Economic evaluation of diagnostic sleep studies for obstructive sleep apnoea: a systematic review protocol

Obstructive sleep apnoea (OSA) is a condition characterised by airflow cessation or obstruction caused by complete or partial collapse of the upper airway, which affects 9 to 38% of the general adult population [1, 2]. Unfortunately, underdiagnosed OSA is common, occurring in 75 to 90% of individuals in the US general adult population and adult surgical patients [3‐5]. In Australia, studies have reported up to 10% of individuals remain undiagnosed and untreated in the community [6, 7]. When left untreated, individuals with OSA are at heightened risks of metabolic syndrome [8‐11], cardiovascular diseases [12‐14], neurocognitive abnormalities [15, 16], mental health problems [17], behavioural alteration [18], traffic collisions [19], workplace accidents and productivity losses [20‐22], reduced quality of life [23, 24] and premature death [25‐28]. The high prevalence and various adverse health outcomes associated with OSA contribute to significant economic costs [29‐31]. In Australia, OSA accounts for $408.5 million of total health system costs and $2.6 billion of indirect financial costs, making it the highest contributor to sleep disorder-related expenses [32].

The health and economic burden related to OSA can be diminished with effective and efficient case-identification and treatment. The latter is readily available and found to be cost-effective [33‐35]. A recent systematic review that investigated the impact of OSA therapies suggested that treating OSA has a positive effect on quality-adjusted life-years (QALYs) and can reduce health care utilisation and costs [36]. Despite the existence of effective treatments, case identification of OSA remains a challenge and creates the need to widen diagnostic access for those with suspected OSA [37]. The current gold standard for diagnosing OSA is full overnight laboratory polysomnography (PSG, i.e. level I study)) [38‐40]. However, laboratory PSG is expensive, time-consuming and labour intensive, which creates further access barriers to OSA diagnosis [38‐40]. Researchers have investigated alternatives to laboratory PSG, which include portable PSG (i.e. level 2 [L2] study) and limited channel testing (i.e. level 3 [L3] and level 4 [L4] studies) that are typically performed in the patient’s home. While full PSG typically (L1 and L2 studies) utilises more than seven recording channels, L3 (polygraphy) measures four to seven channels, while L4 uses only one to three physiological measures to derive a diagnosis [41], making them less labour intensive and less time consuming to report. Previous studies have demonstrated favourable results showing a high level of agreement between limited-channel testing and in-laboratory PSG, particularly when diagnosing moderate to severe OSA [42, 43].

Limited channel tests can advance and widen access to diagnostic services and treatment initiation in OSA, ultimately reducing the substantial economic burden related to OSA. However, to date, research into their cost-effectiveness remains limited and contentious. Scarce evidence exists for the cost-effectiveness of L4 devices despite results suggesting that a single-channel sensor offers a cheaper alternative to polysomnography [44]. Previous economic evaluation models assessing the economic impact of polygraphy (i.e. L3 testing) relative to PSG using published data showed contradictory results [45‐47]. However, several empirical studies found consistent results signifying similar effectiveness and lower costs of L3 studies vs. PSG [48‐53]. Despite the longstanding controversy of the value and efficiency of polygraphy, the test was approved by the US Centres for Medicare and Medicaid Services and the American Academy of Sleep Medicine practice parameters for use in selective populations in 2008 [54, 55]. Following the change in the health insurance policy that enables reimbursement for home sleep polygraphy in the USA, there has been a growing demand for limited-channel testing for OSA [56]. However, L3 tests are still not compensated by the Medicare Benefits Schedule in Australia where access to OSA diagnosis remains restricted [57].

Robust evidence about the potential economic benefit of OSA diagnostic testing is required to guide decision-makers towards better allocation of scarce healthcare resources. Only a few studies have sought to investigate the health economic impact of OSA diagnostic studies [58, 59]. One previous literature review included nine studies, predominantly examining the cost of diagnostic testing [58]. In contrast, the most recent review has focused on constructing a cost-effectiveness model of OSA management based on existing data [59]. However, none have thoroughly examined the evidence related to the cost and effectiveness of limited channel testing compared to PSG. Hence, we aim to address this gap in the literature by conducting a structured systematic review of economic studies to capture, critically appraise and synthesise data related to health economic outcomes of OSA diagnostic tests, focusing on L3, L4 and PSG testings (LI and L2) in adult OSA populations. Our review will generate information that will help decision-makers determine whether implementing limited channel testing provides value for money for OSA case identification.

This protocol has been registered with the PROSPERO database (registration number CRD42020150130) and is being reported in accordance with the reporting guidance provided in the Preferred Reporting items for Systematic Reviews and Meta-Analyses Protocols (PRISMA-P) statement [60] (see checklist in Additional file 1).

The proposed systematic review will be reported in accordance with the reporting guidance provided in the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement [71]. This systematic review will present current evidence relating to the economic impact of OSA diagnostic testing methods (i.e. limited channel versus full sleep study testing) using a reproducible and transparent systematic literature search. We have outlined comprehensive inclusion and exclusion criteria of eligible studies, which encompass study design, context, target population, intervention and comparator, cost and outcomes measures. A search strategy and a list of databases for searching have been determined, followed by a strategy for data extraction and synthesis. Included studies will be subjected to critical appraisal using standardised tools tailored for model and non-model based economic evaluation studies. Some potential limitations of this review include the heterogeneity of findings and cost perspective, which can influence the aggregability, transferability and generalisability of extracted data. Another limitation is information bias originating from our restriction to studies reported in English language and adult populations. However, this review is timely in addressing and identifying gaps in current evidence regarding diagnostic options for OSA. Our findings can potentially inform evidence-based decision making when implementing limited channel sleep study testing, particularly in Australia, where undiagnosed OSA and constrained access to diagnostic services remains an important issue.