The Diagnostics Advisory Committee reviewed the evidence available on the clinical and cost effectiveness of self‑monitoring coagulometers for self‑testing or self‑managing coagulation status in people on long‑term vitamin K antagonist therapy who have atrial fibrillation or heart valve disease.
The Committee considered the clinical evidence on the use of point‑of‑care coagulometers in people with atrial fibrillation or artificial heart valves. The Committee noted that 26 randomised controlled trials compared the use of point‑of‑care coagulometers for self‑monitoring with standard anticoagulation control. The Committee noted that self‑monitoring nearly halved the risk of thromboembolic events and substantially reduced the risk of mortality in people with artificial heart valves compared with standard monitoring. However, the Committee also noted that self‑monitoring did not result in a statistically significant reduction in the number of major and minor bleeding events compared with standard monitoring. The Committee discussed the heterogeneity in the trials and the applicability of the pooled results from the meta‑analysis of the trial data to the UK population. It noted that the meta‑analysis results showed low statistical heterogeneity and concluded that self‑monitoring offered clinical benefit because it was likely to result in a significant reduction in thromboembolic events. The Committee concluded that the pooled effect estimates from the meta‑analysis were likely to be applicable to the UK because there are no confounding biological differences between people receiving vitamin K antagonist therapy in the UK and those in other countries.
The Committee discussed that 22 of the 26 trials included in the assessment investigated the use of the CoaguChek system and considered the different versions of the CoaguChek systems used in these trials. The Committee noted that there are substantial technical differences between the CoaguChek S system and the CoaguChek XS system and heard from clinical specialists and the manufacturer that changes had been made to the different versions to improve reliability and accuracy. The Committee considered the performance of the CoaguChek S and XS systems compared with the gold standard of laboratory‑based INR testing and noted that the precision and accuracy of the 2 CoaguChek versions correlated with that of laboratory‑based measurements. The Committee concluded that results from the CoaguChek XS system were likely to be at least as good as those obtained from trials in which previous versions of the system were used. The Committee also noted that 4 of the 22 trials investigated the use of the CoaguChek XS system and that 2 of these trials demonstrated a significant improvement in time in therapeutic range. The Committee concluded that it was appropriate to pool the results of trials using different versions of the CoaguChek system and that these pooled results could demonstrate the clinical effectiveness of self‑monitoring using the CoaguChek XS version of the system.
The Committee considered the evidence for the 2 different self‑monitoring coagulometers: the CoaguChek XS system and the INRatio2 PT/INR monitor. The Committee noted that 22 of the 26 trials included in the assessment investigated the use of the CoaguChek system and noted that there was no direct randomised controlled trial evidence to show the clinical effectiveness of the INRatio2 PT/INR monitor. The Committee considered the evidence that showed the 2 coagulometers had a broadly similar performance in precision and accuracy with regard to time in therapeutic range measurement when compared with the gold standard of laboratory‑based INR testing and therefore concluded that it was appropriate to extrapolate the clinical‑effectiveness data from the CoaguChek system to the INRatio2 PT/INR monitor.
The Committee discussed the usability of the coagulometers and noted that small differences in devices can sometimes result in large differences in behaviour. The Committee considered the results of the systematic review by Christensen and Larsen (2012) and the unpublished review of over 5000 patients from the National Thrombosis Service in the Netherlands. It noted that although the results of the unpublished review suggested that any potential differences in usability between the 2 monitors did not affect their clinical effectiveness, the unpublished review could not be considered methodologically robust. The Committee concluded that based on the systematic review, any potential differences in the usability of the coagulometers did not appear to affect their ability to measure INR.
The Committee considered the differences in clinical outcomes between people who were self‑managing their anticoagulation control and those who were self‑testing. The Committee noted that there was a statistically significantly greater reduction in thromboembolic events among people who self‑managed compared with those who self‑tested. The Committee also noted that when only minor bleeding events were assessed, a statistically significantly increased risk was seen in self‑testing participants compared with those in standard care. All‑cause mortality was lower through self‑management but not through self‑testing. The Committee discussed possible reasons for the differences in results between self‑managing and self‑testing, and it heard from clinical specialists that people who self‑manage their coagulation control may behave differently to those who self‑test because they have greater responsibility for managing their coagulation control. The Committee noted that the largest trial in the assessment of self‑testing did not show a reduction in clinical adverse events but did show an increase in the time in therapeutic range (Matchar et al. 2010). The Committee also noted that this trial had a high standard of coagulation control in the control arm, which could explain why no statistically significant difference in clinical adverse events was detected between the self‑testing group and the standard care group. The Committee concluded that the high standard of coagulation control in the control arm of the trial may not reflect general UK clinical practice and so it was plausible that the increase in time in therapeutic range would lead to a statistically significant reduction in clinical adverse events if compared with UK standard coagulation control practice. The Committee concluded that self‑testing and self‑managing were likely to be clinically effective and that self‑testing was often a step towards self‑management in clinical practice.
The Committee considered the clinical evidence for using self‑monitoring in the population group with atrial fibrillation. The Committee noted that only 2 trials investigated self‑monitoring in people with atrial fibrillation and 19 trials investigated self‑monitoring in a mixed population that included people with atrial fibrillation. The Committee heard from clinical specialists that the clinical outcomes for people with atrial fibrillation are similar to those for people with artificial heart valves. The Committee also heard that people with artificial heart valves may be a younger population than people with atrial fibrillation. The Committee noted that it was not possible to isolate the data for people with atrial fibrillation from the mixed populations investigated in the 19 trials but concluded that it was likely that self‑monitoring would result in similar clinical benefits in people with atrial fibrillation to those achieved in people with artificial heart valves.
The Committee considered the cost‑effectiveness analysis carried out by the External Assessment Group on self‑monitoring. In the base‑case analysis, self‑monitoring with the CoaguChek XS system resulted in an incremental cost‑effectiveness ratio (ICER) of around £300 per quality‑adjusted life year (QALY) gained (based on the pooled effect estimates from the meta‑analysis) compared with standard monitoring. Self‑monitoring with the INRatio2 PT/INR monitor dominated (that is, was less expensive and more effective than) the CoaguChek XS system and standard care, although the Committee noted that there was no direct randomised controlled trial evidence for the clinical effectiveness of the INRatio2 PT/INR monitor so clinical effectiveness equivalent to the CoaguChek XS system was assumed in the base case. The Committee concluded that self‑monitoring with the CoaguChek XS system is cost effective in light of the reduction in thromboembolic events seen in the pooled results of the trial data. However, although the INRatio2 PT/INR monitor dominated standard monitoring in the base‑case analysis, the Committee did not consider this result to be as robust as that for the CoaguChek XS system because there was no direct evidence of clinical effectiveness for the INRatio2 PT/INR monitor that showed a reduction in thromboembolic events. The Committee noted the similar performance of the CoaguChek XS system and the InRatio2 PT/INR monitor compared with the gold standard of laboratory‑based INR testing and concluded therefore that self‑monitoring with the INRatio2 PT/INR monitor was likely to represent a cost‑effective use of NHS resources.
The Committee considered the cost effectiveness of self‑testing and self‑managing individually. The findings showed that self‑management alone is highly cost effective (dominant) but that self‑testing alone is not cost effective, compared with standard monitoring. The Committee noted that these findings were based on the contrasting pooled‑effect estimates obtained from the meta‑analysis of randomised controlled trials, based on thromboembolic events while self‑testing and self‑managing. The Committee discussed the impact of 1 large trial by Matchar et al. (2010) (see section 6.6) on the cost effectiveness of self‑testing and noted that although this trial did not show a reduction in clinical adverse events, it did show an increase in the time in therapeutic range. The Committee discussed the impact on the ICERs for self‑testing if the economic model was driven by time in therapeutic range rather than adverse events. The Committee concluded that self‑testing may be more cost effective if the model had been based on time in therapeutic range. The Committee also considered the costs of self‑managing and self‑testing and noted that self‑testing was more expensive because of higher administration costs. The Committee heard from the External Assessment Group that if the pooled‑effect estimates from self‑monitoring were applied to self‑testing, self‑testing would become cost effective even with the higher administration costs this incurred. The Committee concluded that it was likely that the increase in time in therapeutic range shown for self‑testing in the trial would lead to a reduction in adverse events compared with standard clinical practice in the UK. The Committee therefore concluded that it was likely that the clinical benefits of self‑testing had been underestimated in the economic analyses and that both self‑testing and self‑managing were cost effective.
The Committee considered the impact on people whose anticoagulation therapy is monitored by standard clinical practice. The Committee acknowledged the additional costs and inconvenience for patients of travelling to specialist clinics or hospital to be monitored. The Committee also noted the loss in productivity through absence from work or school to attend clinic appointments. The Committee acknowledged these costs were incurred outside the healthcare system and therefore not included in the reference case; however, it considered that self‑monitoring may reduce the costs and inconvenience incurred by the patient. The Committee also noted that monitoring anticoagulation therapy can have a substantial impact on the quality of life of patients and their families because of the anxiety associated with the risks of bleeding and the consequent behavioural changes such as the length of time they are willing to spend away from home or the distances they are willing to travel. The Committee concluded that the main benefits of self‑monitoring involve reducing the substantial burden associated with monitoring anticoagulation therapy for the patient and their families.
The Committee considered the different methods of self‑monitoring for people having long‑term vitamin K antagonist therapy. The Committee heard from a clinical specialist that computer algorithms may be used by some services to determine dose adjustments. The Committee noted that the cost of software licensing had not been included in the cost‑effectiveness analyses and it discussed the implications of computer‑based dosing for people who would self‑manage their coagulation status. The Committee heard that there were alternative methods to determine dose adjustments and that a lack of internet access should not restrict a person's access to self‑monitoring. The Committee considered an additional analysis by the External Assessment Group that investigated the impact of an additional cost for dose adjustment software on the base‑case ICERs for self‑managing. The Committee noted that the additional cost of software would need to be greater than £190 per patient per year to substantially affect the cost‑effectiveness of self‑managing with the coagulometers. The Committee concluded that the additional cost of software was unlikely to exceed this value and therefore, even with this potential additional cost, self‑managing with the point‑of‑care coagulometers would still represent a cost‑effective use of resources in the NHS.
The Committee considered the benefits for patients receiving vitamin K antagonist therapy of using point‑of‑care coagulometers. It heard from a patient expert on the Committee that self‑monitoring is important to psychological wellbeing because it provides a sense of control for the patient and removes the need to frequently attend clinics or hospitals, which serve as a constant reminder of their condition. The Committee also heard that self‑monitoring allows people to travel to visit, or act as a carer for, other family members, without having to worry about attending testing appointments or if testing facilities are available in other countries. The Committee also heard that the current variation in access to self‑testing strips on prescription for self‑monitoring was of concern to patients because it restricted their freedom to move GP practice or move house to a different area in case the testing strips would no longer be prescribed. The Committee concluded that the benefits of self‑monitoring for patients were not fully captured in the cost‑effectiveness analyses.
The Committee considered the similarities between self‑monitoring coagulation status and self‑managing diabetes. The Committee heard from a patient expert that some patients are used to self‑testing for conditions such as diabetes, hypertension and heart conditions. The Committee also heard from a clinical specialist that although there were similarities between self‑testing for different conditions, there were intrinsic differences between self‑testing for diabetes and coagulation. Vitamin K antagonists are more sensitive to diet and exercise, and act over a longer period of time than insulin. Therefore, the dose response for vitamin K antagonists is less predictable than for insulin and the risk of adverse events is perceived to be higher. The clinical specialist also reported that some patients were successfully self‑monitoring their coagulation status but not all people receiving vitamin K antagonist therapy will be able to self‑monitor and some may not wish to. The Committee noted that some groups of patients who may have difficulty with self‑monitoring, such as children or those with a disability, may be able to self‑test or self‑manage with the help of a carer. The Committee concluded that there are different considerations for self‑monitoring of coagulation status to those made for self‑testing for diabetes, and that the decision for a patient to self‑monitor should be made after a thorough discussion and subsequent agreement between the patient and the healthcare professional.
The Committee considered the impact of the increasing use of non‑vitamin K antagonist oral anticoagulant drugs, which do not involve monitoring because of their predictable dose response. It heard from clinical specialists that there are factors that may influence clinical decisions and affect the number of people receiving warfarin: people receiving warfarin who have stable INRs may be unlikely to switch to non‑vitamin K antagonist oral anticoagulants, and the non‑vitamin K antagonist oral anticoagulants may be unsuitable for some, such as people with mechanical heart valves, certain people with renal or liver dysfunction or those taking drugs that cannot be taken at the same time as the non‑vitamin K antagonist oral anticoagulants. The Committee concluded that, because the non‑vitamin K antagonist oral anticoagulants would not be suitable for all people who need anticoagulant therapy, and there are many people who will receive warfarin therapy rather than non‑vitamin K antagonist oral anticoagulant therapy, self‑monitoring coagulometers are still of clinical importance to the NHS and patients.
The Committee considered the need for quality control of individual patient coagulometers. The Committee heard from a clinical specialist that the National External Quality Assessment Service runs a scheme to ensure the accuracy of coagulometers used by healthcare professionals, and coagulometers used by patients could be checked against a professional coagulometer to ensure accuracy. The Committee concluded that this was a reliable method of ensuring accuracy of individual coagulometers.