• COPD epidemiology
• health economist
• COPD exacerbations
• emphysema
• respiratory measurement
• smoking cessation
• tobacco and the lung

To screen or to target the missing millions with chronic obstructive pulmonary disease (COPD)? Moreover, what if we decide to invest in case-finding: do we get value for our money? Now, that’s the real question! First, however, we should establish what to pay for. Earlier research showed that one-third to half of COPD patients have not been diagnosed.1 2 Indeed, even if patients are already on the radar of a general practitioner (GP), opportunities to diagnose COPD are frequently being missed.3 COPD is still a major cause of severe patient symptoms (eg, breathlessness), costly hospital admissions and death.4 5 Notably, it is likely that earlier diagnosis and treatment to lessen breathlessness could reduce the risk of comorbidities associated with low physical activity developing and so contribute to healthier ageing.5 Those in favour of case-finding have, therefore, recently highlighted the need for early detection.6 Sceptics, on the other hand, argue that it is only useful if we can effectively slow down or halt the natural course of the disease.7

Over the last decade, multiple COPD case-finding strategies have been described.8 9 Most strategies seem effective, but generally few large, well-designed randomised controlled trials (RCTs) have been performed and outcomes have been very heterogeneous.10 Notably, this heterogeneity has limited our evidence on the optimal population at-risk to target, the best screening tools and diagnostic tests to use and, moreover, the cost-effectiveness of active case-finding. Limited evidence has also resulted in guideline committees and policymakers being hesitant in widely promoting case-finding.

This brings us back to the first question: what do the current guidelines actually tell us? The updated Global Initiative for Chronic Obstructive Lung Disease (GOLD) report discourages spirometry in asymptomatic populations without exposure to tobacco or other noxious stimuli. However, since 2019, it advocates active case-finding in symptomatic and/or at-risk groups by either risk scores based on electronic records, or via screening questionnaires.11 Key evidence for this guideline update came from the UK.

In 2016, Jordan et al reported on the largest case-finding trial that ever took place and in which almost 75 000 UK subjects were randomly assigned to either a targeted strategy (active or opportunistic case-finding) or usual care.12 This TargetCOPD trial showed that targeted case-finding strategies yielded 4% new COPD cases vs 1% with usual care. Within the two targeted strategies, active case-finding was more effective than opportunistic case-finding (5% vs 2%). While providing convincing evidence that case-finding was effective, the cost-effectiveness analysis that was performed alongside the trial, suffered from several limitations.13 Among others, it only included costs and short-term effects during the trial and did not calculate lifetime costs per quality-adjusted life year (QALY) gained as is recommended by the UK’s National Institute for Health and Care Excellence (NICE).

In this issue of Thorax, Lambe et al have filled this longstanding and policy-relevant evidence gap by assessing the long-term cost-effectiveness of COPD case-finding from the UK National Health Service perspective.14 In their economic evaluation, the authors estimated the cost-effectiveness of a 3-yearly systematic case-finding programme targeted to patients aged over 50 years that ever smoked and were routinely managed in the UK primary care. As a basis for their Markov model-based analysis, patient-level effectiveness data from the active case-finding arm of the previously mentioned TargetCOPD trial were used. Active case-finding was defined as actively providing, and additionally mailing out, a respiratory symptom questionnaire to undiagnosed primary care-managed patients at risk for COPD and performing diagnostic spirometry in symptomatic patients. Importantly, while in the TargetCOPD trial this was only a ‘one-off’ intervention, in their model, the authors assumed that this would be repeated every 3 years. Over the course of a lifetime, virtual patients could move between the model states ‘disease-free’, ‘undiagnosed disease’, ‘diagnosed disease’ and ‘dead’. To extrapolate the short-term RCT’s findings and estimate costs and effects beyond the length of the RCT, information on COPD’s long-term natural course of disease, costs and effects of treatment was indirectly estimated from the linked Birmingham COPD cohort, UK primary databases and literature. Of note, this did not include costs for routine care or comorbidities, but the characteristics of the modelled cohort closely matched the trial population. Of the newly diagnosed, 96% was GOLD I or II and one-third was assumed to commence treatment annually. The treatment effect in these newly diagnosed was based on the point estimate of a meta-analysis of long-acting muscarinic antagonist (LAMA) RCTs. Applying these data, in the base-case analysis (ie, using the most likely estimates for costs and effects,15 the incremental cost-effectiveness ratio (ICER) of active case-finding was £16 596 per QALY.

While being the first of its kind and methods generally solid, as any model-based study, this analysis was also surrounded by some extent of uncertainty and did not come without limitations. For example, this includes the assumptions on the effects of LAMA treatment on disease progression and outcomes. We may wonder whether treatment efficacy based on RCTs, usually performed in selected more severe populations, will be similar in the relatively mild newly diagnosed COPD patients identified using real-world case-finding? Even if the case, the assumed effect of LAMA on mortality, although minor (OR: 0.98), has not been convincingly shown in any RCT so far. Another matter of debate: effects on utility gain were based on a single fluticasone study and assumed to be the same across all GOLD states. Is this realistic? More common COPD treatments, such as long-acting beta agonist (LABA)/LAMA combination therapy, may result in better utility and accompanied long-term cost reductions. Most importantly smoking cessation, the cornerstone treatment of COPD, was not taken into account at all. This absence of smoking cessation as a treatment possibility suggests that the cost per QALY could likely be lower than currently specified. Of note, in already diagnosed COPD patients, smoking cessation has an ICER as low as £2000/QALY,16 and is even cost-saving if provided as a general public health intervention.17 Should we not better invest our money in smoking cessation instead of case-finding? Possibly, targeting specific occupations and non-wealthy individuals may further improve the value of screening. It is, however, reassuring to see that the authors were generally well aware of these limitations and have provided several sensitivity and scenario analyses to show the impact of some of their analytical uncertainties.

Taking this uncertainty for granted, back to the second question of this editorial’s introduction: what is the economic value of case-finding? That is, how should we interpret the ICER of £16 696/QALY? A first conclusion is that the cost-effectiveness of active case-finding is at least below NICE’s common threshold to consider an intervention cost-effective (ie, £30 000/QALY). Second, can we say something about its relative value? In 2012, the London Respiratory Network introduced the COPD value pyramid that importantly showed the high value of influenza vaccination, smoking cessation and pulmonary rehabilitation in relation to pharmacotherapy and telehealth, with estimated ICERs ranging between £1000 and £92 000/QALY.18 While keeping in mind that the pyramid’s interventions were only for already diagnosed COPD patients, a simplistic conclusion would be that the ICER of case-finding is somewhere in the middle. An important note here, however, is that we cannot directly compare this ICER of £16 696/QALY to the absolute amounts stated in the pyramid given these numbers were (1) merely based on single studies performed before 2012, (2) were relative to the standard of care within their own segment of care and (3) did not differentiate between the value of each intervention at different stages of the disease, as the authors pointed out in the more detailed IMProving and Integrating RESpiratory Services in the NHS (IMPRESS) guide to the relative value of COPD interventions.19

Another question that remains is whether active case-finding is ready for large-scale adoption in the UK. As many will be aware, effectiveness and cost-effectiveness are not the only two prerequisites for successful scale-up. Already overburdened primary care physicians may be held back by the relatively high number needed to screen to identify one case of COPD, especially if no proper compensation is in place. Earlier stakeholder exploration showed that limited resources were indeed the main barrier for implementation, but also the value of identifying people with mild disease has been questioned.7 Facilitators for adoption included support from medical specialists, financial incentives and comprehensive guidance. Interestingly, as has been established by the same active Birmingham group that performed the current study, even patients perceived that GPs lacked time.20 Regarding this workload, one way forward here may be involving allied health professionals, such as pharmacists and nurses, in the screening process and refer to GPs/specialists only to make the diagnosis.21 Other, more technological options include advanced techniques, such as machine learning and chest CT imaging.6 CT scans to screen for COPD seem a secondary care option only, but may be cost-effective if integrated with lung cancer and cardiovascular screening as currently being investigated.22

Looking beyond the UK, we may wonder about the generalisability of this study’s findings to other settings? If we take a close look at the risk factors that were included in the screening algorithm, we see that smoking was the only risk factor considered. Even in high-income settings, this may already exclude a fair share of patients at risk. Indeed, in a high-income country, such as Denmark, about 20% of COPD patients are never smokers and their disease burden is equally significant.23 In low-income and middle-income countries (LMIC), low availability of resources and additional risk factors to smoking, such as household and ambient air pollution,24 would require additional radically different approaches. Notably, over 90% of global COPD deaths occur in LMIC and in many places patients have no access to spirometry. Therefore, alternative diagnostic and treatment strategies may be needed. Hopefully, the on-going Global Excellence in COPD outcomes study will shine a light on this unmet need, looking at case-finding in Nepal, Peru and Uganda.25 Especially in poorly resourced countries, cost-effectiveness (and affordability!) is of utmost importance and may be much different from the UK.

All in all, while not fully generalisable and still surrounded by some uncertainty, Lambe et al provide compelling evidence on the cost-effectiveness of active COPD case-finding in the UK. While case-finding is cost-effective, public health interventions focusing on primary prevention could, however, have an even larger clinical and economic impact. COPD’s current value pyramid is providing options for already diagnosed (more severe) COPD patients only. Yet, the largest value may be in targeting those people not having COPD yet, specifically focusing on prevention and detection of early signs of lung damage. In my humble opinion, we may, therefore, have to consider turning COPD’s value pyramid into an iceberg (figure 1) and expanding it with options for non-diagnosed people at risk for COPD.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

COPD’s value pyramid: should it include options for non-diagnosed and people at risk for COPD? COPD, chronic obstructive pulmonary disease; QALY, quality-adjusted life year.