thresholds, the analysis demonstrated that patient self-management (PSM) of anticoagulation was not cost-effective in comparison with usual care.
CRD COMMENTARY - Selection of comparators A justification was provided for the comparators used. Usual care seems to have represented standard practice in the authors' setting. You should decide if this represents a valid technology in your own setting.
Validity of estimate of measure of effectiveness The analysis was based on a randomised controlled trial, which seems to have been appropriate given the study question. The study sample was representative of the study population. However, it was not clear whether the patient groups were comparable because no details of the patients were provided. It is not possible to comment on the internal validity of the effectiveness results since the authors referred to a separate clinical paper for details of the clinical study. Appropriate statistical analysis was undertaken to take potential biases and confounding factors into account. However, no power calculations, to ascertain whether the results obtained were due to the intervention or to chance, were reported.
Validity of estimate of measure of benefit The measure of benefit used was the health utility (QALYs), which was measured over 12 months using the EQ-5D questionnaire. This might not have captured all relevant aspects since it is a generic instrument, although it does allow cross-programme comparisons.
Validity of estimate of costs The analysis of the costs was performed from the societal as well as the health care provider (NHS) perspective. It appears that all the relevant categories of costs have been included in the analysis. Although the unit costs were reported for all cost categories, with the exception of transportation and indirect costs, the quantities of resources used were not reported. This will not enable the analysis to be easily reworked for other settings. Statistical and sensitivity analyses on the costs and quantities were carried out to assess the robustness of the estimates used. Since the costs were incurred during less than two years, discounting was not relevant. However, capital costs were appropriately amortised. The price year was reported, which will aid any future reflation exercises.
Other issues The authors did not compare their results with those of published studies, as this was the first economic evaluation of PSM of warfarin in a UK setting. The issue of generalisability of the results to other settings was not directly addressed. The authors do not appear to have presented their results selectively. The study enrolled patients with long-term indication of oral anticoagulation and this was reflected in the authors' conclusions. The authors reported several limitations to their study. For example, self-management was carried out under trial conditions, implying an artificial testing frequency. In addition, follow up of the patients was limited to one year.
Implications of the study The authors did not make explicit recommendations for changes in policy or practice. However, they called for further research to evaluate additional benefits of PSM, namely its impact in relieving pressure on traditional clinic-based care. In addition, future economic evaluations should focus on targeted sub-groups of anticoagulation patients. Lastly, the authors recommended that future research should try to evaluate patients in a real life environment.
Source of funding Supported by the Medical Research Council.
Bibliographic details Jowett S, Bryan S, Murray E, McCahon D, Raftery J, Richard Hobbs F D, Fitzmaurice D. Patient self-management of anticoagulation therapy: a trial-based cost-effectiveness analysis. British Journal of Haematology 2006; 134(6): 632-639 Other publications of related interest Fitzmaurice DA, Murray ET, McCahon D, Holder R, Hussein S, Sandhar H, et al. Self-management of anticoagulation: a randomised trial. BMJ 2005;331:1057.
Indexing Status Subject indexing assigned by NLM MeSH Aged; Anticoagulants /administration & Cost-Benefit Analysis; Drug Costs; Female; Humans; Male; Middle Aged; Patient Education as Topic /economics; Self Administration /economics /methods; Treatment Outcome; dosage /economics /therapeutic use AccessionNumber 22006001809 Date bibliographic record published 31/01/2007 Date abstract record published 31/01/2007 |
|
|
|