Note from the National Guideline Clearinghouse (NGC): The National Institute for Health and Care Excellence (NICE) commissioned an External Assessment Group to perform a systematic literature review on the technology considered in this diagnostics guidance and prepare a Diagnostics Assessment Report (DAR). The DAR for this guidance was prepared by the Aberdeen Health Technology Assessment (HTA) group (see the "Availability of Companion Documents" field).
Assessment Design and Results — Clinical Effectiveness
Methods for Standard Systematic Review of Effectiveness
Data Extraction Strategy
A data extraction form was designed and piloted for the purpose of this assessment (see Appendix 2 in the DAR). One reviewer extracted information on study design, characteristics of participants, settings, characteristics of interventions and comparators, and relevant outcome measures. A second reviewer crosschecked the details extracted by the first reviewer. There was no disagreement between reviewers.
Assessment of Risk of Bias in Included Studies
A single reviewer assessed the risk of bias of the included studies and findings were crosschecked by a second reviewer. There were few disagreements which were resolved by consensus or arbitration by a third reviewer. The reviewers were not blinded to the names of studies' investigators, institutions, and journals. Studies were not included or excluded purely on the basis of their methodological quality. The risk of bias assessment for all included randomised controlled trials (RCTs) was performed using the Cochrane Risk of Bias tool (see Appendix 3 in the DAR). Critical assessments were made separately for all main domains: selection bias ('random sequence generation', 'allocation concealment'), detection bias ('blinding of outcome assessor'), attrition bias ('incomplete outcome data') and reporting bias ('selective reporting'). The 'blinding of participants and personnel' was not considered relevant for this assessment due to the nature of intervention being studied (i.e., patient performing the test themselves or under supervision of health care professionals). However, the Assessment Group collected information related to the blinding of outcome assessors, which was considered relevant to the assessment of risk of bias.
Each included study was judged as 'low risk of bias', 'high risk of bias' or as 'unclear risk of bias' according to the criteria for making judgments about risk of bias described in the Cochrane Handbook for Systematic Reviews of Interventions. Adequate sequence generation, allocation concealment, and blinding of outcome assessor were identified as key domains for the assessment of the risk of bias of the included trials.
For dichotomous data (e.g., bleeding events, thromboembolic events, mortality), relative risk (RR) was calculated. For continuous data (e.g., time in therapeutic range), weighted mean difference (WMD) was calculated. Where standard deviations were not given, the Assessment Group calculated them using test statistics wherever possible. The RR and WMD effect sizes were meta-analysed as pooled summary effect sizes using the Mantel-Haenszel method and the inverse-variance method, respectively. Confidence intervals were also calculated (95% CIs). To estimate the summary effect sizes, both fixed effects and random effects models were used with RR and WMD. In the absence of clinical and/or statistical heterogeneity, the fixed effects model was selected as the model of choice while the random effects model was used to crosscheck the robustness of the fixed effects model. However, in the presence of either clinical or statistical heterogeneity, the random effects model was chosen as the preferred method for pooling the effect sizes, as in this latter situation, the fixed effect method is not considered appropriate for combining the results of included studies. Heterogeneity across studies was measured by means of the Chi-squared statistic and also by the I-squared statistic, which describes the percentage of variability in study effects that is explained by real heterogeneity rather than chance. It is worth noting that, for bleeding and thromboembolic events, the Assessment Group used the total number of participants who were actually analysed as denominator in the analyses. In contrast, for mortality, they used the total number of participants randomised as denominator because participants could have died due to any causes after randomisation but before entering the self-monitoring programme.
Apart from the pre-specified subgroups analysis according to the type of anticoagulation therapy management (self-testing and self-management), the Assessment Group performed a post-hoc subgroup analysis according to the type of the target clinical condition (i.e., atrial fibrillation, heart valve disease, and mixed clinical indication) and one according to the type of service provision for anticoagulation management (i.e., primary care, secondary care, and shared provision). Where trials had multiple arms contributing to different subgroups, the control group was subdivided into two groups to avoid a unit of analysis error.
Sensitivity analyses were planned in relation to some of the study design characteristics. The methodological quality (low/high risk of bias), and the different models of the CoaguChek system were identified at protocol stage as relevant aspects to explore in sensitivity analyses. In addition to those pre-specified in the protocol, a sensitivity analysis was performed by excluding the studies conducted in the UK.
Review Manager software (Review Manager 5.2, 2012) was used for data management and all relevant statistical analyses for this assessment. Where it proved unfeasible to perform a quantitative synthesis of the results of the included studies, outcomes were tabulated and described in a narrative way.
See Section 3 in the DAR for more information on clinical effectiveness analysis.
Assessment of Cost-effectiveness
Independent Economic Assessment
A de novo economic model was developed in TreeAge Pro (TreeAge Software, Williamstown, MA, 2013). The model was designed to assess the cost-effectiveness of self-monitoring (self-testing and self-management) using alternative point-of-care devices: CoaguChek XS system, INRatio2 prothrombin time/international normalized ratio (PT/INR) monitor, and ProTime Microcoagulation system.
The model was populated using data derived from the systematic clinical effectiveness review, other focused reviews to inform key parameters (e.g. baseline risks), routine sources of cost data, and where necessary some study specific cost estimates based on expert opinion. The model was built and analysed in accordance with the NICE reference case for the evaluation of diagnostic tests and devices.
Relevant Patient Population(s)
The model compared the alternative monitoring strategies for a hypothetical cohort of people with atrial fibrillation or an artificial heart valve. These two groups represent the majority of people on long-term vitamin K antagonist therapy. While self-monitoring of INR is relevant to other patient groups, including those with venous thrombotic embolism, there was insufficient data to explicitly model cost-effectiveness for all groups individually. Furthermore, the majority of studies informing the relative effects of alternative monitoring strategies were derived from trials including predominantly people with atrial fibrillation and/or an artificial heart valve. Therefore, the base case modelling exercise was carried out for a mixed cohort consisting of people with one or other of these two conditions.
Monitoring Strategies to Be Evaluated
The economic model incorporated the pathways of care that individuals currently follow under standard practice in the National Health Service (NHS), as well as proposed new pathways for self-testing and self-management (informed by a review of current guidelines and expert opinion). Current practice was dichotomised in the model as standard monitoring in primary care and standard monitoring in secondary care. In the base case analysis, the proportional split between standard primary and secondary care INR monitoring was taken from the manufacturer's submission for NICE technology appraisal 256 (TA256) . Based on a survey of providers in England and Wales carried out in 2011, it was estimated that 66.45% and 33.55% of warfarin monitoring appointments were managed in a primary and secondary care setting, respectively. These figures were accepted by the independent evidence review group and appraisal committee for NICE TA256.
In terms of self-monitoring, the model incorporated both self-testing and self-management strategies using the alternative devices identified in the scope. However, the cost-effectiveness of self-monitoring was assessed as a whole, and it was assumed in the base case analysis that 50% of people would self-test whilst 50% would self-manage. These proportions were varied in sensitivity analysis. Self-testing and self-management strategies were costed separately for each device based on the assumption that self-testing people phone in their results from all tests undertaken, while self-managing people manage their dosing independently. In reality, some self-monitoring people are likely to fall somewhere in between these two strategies, and several alternative scenarios were also assessed.
Framework (Method of Synthesis)
The alternative monitoring pathways, informed by review of previous guidance and expert opinion, were embedded in a Markov model simulating the occurrence of adverse events over time. The adverse events included in the model were ischaemic stroke (minor, non-disabling, and major, disabling or fatal), systemic embolism (SE), minor haemorrhage, and major haemorrhage (intra-cranial haemorrhage [ICH], including haemorrhagic stroke [HS], gastrointestinal [GI] bleed, and others). Systemic embolism was treated as a transient event within the model, such that people surviving this event returned to baseline levels of quality of life and did not incur on-going costs and morbidity. Minor haemorrhage was handled in the same way. Ischemic stroke and ICH were assigned post event states associated with additional costs and quality of life decrements.
The model simulated transitions between the discrete health states, and accumulated costs and quality adjusted life years on a quarterly (three-month) cycle. Within each three-month cycle, the simulated cohort was exposed to a risk of the aforementioned events as well as death from other causes. A constraint was applied whereby simulated people could only experience one event per cycle. A further simplifying structural assumption was applied, such that following a major ischaemic stroke or ICH, no further events were explicitly modelled. However, all-cause mortality was inflated following these events to account for the increased risk of death.
See Section 4 in the DAR for detailed information on cost-effectiveness analysis.