Statistics from Altmetric.com
The BTS guidelines on COPD1 exhort GPs to diagnose COPD early, preferably in the presymtpomatic stage. In effect what is being promoted is “secondary prevention” (screening) through the use of spirometry. In most early cases this would involve two readings at an interval of four to five years. The intended outcome is that the patient stops smoking.
At first sight this appears a sensible proposal (although not one included in the ATS2 or ERS3 guidelines). However, the same outcome can be achieved without the screening process, so before GPs adopt this recommendation wholesale it is worth noting that there is no direct evidence to support screening.
Literature abounds on the costs of persuading smokers to stop smoking. If by screening for COPD one could achieve a lower cost per quitter amongst susceptible smokers, then an argument could be made on cost effective grounds. But as simple advice from a GP is one of the most cost effective interventions in health care, this is unlikely.4
One is therefore left to consider the incremental stop rate achieved by superimposing screening and advice on a programme of advice alone. Most smokers are already aware of the risk of lung cancer and ischaemic heart disease. It is not known what difference the distant threat of COPD makes to smoker quit rates.
If the costs of screening are distributed solely to the increment of true positives who stop with the enhanced programme one arrives at a true marginal cost of this preventative measure. The costs to be considered should include the direct health service costs of equipment and training, and the opportunity cost to society of the time spent by GPs and nurses.
Screening is intended to improve well being (the ethic of “maximising public welfare”), yet there is the paradox of identifying disease (and therefore reducing perceived well being) in pre-symptomatic individuals when the majority derive no benefit. Intangible costs become relevant as do indirect costs (with well people often having to miss work for screening). Intangibles include the anxiety created by screening and the even greater anxiety in the true positives who then fail to stop smoking. One also has to decide what to tell the true negatives (those shown not to be at risk of COPD).
A broader cost-utility analysis would permit comparisons with the benefits of other screening interventions in primary care. Cervical cytology, mammography, and newer technologies such as colorectal cancer screening compete for development monies.
Lieberman offers a model for measuring cost effectiveness of colorectal cancer screening programmes.5 The sensitivity analyses considered are patient compliance, varying costs of procedures, frequency of surveillance, costs of downstream care, cancer detection rate, and cancer prevention rate. The parallels for COPD screening might include patient response rate to invitation, varying costs of equipment and staff time, frequency of surveillance, costs of treating diagnosed COPD, COPD detection rate, and smoker quit rates. A full analysis might also consider the discount rate for costs and benefits over time. In a programme that detects disease 10 years before it becomes symptomatic, an accepted discount rate of 6% per annum compounds considerably.
Spirometry is invaluable in the diagnosis and management of COPD. GPs should welcome the guidelines but must consider the opportunity costs to their activities before embarking on screening pre-symptomatic patients.
authors’ reply With regard to the points raised by Dr Perry we would comment that there may be little direct evidence about screening for early COPD, but the indirect evidence was sufficient to persuade the guidelines committee. The natural history of COPD, elegantly described by Fletcher and Peto,1-1 is that a proportion of cigarette smokers (15–20%) have an annual loss of FEV1 that is larger than for non-smokers. Initially such losses, while measurable in large group studies, are too small to be detected in the individual. After 20–40 years the cumulative excess loss of FEV1 is large enough for an individual’s FEV1 to be below the lower limit of the predicted normal value and thus be detectable. Thereafter, with continuing smoking, the subject will progress from mild impairment through moderate to severe impairment. The only proven treatment that slows this inevitable decline in function is stopping smoking.1-1 1-2
Most patients only present to the health service with moderate to severe disease at a stage when jobs are threatened or lost, and lifestyle limited. Since the damage is irreparable, it is logical to consider prevention at an earlier stage and preferably before symptoms develop.
Dr Perry is correct to point out that cost effectiveness studies should consider the wider picture, but it is important not to extrapolate from studies based on different principles. Simple advice from a GP is effective in a very small proportion of subjects (2–5%) and is cost effective only because it is so cheap. The health benefit of adding nicotine patches and nurse counsellor advice,1-3 which increases the quit rate significantly (21%), are both clear and desirable, but the extra treatment costs have left doubt about the cost effectiveness of such additional work. However, studies in unselected smokers will include those not susceptible to COPD, cardiac disease, or lung cancer (perhaps half of all smokers).1-4 The extra costs of treating all smokers (30% of the population) when only some can benefit makes the cost effectiveness equation less favourable. It also presupposes that non-symptomatic patients will agree to come and see their GP. Targeting smokers in their 50s with a reduced FEV1 (who amount to less than 1% of the population) would result in a very different calculation. Now only 15% of smokers in a defined age range are being targeted for non-smoking advice/support which reduces the cost implication of the more effective treatment and, moreover, the health benefit is greater because every person who gives up smoking is being prevented from developing symptomatic COPD, with its costs to both the individual and the health service. A formal study should be done to confirm such estimates, but progress cannot always wait for absolute confirmation, particularly when the health benefits to this defined subgroup are so clear cut.
Dr Perry is concerned at the adverse effects of screening when many will derive no benefit. Those with a normal FEV1 should still be advised to quit on the grounds of the risk from heart disease and lung cancer, but on cost effectiveness grounds they may not qualify for more intensive help. Their anxiety levels are unlikely to be raised by this any more than by the frequent publicity about smoking in the media.
Spirometry is not an expensive procedure, costing less than a chest radiograph and probably less than a fasting lipid measurement. Radiographs are of limited value in COPD yet are often requested, whereas measurement of the FEV1 informs diagnosis, treatment and prognosis and has been grossly underused in both hospital and general practice. Finally, it should be pointed out that a recent paper in the BMJ describing the prospective Renfrewshire population1-5 showed that FEV1 was a stronger predictor of premature death than serum cholesterol measurements and as strong a predictor as cholesterol when only heart disease was considered. It concluded with the recommendation that anti-smoking activity be targeted on those with a low FEV1.