Comments regarding cost-effectiveness of POC monitoring in general practice
Leanne Chalmers, Unit for Medication Outcomes Research and Education (UMORE), School of Pharmacy, University of Tasmania
22 July 2010
To the Editor,
Having conducted several projects utilising point-of-care (POC) INR monitors in a variety of settings over recent years (1-5)*, we read with interest Caroline Laurence’s paper. With international studies having yielded conflicting results regarding the economic feasibility of POC testing, we had hoped that this study would offer some clarity in an Australian context. Unfortunately, given the acknowledged limitations and the fact that the data are now over three years old, we are hesitant to place too much emphasis on its findings.
We have three major concerns regarding the findings of this study. The first relates to the health outcome indicator chosen for the cost-effective analysis. The proportion of patients within the therapeutic range for each condition at the end of the Trial was not only a surrogate outcome marker, as acknowledged by the authors, but was also only a point estimate. The clinically meaningfulness of this indicator is highly questionable, especially in the setting of progressive diseases such as diabetes and dyslipidaemia, where a worsening of control over time is not unusual. Furthermore, it is recognised that the wide range of pharmacodynamic and pharmacokinetic factors that influence the INR render a single result poorly indicative of a patient’s overall INR control. Why was the time in therapeutic range (TTR), a standard measure used in POC trials involving warfarin (6), not used? The TTR is strongly correlated with the clinical outcomes of warfarin therapy and is therefore more appropriate for an analysis of cost-effectiveness. Alternatively, why was health-related quality of life not investigated as the health outcome indicator to allow comparability to other economic studies?
Our second concern relates to the assumption that the method of monitoring directly influences patient outcome, which we cannot accept is necessarily the case. Optimal management of a condition requires not only appropriate monitoring, but appropriate communication of, and action in response to, the result on the part of the doctor and the patient (e.g. in adherence to changes to their medication regimens and lifestyle modifications). In practice, communication issues are particularly relevant to patient safety. No consideration of these factors appears to have occurred within the cost-effectiveness analysis.
Our final concern relates to the assumed association between the monitoring of the patients’ chronic conditions and the costs associated with their health care. A number of patients involved had several of the studied medical conditions and were thus included in more than one arm of the Trial. Was it possible to attribute their health-related costs, especially hospitalisation costs, to a particular disease state, and thus to the efficacy of monitoring of that condition? It appears that hospitalisation costs were extensively modelled based only data from only a sample of patients, so we would suggest not. For example, if a diabetic patient taking warfarin for atrial fibrillation (and thus in both the HbA1c and INR arms of the Trial) was admitted to hospital for an elective total hip replacement, were these costs factored into the costs of their diabetic-related health care, their warfarin-related health care, or both? Was it appropriate to attribute them to either? Given that in the presented comparison of costs of HbA1c testing (as an example), hospitalisation costs represented the single biggest difference in costs between the groups (albeit using the modelled data with a very large confidence interval), any variations in these particular costs could have significantly influenced the results.
Given these limitations and the acknowledged failure of the study to recruit adequate numbers of patients for the desired statistical power, we agree with the authors that further consideration of the range of issues around POC monitoring is required before a definite decision can be made regarding the feasibility of its national implementation. At the very least, a more robust economic analysis utilising a less controversial health outcome indicator would assist in these future discussions.
Miss Leanne Stafford Dr Luke Bereznicki Professor Gregory Peterson
Unit for Medication Outcomes Research and Education (UMORE), School of Pharmacy, University of Tasmania
References 1. Jackson SL, Bereznicki LR, Peterson GM, Marsden KA, Jupe DM, Vial JH, et al. Accuracy and clinical usefulness of the near-patient testing CoaguChek S international normalised ratio monitor in rural medical practice. Aust J Rural Health 2004;12(4):137-42. 2. Jackson SL, Peterson GM, House M, Bartlett T. Point-of-care monitoring of anticoagulant therapy by rural community pharmacists: description of successful outcomes. Aust J Rural Health 2004;12(5):197-200. 3. Jeffrey E, Bereznicki L, Stafford L, Peterson G. Evaluation of a clinical pathway to enable patient self-monitoring of anticoagulation. In: HAA Handbook and Final Programme, 2009 19-21 October; Adelaide: p. 276. 4. Peterson G, Stafford L, Bereznicki L, van Tienen E, Jackson S, on behalf of the Project Team. The role of community pharmacy in post hospital management of patients initiated on warfarin. In: Pharmacy Practice Research Summit, 2010 2-4 March; Canberra. 5. Peterson G, Jackson S, Gee P, Bereznicki L, Bereznicki B, Fitzmaurice K, et al. Improving the management of warfarin in aged care facilities utilising an innovative IT solution. In: Australasian Pharmaceutical Science Association Conference, 2007 8-11 December; Manly. 6. Samsa GP, Matchar DB. Relationship between test frequency and outcomes of anticoagulation: A literature review and commentary with implications for the design of randomized trials of patient self-management. J Thromb Thrombolysis 2000;9(3):283-92.
* Several of these programs were funded by the Australian Government Department of Health and Ageing as part of the Fourth Community Pharmacy Agreement through the Fourth Community Pharmacy Agreement Grants Program managed by the Pharmacy Guild of Australia.
Competing interests
UMORE has received support from Roche Diagnostics Australia for warfarin-related research, principally in the supply of POC INR monitors.
Comments regarding cost-effectiveness of POC monitoring in general practice
22 July 2010
To the Editor,
Having conducted several projects utilising point-of-care (POC) INR monitors in a variety of settings over recent years (1-5)*, we read with interest Caroline Laurence’s paper. With international studies having yielded conflicting results regarding the economic feasibility of POC testing, we had hoped that this study would offer some clarity in an Australian context. Unfortunately, given the acknowledged limitations and the fact that the data are now over three years old, we are hesitant to place too much emphasis on its findings.
We have three major concerns regarding the findings of this study. The first relates to the health outcome indicator chosen for the cost-effective analysis. The proportion of patients within the therapeutic range for each condition at the end of the Trial was not only a surrogate outcome marker, as acknowledged by the authors, but was also only a point estimate. The clinically meaningfulness of this indicator is highly questionable, especially in the setting of progressive diseases such as diabetes and dyslipidaemia, where a worsening of control over time is not unusual. Furthermore, it is recognised that the wide range of pharmacodynamic and pharmacokinetic factors that influence the INR render a single result poorly indicative of a patient’s overall INR control. Why was the time in therapeutic range (TTR), a standard measure used in POC trials involving warfarin (6), not used? The TTR is strongly correlated with the clinical outcomes of warfarin therapy and is therefore more appropriate for an analysis of cost-effectiveness. Alternatively, why was health-related quality of life not investigated as the health outcome indicator to allow comparability to other economic studies?
Our second concern relates to the assumption that the method of monitoring directly influences patient outcome, which we cannot accept is necessarily the case. Optimal management of a condition requires not only appropriate monitoring, but appropriate communication of, and action in response to, the result on the part of the doctor and the patient (e.g. in adherence to changes to their medication regimens and lifestyle modifications). In practice, communication issues are particularly relevant to patient safety. No consideration of these factors appears to have occurred within the cost-effectiveness analysis.
Our final concern relates to the assumed association between the monitoring of the patients’ chronic conditions and the costs associated with their health care. A number of patients involved had several of the studied medical conditions and were thus included in more than one arm of the Trial. Was it possible to attribute their health-related costs, especially hospitalisation costs, to a particular disease state, and thus to the efficacy of monitoring of that condition? It appears that hospitalisation costs were extensively modelled based only data from only a sample of patients, so we would suggest not. For example, if a diabetic patient taking warfarin for atrial fibrillation (and thus in both the HbA1c and INR arms of the Trial) was admitted to hospital for an elective total hip replacement, were these costs factored into the costs of their diabetic-related health care, their warfarin-related health care, or both? Was it appropriate to attribute them to either? Given that in the presented comparison of costs of HbA1c testing (as an example), hospitalisation costs represented the single biggest difference in costs between the groups (albeit using the modelled data with a very large confidence interval), any variations in these particular costs could have significantly influenced the results.
Given these limitations and the acknowledged failure of the study to recruit adequate numbers of patients for the desired statistical power, we agree with the authors that further consideration of the range of issues around POC monitoring is required before a definite decision can be made regarding the feasibility of its national implementation. At the very least, a more robust economic analysis utilising a less controversial health outcome indicator would assist in these future discussions.
Miss Leanne Stafford
Dr Luke Bereznicki
Professor Gregory Peterson
Unit for Medication Outcomes Research and Education (UMORE), School of Pharmacy, University of Tasmania
References
1. Jackson SL, Bereznicki LR, Peterson GM, Marsden KA, Jupe DM, Vial JH, et al. Accuracy and clinical usefulness of the near-patient testing CoaguChek S international normalised ratio monitor in rural medical practice. Aust J Rural Health 2004;12(4):137-42.
2. Jackson SL, Peterson GM, House M, Bartlett T. Point-of-care monitoring of anticoagulant therapy by rural community pharmacists: description of successful outcomes. Aust J Rural Health 2004;12(5):197-200.
3. Jeffrey E, Bereznicki L, Stafford L, Peterson G. Evaluation of a clinical pathway to enable patient self-monitoring of anticoagulation. In: HAA Handbook and Final Programme, 2009 19-21 October; Adelaide: p. 276.
4. Peterson G, Stafford L, Bereznicki L, van Tienen E, Jackson S, on behalf of the Project Team. The role of community pharmacy in post hospital management of patients initiated on warfarin. In: Pharmacy Practice Research Summit, 2010 2-4 March; Canberra.
5. Peterson G, Jackson S, Gee P, Bereznicki L, Bereznicki B, Fitzmaurice K, et al. Improving the management of warfarin in aged care facilities utilising an innovative IT solution. In: Australasian Pharmaceutical Science Association Conference, 2007 8-11 December; Manly.
6. Samsa GP, Matchar DB. Relationship between test frequency and outcomes of anticoagulation: A literature review and commentary with implications for the design of randomized trials of patient self-management. J Thromb Thrombolysis 2000;9(3):283-92.
* Several of these programs were funded by the Australian Government Department of Health and Ageing as part of the Fourth Community Pharmacy Agreement through the Fourth Community Pharmacy Agreement Grants Program managed by the Pharmacy Guild of Australia.
Competing interests
UMORE has received support from Roche Diagnostics Australia for warfarin-related research, principally in the supply of POC INR monitors.