SEARCH

SEARCH BY CITATION

Keywords:

  • anticoagulation;
  • patient self-management;
  • cost;
  • cost-effectiveness

Summary

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Authors contribution
  7. Acknowledgements
  8. Competing interests
  9. Independence
  10. References
  11. Appendix

Demand for anticoagulation management is increasing due to an expansion of clinical indications for therapy. One possible model of care to meet demand is patient self-management (PSM), beneficial to patients who need control over their condition. This study aimed to determine the cost and cost-effectiveness of PSM of anticoagulation compared with routine clinic-based care for patients receiving long-term anticoagulation. A cost–utility analysis was conducted alongside a randomised controlled trial; 617 patients were recruited and followed up for 12 months. There was no significant difference in mean quality-adjusted life years (QALYs) between groups – after adjusting for baseline, the mean difference in QALYs was 0·009 (95% CI, −0·012 to 0·030). Overall mean healthcare costs in the PSM arm were significantly higher at £417 (CI £394–£442) compared with £122 (CI £103–£144) in the control arm. Therefore, using a formal cost-effectiveness analysis, PSM of anticoagulation does not appear to be cost-effective. However, PSM may have other benefits in relieving pressure on traditional clinic-based care, and the cost-effectiveness of this model of care for some subgroups of anticoagulation patients needs to be explored further.

The expansion of clinical indications for oral anticoagulation therapy (primarily warfarin in the UK) (Lowe, 1992; Sweeney et al, 1995), particularly non-rheumatic atrial fibrillation (Gustafson et al, 1992; Sandercock et al, 1992) has led to a big increase in patient numbers, which, in turn has increased the pressure on traditional warfarin clinics. Alternative models of care are needed (Taylor et al, 1993; Sudlow et al, 1995). The traditional model of care involves patients attending a hospital outpatient clinic for their international normalised ratio (INR) estimation, and the main alternatives are principally primary care-based clinics (Fitzmaurice et al, 1996), where blood samples are taken and sent to the laboratory or wholly at the practice using near-patient INR testing.

Near-patient testing for INR estimation provides the option of selected patients undertaking self-management (Rink et al, 1993; Hobbs, 1996) and reliable, portable machines are available. Patient self-management (PSM) requires the patient to conduct the INR test themselves and change their dose using a dosing algorithm. Previous data have demonstrated reliable performance of these machines when used by selected patients (Hasenkam et al, 1997). The move towards PSM has occurred due to patients demanding more autonomy and control over their condition (Kurnitz, 1998). Although convenience and patient independence are important, the relative clinical safety and effectiveness of PSM in comparison with existing models of care in the UK must first be established.

The cost-effectiveness of PSM needs further investigation. A US cohort study investigated cost (Anderson et al, 1993), where the cost of home monitoring was found to be approximately half that of routine care over an 8-week period. However, the cost analysis was not comprehensive. A German cohort study (Taborski et al, 1999) demonstrated a 50% reduction in costs associated with self-management compared with routine care, but routine care appeared to be associated with poor INR control and high rates of complications. From the UK perspective, comparison with usual care, a mix of hospital clinic attendance and primary care-based management, must both be considered.

In Germany and the USA, some insurers fund PSM once individual competence has been established, partly due to evidence of some improvement in therapeutic control (White et al, 1989; Korfer & Kortke, 2001). It is not clear whether this level of improvement is possible in the UK. If it is not, then the capital costs associated with self-management, increased frequency of testing and the cost of training would need to be offset against the savings due to reduced patient contact with health professionals.

This paper reports a cost-effectiveness analysis conducted alongside the first UK-based randomised controlled trial of non-selected patients in primary care, randomised to either self-management or usual practice (a mix of hospital and primary care clinics). The study aimed to determine the clinical and cost-effectiveness of PSM in comparison with standard care in patients on long-term oral anticoagulation therapy.

Methods

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Authors contribution
  7. Acknowledgements
  8. Competing interests
  9. Independence
  10. References
  11. Appendix

Overview

A trial-based cost-effectiveness analysis was conducted with a 1-year time horizon. Costs were estimated from a National Health Service (NHS) perspective and a wider societal perspective, taking into account NHS costs and costs borne by the patients themselves. Clinical benefits were expressed in terms of INR control (time in range). The outcome for the cost-effectiveness analysis was quality-adjusted life years (QALYs), and so a cost–utility analysis was undertaken. QALYs take into account the survival and quality of life of an individual – the focus here was on the potential for quality of life gains from self-management.

Trial design

Full details of the trial have been reported (Fitzmaurice et al, 2005). The trial recruited 617 patients with a long-term indication for oral anticoagulation, randomised to the intervention arm (PSM) or usual care. Intervention arm patients were provided with training and, if capable of undertaking PSM, managed their own anticoagulation for a period of 12 months. PSM patients were asked to perform their INR test once every 2 weeks unless the dosage changed, when they were asked to perform the test after 1 week. They attended a practice-based study clinic every 3 months for assessment of their progress. If not considered capable of self-management, they returned to routine care. Patients could also withdraw from the study if they were no longer happy self-managing or if they suffered a serious adverse event. If a patient died, discontinued warfarin or moved away from the area, they completed the study early and no further data were collected. Control patients continued with their pre-study care, either attending hospital or practice-based anticoagulant clinics. This study was approved by the West Midlands research ethics committee and all relevant local research committees

Health outcomes

Clinical events.  Baseline INR data were collected on all patients at study entry. During the study period, INR data were collected using a case report form. The INR data were then analysed to calculate pre-study and study time in range.

Quality of life.  Postal questionnaires were sent to study participants at baseline, 6 and 12 months. The questionnaire contained the EQ-5D (EuroQol Group, 1990), a tool that measures and values broad aspects of quality of life (Kind et al, 1998). The EQ-5D contains five dimensions: mobility, self-care, usual activities, pain/discomfort and anxiety/depression. Utility values derived from a UK general population survey were used (Dolan et al, 1995). The QALY score for each study patient over the 12 months was estimated by calculating the area under each patient's health utility curve using linear interpolation.

Due to the problem of missing EQ-5D data at one or more of the time points for a proportion of patients, multiple imputation was used (Schafer, 1999). Multiple imputation is a simulation based-technique that estimates missing data points by using information available within the data set to estimate the missing value. In order to avoid bias, adjustment for differences in baseline EQ-5D scores was undertaken (Manca et al, 2005) using a regression-based adjustment. All QALY scores reported were adjusted for baseline differences. Both the complete case and imputed analyses are presented.

Cost analysis

Overview.  The cost analysis adopted a societal perspective to include costs incurred within the health sector and by patients and carers. Data collection was undertaken on all trial patients and a stochastic cost analysis was conducted. PSM patient costs included the machine, consumables used, provision of training and assessment and telephone contacts for advice specific to PSM (Table I). Control patient costs included anticoagulation clinic attendances in primary or secondary care, taking into account staff, equipment, consumables and overheads. Anticoagulation clinic costs per patient visit were collected from a sample of NHS providers in both primary and secondary care. Data were also collected on anticoagulation clinic attendances for PSM patients who reverted back to routine care. For all trial patients, all primary and secondary care (inpatient and outpatient) contacts relating to minor and serious adverse events were recorded.

Table I.   Anticoagulation unit costs.
VariableCost (£)Source
  1. BNF: British National Formulary London, Pharmaceutical Press (http://www.bnf.org/bnf/); Roche: Roche Diagnostics Limited, Lewes, East Sussex, UK; Whitely Council: Whitley Council, Department of Health, London; GP, general practitioner; MLSO, Medical Laboratory Scientific Officer.

  2. *Excluding professional training costs.

Patient self-management
 Machine468·83Roche
 Test strip2·50Roche
 Internal quality control test5·00Roche
 Lancets (each)0·03BNF
 Sharps bin1·05NHS diagnostics
 Tissues (box of 150)0·46NHS diagnostics
 Alcowipes (each)0·02NHS diagnostics
 Laminated dosing card1·00Study data
 Practice nurse phone call (10 min)4·83Netten and Curtis (2003)
 Practice nurse salary (per clinic hour)*23·00Netten and Curtis (2003)
 Practice nurse salary (per clinic hour)29·00Netten and Curtis (2003)
 GP salary (per clinic hour)*98·00Netten and Curtis (2003)
 GP salary (per clinic hour)116·00Netten and Curtis (2003)
 Specialist registrar40·00Netten and Curtis (2003)
 Consultant109·00Netten and Curtis (2003)
 Administration10·93Whitley Council
 Clinic overheads (per hour)13·16Study data
 Training session 1 (per person)42·52Study data
 Practice test strips/lancets25·30Study data
 Training session 2 (per person)28·67Study data
 Assessment clinic (per session)15·73Study data
Routine care (cost per visit)
 Hospital clinic6·75Study data
 GP blood sample, hospital analysis and dosing9·38Study data
 GP blood sample and dosing, hospital analysis10·69Study data
 Practice near-patient testing clinic14·16Study data
 Pharmacist-led practice clinic17·66Study data
 MLSO-led practice clinic11·62Study data

Self-management costs.  Patients attended either two or three training sessions depending on the level of need for further training. All main resources used in training were recorded along with the number of patients per training session. Machine costs were estimated for those patients trained to carry out PSM, even if they did not continue with PSM. Therefore, those randomised to PSM but who did not attend the training were not attributed a machine cost. The cost of the machine was amortised over 3 years. The assumption was made that if a machine was not used for a full year it could be used by another patient. If a trained patient undertook <12 months self-management, the cost of the machine was taken to be proportional to the time spent using it. Data on use of consumables (e.g. test strips, lancets, etc.) were also collected. The number of phone calls per patient concerning INR queries, medical queries and problems with testing equipment was recorded and included in the cost of self-management and it was assumed that calls were to a practice nurse specialising in anticoagulation.

Assessment sessions were assumed to be 15-min long and carried out by a practice nurse. For patients who did not carry out 12 months of self-management, the number of assessment sessions attended was recorded and costs attributed to these. The number of internal quality control tests per patient was also recorded. For any training session or clinic requiring the use of a room in general practitioner surgery, an overhead cost per hour was also taken into consideration. If a patient withdrew from PSM and returned to their previous anticoagulation clinic, data on the total number of clinic visits was recorded.

Patients’ private costs.  A random subsample of 209 study patients (33% of all patients), from both the PSM and control groups were asked to complete a patient cost questionnaire for their most recent clinic visit. For logistical reasons it was not possible to collect data from all study patients. In the control arm, the sample was stratified to contain patients who attended primary care and secondary care anticoagulation clinics. The purpose of the questionnaire was to collect information on private travel and time costs. Data were collected on the time taken to travel to the clinic, the mode and cost of transport where applicable, time spent waiting before the appointment and if the patient was accompanied. Information was also collected on activities foregone in order to attend the clinic, including leisure activities, care of relatives and employment. Using the data from the subsample, an average patient cost was calculated for a primary care clinic visit, a secondary care clinic visit and a PSM practice visit (for training or assessment). Total private costs for a control patient were calculated from the number of clinic visits and the average private cost by clinic type. Total PSM private costs took into account training and assessment sessions attended and routine clinic visits, if applicable.

Unit costs.  All unit costs were valued at 2003 prices in UK £ Sterling. Staff costs were calculated on the basis of salary scale mid-points, plus any additional employer costs where applicable (Netten & Curtis, 2003). Capital (equipment) costs were based on purchase prices and amortised over a 3- or 5-year period (where appropriate) using a 3·5% discount rate (HM Treasury, 2003). Straight line depreciation, and no residual value, was assumed. Unit costs of routine care and PSM are given in Table I. Unit costs of contacts with primary and secondary care, including medical procedures for serious adverse events were obtained from published sources (Netten & Curtis, 2003; NHS Reference Costs 2003 and National Tariff 2004, 2004) (see Appendix).

In the calculation of patient out-of-pocket expenses, where travel was by private car, mileage was obtained using the relevant postcodes and costs were calculated using published Automobile Association motoring costs (http://www.theaa.com/motoring_advice/running_costs/index.html). The value attached to patient time was dependent on the activity foregone – time lost from work was valued as the mean gross weekly wage and for patients not in paid employment, activity foregone was classified as ‘leisure time’ and valued at 40% of the mean average wage. This was based on previous estimates reported in the literature (Henderson et al, 2002).

Cost-effectiveness analysis

An incremental cost-effectiveness analysis was conducted (Drummond et al, 1997) based on intention-to-treat. The aim of the incremental cost-effectiveness analysis was to determine the difference in costs and the difference in outcomes between standard care and self-management. Dividing the difference in costs by the difference in outcomes gives the incremental cost-effectiveness ratio (ICER), which shows the additional cost required to gain one additional unit of benefit (e.g. cost per QALY gained).

The cost data were skewed and so the arithmetic mean was used along with its non-parametric 95% CI (Barber & Thompson, 1998). Significance tests of differences between mean values were carried out using the independent sample t-test. To account for uncertainty due to sampling variation in cost-effectiveness, cost-effectiveness acceptability curves (CEACs) were plotted (van Hout et al, 1994). CEACs show the probability of PSM being more cost-effective than routine anticoagulation clinics for a range of cost per QALY threshold values.

Sensitivity analyses were undertaken to explore uncertainties in the analysis. A complete case analysis was undertaken using patient data, where both cost and quality life information were available, in order to determine the impact of using multiple imputation. Further sensitivity analysis explored the robustness of the results to:

  • 1
    an extended machine lifetime of 5 or 10 years;
  • 2
    the use of alternative health care professionals for training; and
  • 3
    the exclusion of training costs from the analysis.

The Statistical Package for the Social Science for Windows, version 13.0.1 (SPSS Inc., Chicago, IL, USA), was used to perform the main statistical analysis, bootstrapping was carried out using stata version 8.2 (StataCorp LP, College Station, TX, USA) and Microsoft excel 2002. SP3 was used to construct the CEACs. Multiple imputation was performed using norm version 2.03 (Schafer, 2000). P-values < 0·05 were considered significant.

Results

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Authors contribution
  7. Acknowledgements
  8. Competing interests
  9. Independence
  10. References
  11. Appendix

A total of 617 patients were entered into the study (PSM 337 versus control 280) – 400 (65%) of randomised patients were male and the mean ages were 64 years in the PSM arm and 66 years in the control arm. Figure 1 shows the flow of patients through the study, indicating patients for whom complete resource use data was not available. Exclusion of patients was due to drop out before data collection commenced, loss to follow up or discontinuation of warfarin where subsequent treatment was unknown. Complete NHS resource-use data were available for 326 PSM patients and 265 control patients.

image

Figure 1.  Flowchart of study.

Download figure to PowerPoint

Health outcomes and quality of life

There were no statistically significant differences in INR percentage time in range between the PSM and control arms (Fitzmaurice et al, 2005). INR control, based on percentage time in range within the study, was near equivalent between intervention and control groups (PSM 70%versus control 68%). The observed difference in mean total QALYs between groups was in favour of PSM but the difference was very small and not statistically significant, for both the complete case and imputation-based analyses (Table II). This finding is consistent with the lack of difference in the more clinically oriented measures. The proportions of patients in each trial arm reporting some or extreme problems for each EQ-5D dimension at baseline, 6 and 12 months are presented in Table III. Concentrating on the anxiety/depression dimension of the EQ-5D, no statistically significant difference was observed between trial arms in the change from baseline to 6 or 12 months.

Table II.   Quality-adjusted life years over 12 months by treatment group.
 MeanDifference in mean values (PSM − control)Bootstrapped 95% CI
Control arm (n = 265)PSM arm (n = 326)
  1. PSM, patient self-management.

  2. *Control arm n = 165, PSM arm n = 204.

Complete case*0·7380·7390·001−0·027 to 0·032
Imputed0·7120·7210·009−0·012 to 0·030
Table III.   EQ-5D dimensions by trial arm over time.
DimensionPSM arm (% with problems in that dimension)*Control arm (% with problems in that dimension)*
Baseline (n = 204)6 months (n = 200)12 months (n = 198)Baseline (n = 165)6 months (n = 162)12 months (n = 162)
SomeExtremeSomeExtremeSomeExtremeSomeExtremeSomeExtremeSomeExtreme
  1. PSM, patient self-management.

  2. *Reduction in sample as a result of patient deaths.

Mobility41·7044·0047·0049·7054·3056·80
Self-care10·3010·509·6012·70·616·0020·40
Usual activities30·45·441·54·536·93·037·03·642·64·347·51·9
Pain/discomfort45·15·445·06·543·96·652·16·151·96·251·96·2
Anxiety/depression20·11·026·52·024·21·521·81·224·12·526·52·5

Cost-effectiveness analysis

The study results are summarised in Tables II and IV. At 12 months, mean PSM arm anticoagulation costs (£382) were significantly higher (P < 0·001) than the mean control arm costs (£90), and additional (i.e. non-anticoagulation) NHS costs for each arm were similar. Overall, total NHS costs in the PSM arm were £417 and in the control arm were £122 (P < 0·001). Control arm costs were skewed with a few patients having large costs due to serious adverse events. PSM costs were bi-modal as some patients dropped out of PSM and returned to their usual care and so incurred a lower cost.

Table IV.   Costs over 12 months by treatment group.
 Mean (bootstrapped 95% CI)
Control arm (n = 265)PSM arm (n = 326)
  1. PSM, patient self-management; NHS, National Health Service.

Anticoagulation costs (£)89·89 (83·15–97·32)381·53 (365·63–397·90)
Additional NHS costs (£)32·43 (14·65–55·05)35·23 (17·35–58·88)
Total NHS cost (£)122·32 (103·48–143·90)416·76 (393·95–441·81)
Patient costs (£)57·48 (53·65–61·49)45·97 (42·98–49·31)
Overall societal costs (£)179·80 (160·09–202·58)462·73 (439·28–489·15)

In total, 81% of the patient cost questionnaires were returned. The cost to the patient was £8 for a hospital clinic visit, £4 for a control patient practice visit and £4 for a PSM patient practice visit. The mean unit cost for attending the training sessions was £11 for session 1 and £6 for session 2. Taking into account all visits, training sessions and assessments, mean costs to the patient were significantly higher (P < 0·001) in the control arm (£57) than in the PSM arm (£46). Thus, the overall mean societal costs were significantly higher (P < 0·001), and considerably larger, in the PSM arm (£463) than the control arm (£180).

From a healthcare perspective and using the imputation-based analysis results, the ICER for PSM was £32 716 per QALY gained. Using societal costs, the ICER for PSM was £31 437 per QALY gained. Whilst these are based on a QALY difference that is not statistically significant, cost-minimisation analysis is not appropriate as the trial was not conducted to test a hypothesis of equivalence (Briggs & O'Brien, 2001).

The CEACs are shown in Fig 2. The CEACs showed the probability that PSM was more cost-effective than routine anticoagulation clinics for a range of CE threshold maximum values. At a ceiling of £20 000 per QALY gained, the probability that PSM was cost-effective using the imputed data set was 30%, and 16% for the complete case analysis. At £30 000 per QALY the probability rose to 46% for the imputed data set and 26% for the complete case analysis. The inclusion of patient costs (using the imputed data set) increased the probabilities to 32% and 49% at £20 000 per QALY and £30 000 per QALY respectively.

image

Figure 2.  Cost-effectiveness acceptability curve for patient self-management of anticoagulation compared with routine care.

Download figure to PowerPoint

Sensitivity analysis

Results of the sensitivity analysis are given in Table V. The mean PSM anticoagulation and mean total NHS costs were still significantly higher than routine care even when changing the lifetime of the near-patient testing machine to 5 or 10 years. The exclusion of training costs also did not have a marked affect on the results.

Table V.   PSM sensitivity analysis.
AnalysisMean (£) (bootstrapped 95% CI)
  1. PSM, patient self-management; NHS, National Health Service.

Base case anticoagulation cost381·53 (365·63–397·90)
Anticoagulation cost with machine over 5 years338·21 (324·59–351·80)
Anticoagulation cost with machine over 10 years305·83 (294·42–317·65)
Anticoagulation cost without training291·69 (276·04–306·10)
Base case total NHS cost416·76 (393·95–441·81)
Total NHS cost with machine over 5 years373·44 (350·70–398·67)
Total NHS cost with machine over 10 years341·06 (320·11–366·06)
Total NHS cost without training326·92 (304·07–353·69)

Discussion

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Authors contribution
  7. Acknowledgements
  8. Competing interests
  9. Independence
  10. References
  11. Appendix

Patient self-management was more expensive than standard care, with NHS costs around £420 per patient per year, compared with approximately £120 for routinely managed patients. The CEAC demonstrated that, at a CE threshold of £20 000 per QALY, PSM had a probability of only 30% of being cost-effective, and 46% at £30 000 per QALY. Therefore, in general terms, and using the cost-effectiveness criteria applied by the National Institute for Health and Clinical Excellence in the UK (i.e. £30 000 per QALY) (Rawlins & Culyer, 2004), it seems unlikely that this technology represents good value for money within the UK healthcare system.

However, the findings of this study need to be balanced against the capacity constraint issue that currently exists within clinic-based anticoagulation delivery systems. Whilst it is more expensive, the PSM approach would, in theory, remove many patients (possibly 25% of patients) from outpatient clinics, potentially freeing up clinician time to manage other, possibly more complicated patients. It should also be noted that, within the trial, self-management patients tested at least once every 2 weeks, whereas in routine care testing was less frequent, except where dose changes were required. Partly because of the cost of test strips, PSM will always be more expensive if such frequent testing is undertaken. However, the cost of PSM could be reduced if self-management testing is more in line with traditional care.

The increased convenience that self-management brought to patients, reflected in part through lower private costs, might also be remembered. The patient cost questionnaire revealed differences in level of convenience between the groups, shown in waiting times, travel time and distance to clinic. Patient costs were slightly lower in the PSM arm as no routine clinic visits were required, but the costs of training and assessment clinics were included. The costs in subsequent years for PSM patients will be lower than that reported in this paper as there will be no training costs and probably fewer assessment visits per year. However, the lower private costs in future years are unlikely to have a big impact on the overall results as the patient cost component is small.

The ICER reported in this study was relatively high, which was partly due to the very small difference, if any, in QALYs between groups. This raises the issue of the appropriateness of using the EQ-5D questionnaire in an intervention of this type. However, this instrument is now widely used in cost–utility analyses, as a generic measure of health-related quality of life, allowing broad efficiency comparisons across disease areas and across different interventions. Given that self-management programmes are competing for limited ‘healthcare’ resources then, in our opinion, it is right and appropriate to judge its benefits in relation to a broad health measure. Although PSM of warfarin may not bring about a large increase in QALYs, there may be other less tangible benefits, such as increased independence and patient empowerment that are not captured by the QALY measure. In addition, it may be that PSM is more cost-effective for particular subgroups of patients, such as those with poor INR control who are more problematic to manage and are at risk of adverse events. However, such subgroup analysis was not part of this study and so further research is required to investigate this.

This is the first UK cost-effectiveness analysis of PSM of warfarin. A major strength of this study is that it was conducted alongside a large randomised controlled trial. In addition, the study achieved a high level of data completeness on the use of health care resources – this was missing for only a small number of patients. However, there are some limitations to this work. The self-management was undertaken within trial conditions and so an artificial testing frequency was imposed on patients and only a 1-year follow up was possible. Therefore, it will be important for future work to follow-up patients outside trial conditions in order to observe the actual frequency of testing. In addition, there were missing EQ-5D data at one or more time points for a number of study patients and so multiple imputation was used to address this problem. Both the complete case and imputed data analyses were reported here – the results when imputation was used were marginally more favourable to PSM compared with those when only complete case data were analysed.

In conclusion, using a cost-effectiveness threshold of £30 000 per QALY, PSM was not highly cost-effective. However, it may be that, with careful selection and targeting of patients, self-management is an attractive approach for a significant minority of patients receiving oral anticoagulation therapy. Giving such patients the option of self-management would also be in line with the current UK Government's health strategy where patients are given more choice about how, when and where they receive treatment.

Authors contribution

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Authors contribution
  7. Acknowledgements
  8. Competing interests
  9. Independence
  10. References
  11. Appendix

SJ undertook the economic evaluation, wrote the first draft of the article and is the guarantor. SB designed and supervised the economic analysis. EM was the project manager of the trial. DM contributed to the data collection and data management. JR provided health economic input. FDRH contributed to the study design. DF was the principal investigator. All authors contributed to the drafting of the final manuscript.

Acknowledgements

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Authors contribution
  7. Acknowledgements
  8. Competing interests
  9. Independence
  10. References
  11. Appendix

The authors gratefully acknowledge the Medical Research Council for funding this study (grant no. G9900263). DF was funded by an NHS career scientist award, EM was funded by an MRC health services research fellowship.

References

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Authors contribution
  7. Acknowledgements
  8. Competing interests
  9. Independence
  10. References
  11. Appendix
  • Anderson, D.R., Harrison, L. & Hirsh, J. (1993) Evaluation of a portable prothrombin time monitor for home use by patients who require long-term oral anticoagulant therapy. Archives of Internal Medicine, 153, 14411447.
  • Barber, J.A. & Thompson, S.G. (1998) Analysis and interpretation of cost data in randomised controlled trials: review of published studies. British Medical Journal, 317, 11951200.
  • Briggs, A.H. & O'Brien, B.J. (2001) The death of cost-minimization analysis? Health Economics, 10, 179184.
  • Dolan, P., Gudex, C., Kind, P. & Williams, A. (1995) A Social Tariff for Euroqol: Results from a UK General Population Survey. CHE Discussion Paper 138, Centre for Health Economics, University of York. Available at http://www.york.ac.uk/inst/che/pdf/DP138.pdf [last accessed 21 July 2005].
  • Drummond, M.F., O'Brien, B., Stoddart, G.L. & Torrance, G.W. (1997) Methods for the Economic Evaluation of Health Care Programmes, 2nd edn. Oxford University Press, Oxford.
  • EuroQol Group (1990) EuroQol: a new facility for the measurement of health-related quality of life. Health Policy, 16, 199208.
  • Fitzmaurice, D.A., Hobbs, F.D.R. & Murray, J.A. (1996) Monitoring oral anticoagulation in primary care. British Medical Journal, 312, 14311432.
  • Fitzmaurice, D.A., Murray, E.T., McCahon, D., Holder, R., Hussein, S., Sandhar, H., Hobbs, F.D.R. & Raftery, J.P. (2005) Self-management of anticoagulation: a randomised trial. British Medical Journal, 331, 1057.
  • Gustafson, C., Asplund, K., Britton, M., Norrving, B., Olsson, B. & Marke, L.A. (1992) Cost effectiveness of primary prevention in atrial fibrillation: Swedish national perspective. British Medical Journal, 305, 14571460.
  • Hasenkam, J.M., Knudsen, L., Kimose, H.H., Gronnesby, H., Atterman, J., Anderson, N.T. & Pilegaard, H.K. (1997) Practicability of patient self testing of oral anticoagulant therapy by the international normalized ratio (INR) using a portable whole blood monitor. A pilot investigation. Thrombosis Research, 85, 7782.
  • Henderson, J., Bricker, L., Roberts, T., Mugford, M., Garcia, J. & Neilson, J. (2002) British National Health Service's and women's costs of antenatal ultrasound screening and follow-up tests. Ultrasound in Obstetrics and Gynecology, 20, 154162.
  • HM Treasury (2003) The Green Book. Appraisal and Evaluation in Central Government. The Stationery Office, London.
  • Hobbs, F.D.R. (1996) Near patient testing in primary care. British Medical Journal, 312, 263264.
  • Van Hout, B.A., Al, M.J., Gordon, G.S. & Rutten, F.F.H. (1994) Costs, effects and c/e-ratios alongside a clinical trial. Health Economics, 3, 309319.
  • Kind, P., Dolan, P., Gudex, C. & Williams, A. (1998) Variations in population health status: results from a United Kingdom national questionnaire survey. British Medical Journal, 316, 736741.
  • Korfer, R. & Kortke, H. (2001) ESCAT Early Self Controlled Anticoagulation Trial. Annals of Thoracic Surgery, 72, 4448.
  • Kurnitz, J. (1998) The view from the inside: personal reflections. Journal of Thrombosis and Thrombolysis, 5, S49S51.
  • Lowe, G.D.O. (1992) Anti-thrombotic treatment and atrial fibrillation. Editorial. British Medical Journal, 305, 14451446.
  • Manca, A., Hawkins, N. & Sculpher, M.J. (2005) Estimating mean QALYs in trial-based cost-effectiveness: the importance of controlling for baseline utility. Health Economics, 14, 487496.
  • Netten, A. & Curtis, L. (2003) Unit Costs of Health and Social Care 2003. Personal Social Services Research Unit, University of Kent, Canterbury.
  • NHS Reference Costs 2003 and National Tariff 2004 (2004) Payment by Results Core Tools 2004. Department of Health, London.
  • Rawlins, M.D. & Culyer, A.J. (2004) National Institute for Clinical Excellence and its value judgments. British Medical Journal, 329, 224227.
  • Rink, E., Hilton, S., Szczepura, A., Fletcher, J., Sibbald, J., Sibbald, B., Davies, C., Freeling, P. & Stilwell, J. (1993) Impact of introducing near patient testing for standard investigations in general practice. British Medical Journal, 307, 775778.
  • Sandercock, P., Bamford, J., Dennis, M., Burn, J., Slattery, J. & Jones, L. (1992) Atrial fibrillation and stroke: prevalence in different types of stroke and influence on early and long term prognosis. Oxfordshire Community Stroke Project British Medical Journal, 305, 14601465.
  • Schafer, J.L. (1999) Multiple imputation: a primer. Statistical Methods in Medical Research, 8, 315.
  • Schafer, J.L. (2000) NORM: multiple imputation of incomplete multivariate data under a normal model. (2003). Pennsylvania, Pennsylvania State University, USA.
  • Sudlow, C.M., Rodgers, H., Kenny, R.A. & Thomson, R.G. (1995) Service provision and use of anticoagulants in atrial fibrillation. British Medical Journal, 311, 558561.
  • Sweeney, K.G., Pereira-Gray, D., Steele, R. & Evans, P. (1995) Use of warfarin in non-rheumatic atrial fibrillation: a commentary from general practice. British Journal of General Practice, 45, 153158.
  • Taborski, U., Wittstamm, F.J. & Bernada, A. (1999) Cost-effectiveness of self-management anticoagulant therapy in Germany. Seminars in Thrombosis and Haemostasis, 25, 10107.
  • Taylor, F., Ramsey, M., Voke, J. & Cohen, H. (1993) GPs not prepared for monitoring anticoagulation. British Medical Journal, 307, 1493.
  • White, R., McCurdy, S.A., Von Marensdorff, H., Woodruff, D.E. & Leftgoff, P.A. (1989) Home prothrombin time monitoring after the initiation of Warfarin therapy. Annals of Internal Medicine, 111, 730736.

Appendix

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Authors contribution
  7. Acknowledgements
  8. Competing interests
  9. Independence
  10. References
  11. Appendix
Table A1.   National Health Service (NHS) resource-use unit costs.
Adverse event/procedure/visitMean cost (£) (range 50% trusts)
  1. ITU/ICU, intensive therapy unit/intensive care unit; A&E, accident and emergency; GP, General Practitioner.

  2. *NHS Reference Costs (2003), non-elective unless stated, NHS Trusts.

Thrombotic events*
 Pulmonary embolism1132 (932–1428)
 Thrombotic stroke (9 days hospital stay)1647 (1279–2019)
 Minor thrombotic stroke594 (416–770)
 Transient ischaemic attack, day case384 (309–480)
 Thrombectomy2085 (1235–2203)
Haemorrhagic events*
 Gastrointestinal bleed with major procedures3498 (2102–4100)
 Gastrointestinal bleed with endoscopy1126 (786–1498)
 Gastrointestinal bleed >69 years old or with complications and comorbidities918 (798–1221)
 Gastrointestinal bleed <70 years old without complications and comorbidities569 (426–757)
 Epistaxis day case521 (416–657)
 Epistaxis non-elective inpatient833 (609–1082)
 Rectal bleed597 (398–898)
 ITU/ICU per bed day1265 (1105–1478)
Secondary care inpatient*
 Dilation and curettage, gynaecology, elective inpatient738 (473–955)
 Flexible cystoscopy, elective662 (529–898)
Secondary care outpatient*
 A&E (high-cost imaging, referred/discharged)118 (103–190)
 A&E (lower cost investigation, admitted)112 (93–136)
 A&E (lower cost investigation, referred/discharged)65 (54–81)
 Endoscopy164 (116–201)
 X-ray14·97 (11·00–18·14)
 Arthroscopy767 (607–911)
 Surgical gastroenterology (with investigation)87 (74–105)
 Medical gastroenterology (with investigation)96 (76–127)
 Medical gastroenterology (without investigation)75 (51–97)
 Ophthalmology59 (44–69)
 Flexi sigmoidoscopy401 (332–516)
 Flexible cystocopy351 (291–460)
 Vascular surgery61 (44–72)
 Cardiology68 (48–79)
 Gynaecology68 (48–79)
 Cautery74 (59–90)
Primary care†
 GP consultation15
 GP home visit17
 GP telephone consultation8
 Practice Nurse consultation47
Table A2.   Resource use.
 Mean cost per patient per year (£)
Control arm (n = 265)PSM arm (n = 326)
  1. PSM, patient self-management; NHS, National Health Service.

Anticoagulation costs
PSM costs
 Consumables0·0077·82
 Machine0·00114·16
 Internal quality control0·0017·15
 Telephone calls0·008·79
 Training0·0089·84
 Assessment0·0040·92
Routine anticoagulant clinic visits89·8932·87
Additional NHS costs
 Primary care2·823·18
 Secondary care29·6232·05
Table A3.   Unit patient private costs (£)*.
 PSM assessment visit (n = 50) Practice (n = 55) Hospital (n = 50)
  1. PSM, patient self-management.

  2. *Costs for two domiciliary patients £0·50 per visit.

Mean (95% CI)4·08 (3·26–4·89)4·18 (3·33–5·03)8·12 (6·82–9·43)
Median3·223·276·52
Range1·06–13·710·96–14·862·77–25·22
Missing534