Systematic review and cost-effectiveness evaluation of 'pill-in-the-pocket' strategy for paroxysmal atrial fibrillation compared to episodic in-hospital treatment or continuous antiarrhythmic drug therapy

Abstract

Background: Atrial fibrillation (AF) is a tachyarrhythmia characterised by uncoordinated atrial activation with consequent deterioration of impairment of atrial function and a rapid, irregular heartbeat. The annual incidence rate of paroxysmal AF (PAF) has been estimated at 1.0 per 1000 person-years (95% confidence interval 0.9 to 1.1), and reported prevalence rates show wide variations depending on age and country. Conventional treatment strategies for PAF focus on the suppression of paroxysms of AF and return to normal sinus rhythm.

Objectives: To summarise the results of the rapid reviews of the clinical effectiveness and cost-effectiveness literature describing the pill-in-the-pocket (PiP) approach for the treatment of patients with PAF; and to develop an economic model to assess the cost-effectiveness of PiP compared with in-hospital treatment (IHT) or continuous antiarrhythmic drugs (AADs) for the treatment of patients with PAF.

Data sources: Ovid MEDLINE and Ovid OLDMEDLINE 1950 to present with Daily Update were searched. The following electronic databases were searched for ongoing trials: Health Services Research Projects in Progress, ClinicalTrials.gov, metaRegister of Current Controlled Trials, BioMed Central, World Health Organization International Clinical Trials Registry Platform, ClinicalStudyResults.org and the National Library of Medicine Gateway.

Review methods: Inclusion criteria, which included patients suffering from PAF, were independently applied to all identified references by two reviewers (JH and CMS). Electronic searches were conducted to identify clinical effectiveness and cost-effectiveness evidence describing the use of a PiP strategy for the treatment of PAF, published since the release of the Royal College of Physicians' national guidelines on AF in June 2006. A Markov model was constructed to examine differences between three PAF strategies (PiP, AAD and IHT) in terms of cost per quality-adjusted life-year (QALY). A Markov model structure was chosen because it is assumed that PAF is a condition that causes patients to move between a limited number of relevant health states during their lives.

Results: The search strategies for clinical studies identified 201 randomised controlled trials (RCTs). Of the 201 RCTs identified, 12 were deemed to be relevant to the decision problem as they included drugs used to treat PAF; summary data were abstracted from these studies in order to inform the development of the economic model only. The model results indicate that the PiP strategy is slightly less effective than the other two strategies, but also less costly (incremental cost-effectiveness ratio of 45,916 pounds per QALY when compared to AAD, and 12,424 pounds per QALY when compared to IHT). The one-way sensitivity analyses performed do not show substantial changes in relative cost-effectiveness except in relation to the age of patients, where PiP dominates AAD in men over 65 years and in women over 70 years. At a threshold of 25,000 pounds per QALY, IHT has the maximum probability of being cost-effective at this threshold. For threshold values between 0 pounds and 9266 pounds per QALY, PiP is the option exhibiting the maximum probability of being cost-effective. The AAD strategy has a very poor probability of being cost-effective under any threshold. However, none of the strategies considered has more than a 40% probability of being cost-effective at a threshold of 25,000 pounds per QALY at any threshold level. This demonstrates the uncertainty around the parameters and its effect on the decision to choose any one strategy over the others.

Limitations: Most of the data used to populate the model have been taken from studies with populations that do not match the patient population specified in the decision problem. Populating the model in this way was unavoidable as there was a paucity of published clinical effectiveness and cost-effectiveness data describing a PiP strategy for this highly specific group of patients.

Conclusions: Overall, a PiP strategy seems to be slightly less effective (i.e. fewer QALYs gained) than AAD and IHT, but is associated with cost savings. A PiP strategy seems to be more efficacious and cost-effective than an AAD strategy in men over 65 years and women over 70 years, but this is principally due to a very slight difference in QALY gained by the PiP strategy. A change in clinical practice that includes the introduction of PiP may save costs, but also involves a reduction in clinical effectiveness compared to existing approaches used to treat patients with PAF. Uncertainty in the available clinical data means there was insufficient evidence to support a recommendation for the use of PiP strategy in patients with PAF. Further research should identify outcomes of interest such as adverse events and recurrent AF episodes in an RCT setting because the only clinical study addressing these issues, even partially, is not an RCT but a descriptive analysis. Patient preferences also need to be considered in any future research designs.