National Survey of Guideline-Compliant COPD Management Among Pneumologists and Primary Care Physicians

Abstract

The aim of this survey was to investigate guideline-compliant COPD management among pneumologists and primary care physicians (PCPs). A multiple-choice questionnaire was sent out to 1836 PCPs and 863 pneumologists in Germany. The questions focused on the key aspects of current national and international COPD guidelines. Four hundred eighty-six PCPs and 359 pneumologists participated in the study. It was found that pneumologists held the GOLD guideline in high regard (60.4%), while PCPs tended to follow the German National COPD guideline (66.5%). Differences were also found with regard to diagnosis and classification of COPD on the basis of spirometric and clinical criteria. The current GOLD classification of moderate and severe COPD was used by 36.2% and 23.4% of the pneumologists, respectively, and by 32.1% and 20.2% of the PCPs. Although PCPs and pneumologists endorsed educational measures to help patients quit smoking, implementation was still inadequate. The two most important therapeutic goals were to improve quality of life and prevent exacerbations. Except for the criteria for the use of steroids and the implementation of pulmonary rehabilitation measures, treatment of COPD based on severity class was largely in compliance with guidelines. However, appreciably more PCPs than pneumologists incorrectly assessed the evidence-based clinical benefits of various therapeutic measures. The study shows that, despite the popularity of COPD guidelines, deficits exist among pneumologists and PCPs with respect to diagnosis and treatment of COPD and practical implementation of educational measures. These deficiencies in guideline conformity might be best addressed through targeted continuing-education measures.

Key words: :

• COPD
• Guidelines
• Disease management
• Primary care.

INTRODUCTION

Chronic Obstructive Pulmonary Disease (COPD) is one of the most prevalent chronic diseases throughout the world. It is a growing cause of morbidity and mortality and is expected to become the third most common cause of death worldwide by 2020 [[2], [4], [16]]. In addition, COPD imposes an immense economic burden resulting from hospital and outpatient care and work incapacity. Nevertheless, the prevalence of COPD is probably underestimated in Germany, as elsewhere [[14]]. This is due, among other reasons, to the fact that the disease often only becomes symptomatic in advanced stages and therefore tends to be diagnosed late. In recent years intensive national and international efforts have been undertaken to establish clinical evidence-based guidelines in order to assist physicians in the early diagnosis and treatment of this disease [[2], [4], [16]].

Although the guidelines are widely available, a current local Swiss study has shown marked deficits in the knowledge and implementation of evidence-based COPD guidelines [[9]]. The spirometric classification and treatment algorithm of the current German COPD-guidelines are based on the GOLD guidelines of 2001 as follows (modified to [[16]]): Stage 0 (at risk): lung function: FEV₁/VC ≥ 70%; FEV₁ ≥ 80% pred., pharmacotherapy: none. Stage 1 (mild): lung function: FEV₁/VC < 70%; FEV₁ ≥ 80% pred., pharmacotherapy: short-acting ß2-agonists and/or anticholinergics as needed. Stage II (moderate): lung function: FEV₁/VC < 70%; FEV₁ < 80% pred; pharmacotherapy: long-acting ß2-agonists and /or anticholinergics, if limited benefit: add oral theophylline; if limited benefit: add inhaled steroids for at least 3 months. Stage III (severe): lung function: FEV₁/VC < 70%; FEV₁ < 30% pred., pharmacotherapy: check if long-term oxygen therapy (LTOT) is indicated.

Investigations into various diseases indicate that specialists possess better knowledge and exhibit better guideline adherence in their field of discipline [[3]]. However, no study data comparing the adherence of primary care physicians (PCPs) and pneumologists with COPD guidelines are currently available.

The aim of this prospective, cross-sectional study was therefore to investigate the knowledge and acceptance of guideline-compliant COPD management both among PCPs and pneumologists and to identify any deficits in this respect. The study results thus provide a first comparative overview of outpatient COPD management as practiced by pneumologists and PCPs.

METHODS

The present study was a prospective cross-sectional survey in the form of an anonymized self-reported questionnaire sent out, together with a cover letter and stamped addressed envelope, to 1836 primary care physicians (general practitioners (GPs) and internists) and 863 pneumologists in private practice. The survey was conducted between mid-April and mid-June 2005. 863 pneumology practices across Germany were addressed. The PCPs who were approached were selected randomly from a register of physicians. A distribution between PCP practices in major cities (> 200,000 inhabitants) and those in the surrounding area was 50:50 for the following regions (major city/state: [[1]] Berlin/Brandenburg, [[2]] Hamburg/Lower Saxony, [[3]] Cologne/North Rhine-Westphalia, [[4]] Frankfurt/ Hesse, [[5]] Leipzig and Halle/Saxony and Saxony-Anhalt, [[6]] Stuttgart/Baden-Wurttemberg, [[7]] Munich/Bavaria. The questionnaire, which contained 40 product-neutral, multiple-choice questions, investigated the following key points relating to current national and international COPD guidelines (German COPD guideline [[16]], GOLD guideline (Update 2004) [[2]]:

1. Epidemiology and diagnosis of COPD (n = 10 questions)

2. Patient education and prevention (n = 3 questions)

3. Treatment (n = 19 questions)

4. Knowledge and acceptance of current COPD guidelines (n = 2 questions).

The final questionnaire section contained questions on the respondent's practice (n = 7 questions).

The study was funded by GlaxoSmithKline (GSK) and carried out in cooperation with the German National Association of Pneumologists (BDP), a professional specialty organization of pneumologists. The survey was approved by the ethics committee of the Bayrische Landesqerzte Kammer. The physicians received a modest honorarium of 50 euros for taking part in the survey.

STATISTICS

The data from the completed anonymized, machine-readable questionnaires were statistically analyzed using SAS. The text and tables present the percentages in each category with respect to the total number of responses. We used χ²-tests or, if indicated, FISHER's exact tests to assess differences between the two groups of physicians. A saturated 2-population repeated measures analysis of variance model was used (SAS PROC CATMOD, SAS version 9.1.3) to investigate the ratings of the benefits of several different therapeutic interventions (as presented in Table 3). Statistical significance was set at P < 0.05 (2-sided test).

Table 3 Assessment of the benefits of various therapeutic measures

	Improves mortality		Reduces exacerbations		Slows Progression		Improves symptoms		Improves quality of life		No effect
Therapy	PNEU	PCP	PNEU	PCP	PNEU	PCP	PNEU	PCP	PNEU	PCP	PNEU	PCP
SABD	5.3%	8.4%	7.2%	17.3%	4.5%	7.6%	95%	91.8%	60.7%	65%	3.9%	3.1%
LABD	19.8%	35%	47.6%	57.4%	26.2%	40.9%	94.7%	86.8%	91.1%	86.4%	0%	1%
ICS	15%	34.4%	90.8%	74.5%	28.7%	51.9%	63.5%	83.1%	60.4%	72.6%	2.5%	1.9%
OCS	9.7%	18.5%	41.2%	47.7%	14.8%	29.2%	73%	78%	50.7%	52.5%	13.1%	9.5%
Oxygen	81.9%	54.9%	10%	17.5%	8.4%	10.5%	68.2%	80%	82.5%	78.4%	1.1%	6%
Reha	32.9%	37.4%	44.3%	51.9%	29%	42%	71.3%	70.4%	88.9%	81.9%	3.3%	4.9%
SMOK	84.1%	80.9%	75.2%	76.5%	91.6%	84.4%	76%	72%	76.9%	75.7%	0.6%	1%
p-Value	0.053		0.01		< 0.0001		0.054		0.90		0.61

Question: What consequences for patients do you associate with the following therapeutic measures? For the assessment questions multiple answers were possible. SABD: short-acting bronchodilators (sympathomimetics and/or anticholinergics), LABD: long-acting bronchodilators (sympathomimetics and/or anticholinergics); ICS: inhaled corticosteroids; OCS: oral corticosteroids, Reha: pulmonary rehabilitation programs, SMOK: smoking cessation. Significant differences between pneumologists (PNEU) and PCP were found for evaluation of the therapeutic impact on reduction of exacerbations and progression of the disease. There was also a trend for group differences in the evaluation of the therapeutic benefits on survival and symptoms. N = 486 PCP, 359 pneumologists.

RESULTS

Demography

Fortyone point six percent (n = 359) of the pneumologists and 26.5% (n = 486) of the PCPs returned the questionnaires. The ratio of general practitioners (GPs) to internists in the latter group was 70:30. The respondents' demographic data are listed in Table 1. The PCPs treated on average 5–20 COPD patients a month, mainly mild to moderate cases. The pneumologists, by contrast, treated on average about 150–200 COPD patients a month, chiefly patients with moderate COPD.

Table 1 Demographic data of the participating physicians

Variable	Pneumologists	Primary care physicians (PCP)
Age
30–40 years	5.6%	14.8%
41–50 years	45.4%	40.1%
51–60 years	34.8%	30.2%
Over 60 years	13.6%	14.0%
Gender
Male/female	81.3%/17.8%	60.7%/38.7%
Practice
Single-physician practice/group practice	64.6%/34.9%	64.6%/33.9%
Location
City (> 200,000 inhabitants)/ surrounding area	46.5%/51.8%	44.2%/54.3%
Length of time in practice
Less than 10 years	38.2%	35.2%
11–20 years	47.6%	45.1%
More than 20 years	13.4%	18.9%

486 PCP and 359 pneumologists participated in the study. Information was unavailable in fewer than 2% of cases.

Epidemiology and risk factors

The questions concerning epidemiology and risk factors showed that 84.2% of the PCPs and 94.7% of the pneumologists believe that COPD is becoming more prevalent (p < 0.0001). Most of the physicians (PCPs 57.8%, pneumologists 51.8% (p = 0.71)) regarded COPD as a relevant public health issue. 42.3% of the pneumologists versus 20.8% the PCPs saw it as a major problem (p < 0.0001). 20.8% of the PCPs did not believe that COPD is a more pressing problem than other public health issues. Nearly all the respondents (PCPs: 92.4%, pneumologists 97.5%) regarded smoking as the foremost risk factor for COPD.

Diagnosis of COPD

Eighty six percent of the PCPs and 94.4% of the pneumologists indicated that suspicion of COPD is warranted in the presence of symptoms such as cough, sputum production and dyspnea and/or existing risk factors (p < 0.0001). 76.5% of the PCPs and 92.7% of the pneumologists regarded pulmonary function tests as the primary diagnostic tool for COPD (p < 0.0001). The pneumologists preferred whole-body plethysmography (54.3%) over spirometry (38.4%). 90.1% of the PCPs had access to a spirometer in their practice.

However, only 39.5% of the PCPs and 53.8% of the pneumologists applied spirometric criteria (Tiffenau index < 70%) to COPD diagnosis. Instead, 49.6% of the PCPs and 38.2% of the pneumologists used the diagnostic criteria for chronic bronchitis (Table 2).

Table 2 Diagnostic criteria for classifying COPD

	Pneumologists	PCP	P-value
In your opinion the diagnosis of COPD is based on:
Cough/sputum production for three months over a two-year period	38.2%	49.6%	0.001
Tiffenau index less than 70%	53.8%	39.5%	< 0.0001
FEV1 less than 1.5 liters	1.7%	3.9%
Tiffenau index less than 90%	2.5%	1.6%
Chest x-ray with signs of hyperinflation	0.3%	1.0%
Blood gas analysis showing hypoxemia and/or hypercapnia	1.4%	3.5%
In your opinion moderate COPD is defined as follows:
Tiffenau index less than 70% and FEV1 less than 80% of normal/**	36.2%	32.1%	0.21
Primarily on the basis of clinical symptoms	8.4%	23.7%	< 0.0001
FEV1 less than 1.0 liters	1.1%	1.2%
Tiffenau index less than 50%	3.3%	6.8%
Tiffenau index less than 70% and FEV1 less than 50% of normal**	49.3%	35.2%	< 0.0001
In your opinion severe COPD is defined as follows:
Tiffenau index less than 70% and FEV1 less than 30% of normal**	56.5%	43.4%	0.0002
Primarily on the basis of clinical symptoms	8.9%	24.5%	< 0.0001
FEV1 less than 1.0 liters	2.8%	2.7%
Tiffenau index less than 50%	7.2%	8.8%
Tiffenau index less than 70% and FEV1 less than 50% of normal	23.4%	20.2%	0.26

Tiffenau index: FEV₁/VC. FEV₁: forced expiratory volume in one second, VC: vital capacity. Only one answer to each of the questions on COPD diagnostic criteria was possible. *GOLD Guideline (2004). **GOLD Guideline (2001) and German Guideline. Moderate COPD was defined as Tiffenau index less than 70% and FEV₁ < 80% and ≥30% predicted. Statistically significant group differences (p ≤ 0.0001) were found for each category of question (basis of diagnosis, definition of moderate and severe COPD). N = 486 PCP, 359 pneumologists.

60–65% of the PCPs and 80–85% of the pneumologists were able to correctly cite the spirometric criteria for classifying the severity of COPD into moderate and severe forms. This value was derived by adding the various spirometric criteria of the current national and international guidelines. Among the PCPs there was a clear tendency to classify severity on the basis of the current German COPD guideline. In terms of the current GOLD guideline, 36.2% and 23.4% of the pneumologists and 32.1% and 20.2% of the PCPs applied the correct spirometric criteria for moderate and severe COPD, respectively. Just under 25% of the PCPs and fewer than 9% of the pneumologists based their classification of moderate and severe COPD solely on the clinical picture. Table 2 shows an overview of the diagnostic approaches to COPD.

Educational measures and prevention

Sixty three percent of the PCPs and 95.5% of the pneumologists (p < 0.0001) reported that the patients' inhalation technique is regularly checked by the physician him/herself or by trained staff. The rest believed it sufficient to instruct the patients if required. Influenza vaccination of patients with moderate or severe COPD was often given routinely by PCPs and therefore might be less frequently given by pneumologists (Figure 2).

Figure 2 PCPs' and pneumologists' treatment for ≥50% of their patients with moderate COPD (Figure 2a) and severe COPD (Figure 2b). OCS: oral corticosteroids, SABD: short-acting bronchodilators (ß2-agonists and/or anticholinergics), LABA: long-acting ß2-agonists; LAMA: long-acting anticholinergics; ICS: inhaled corticosteroids, LABD: long-acting bronchodilators (ß2-agonists and/or anticholinergics), Pulm. Reha: pulmonary rehabilitation, Influenza vacc.: Influenza vaccination. N = 486 PCP (grey columns), 359 pneumologists (black columns). *p < 0.001, ⁺p < 0.05.

Measures to quit smoking were not routinely discussed at every patient visit, especially by PCPs (PCPs: moderate COPD: 47.9%, severe COPD: 58.8%, pneumologists: moderate COPD: 69.4%, severe COPD: 74.4%, p < 0.0001 vs. PCP). In addition, 59.7% of the PCPs and 65.5% of the pneumologists (p = 0.09) were of the opinion that appropriate training measures to help patients quit smoking are difficult to implement. Nevertheless, most of the respondents regarded cessation of smoking as by far the most effective measure to prolong life and slow disease progression (Table 3).

Treatment of COPD

The two physician groups agreed that the two main goals of effective treatment are to improve quality of life and prevent exacerbations (Figure 1).

Figure 1 Question: What are the two main features of COPD therapy to be improved for you personally? With regard to treatment goals, an improvement in clinical parameters, such as quality of life and reduction of exacerbations, was rated higher by the physicians than improvement in pulmonary function. On the other hand, a favorable effect on cough/sputum production and mortality played a relatively minor role. N = 486 PCP (grey columns), 359 pneumologists (black columns). *p < 0.001, ⁺p < 0.005.

As Figure 2 shows, most patients with moderate or severe COPD were prescribed long-acting bronchodilators. Combinations of long-acting ß2-agonists and long-acting anticholinergics were used especially in severe COPD. Short and long-acting bronchodilators (ß2-agonists, anticholinergics) were thought to be beneficial particularly in improving quality of life and symptoms (Table 3).

By comparison, the proportion of patients with moderate or severe COPD who were treated with theophylline was relatively small, but increased with severity. Another frequently used therapy for patients with severe COPD was to give inhaled steroids alone or in combination with inhaled long-acting bronchodilators. Yet in this context 30.4% of the pneumologists and 53.9% of the PCPs did not follow the recommendations for the use of inhaled steroids given in the current GOLD guideline (Table 4). The main reason for the use of inhaled steroids was to improve symptoms. Hence, the chief positive effects of inhaled steroid therapy were viewed in the context of improved symptoms and quality of life and a reduction of the exacerbation rate (Table 3).

Table 4 Criteria for long-term therapy (> 3 months) with inhaled and oral steroids

	Inhaled steroids		Oral steroids
	Pneumologists	PCP	Pneumologists	PCP
COPD with FEV1 < 50% and repeated exacerbations	69.6%*	46.1%	27.6%	31.1%
Good response to inhaled or oral steroids in pulmonary function tests	35.9%	34.2%	18.7%	17.3%
Symptomatic improvement achieved with inhaled steroids	87.2%	81.7%	—	—
Only in cases of deteriorations	5.3%	10.9%	61%	53.7%
Problems with the use of inhaled steroids	—	—	17.8%	22.2%
No long-term use of inhaled or oral steroids	1.1%	0.9%	33.7% +	25.3%

Question: What criteria do you apply when you prescribe long-term treatment with inhaled/oral steroids? Multiple answers were permitted to the questions concerning criteria. From the physicians' point of view, symptomatic improvement is the leading criterion for long-term therapy with inhaled steroids, followed by the current guideline recommendations (*p < 0.0001). Only a subset of the physicians do not use any oral steroids in long-term therapy in compliance with the guidelines, (+p = 0.0078). N = 486 PCP, 359 pneumologists.

Interestingly, long-term therapy (> three months) with oral steroids, especially in severe COPD, still played a certain role. Symptomatic deterioration was cited among the criteria applied to the use of long-term oral steroid therapy. Only 25.3% of the PCPs and 33.7% of the pneumologists did not use oral steroids in long-term treatment. Long-term oxygen therapy (> 15 hours daily) was generally recognized as being beneficial and was believed to be an effective measure for improving quality of life and symptoms. However, the positive effect of long-term oxygen therapy on survival was rated markedly higher by the pneumologists than by the PCPs (p < 0.0001).

Pulmonary rehabilitation measures were carried out only in a small percentage of patients with moderate and severe COPD. The chief benefit of pulmonary rehabilitation was thought to be improved quality of life.

Knowledge and acceptance of COPD guidelines

The recommendations of national and international guidelines on the diagnosis and treatment of COPD were deemed to be very helpful by 48.8% of PCPs and 46.5% of pneumologists and as an orientation aid by 48.8% and 52.1%, respectively. In this context, the German national COPD guidelines served as the standard for PCPs and the GOLD guidelines for pneumologists (Figure 3).

Figure 3 Question: Which national and international guidelines on the diagnosis and treatment of COPD are most relevant for you? This was a single-answer question. Pneumologists (black columns) were more adherent to guidelines than PCP (grey columns) and preferred using the GOLD guidelines whereas PCP tended to apply the German national COPD guideline. The current British NICE guidelines had no relevance for German physicians. N = 486 PCP, 359 pneumologists. *p < 0.0001, +p = 0.001.

Influencing factors

Overall, we found no consistent relationship of physician characteristics (gender, age, years of experience, practice, location) beyond speciality that were associated with better or worse self-reported levels of COPD care in additional descriptive statistical analyses including FISHER's exact tests. Slight variations of around 10% were found only with regard to the criteria for the use of inhaled steroids (positive pulmonary function test with inhaled steroids, city: 28.4%, surrounding area: 38.7%) and in the evaluation of the efficacy of long-acting bronchodilators (reduction of the exacerbation rate: city: 63.7%, surrounding area: 52.4%).

Among the pneumologists individual discrepancies were found in that pneumologists in major cities were more likely to set up group practices within the same discipline (city: 26.9%, surrounding area: 14.1%), were less likely to apply the German national COPD guidelines (city: 26.9%, surrounding area: 39.1%) and tended to evaluate the efficacy of systemic steroids somewhat differently (reduction of the exacerbation rate: city: 48.5%, surrounding area: 34.9%).

DISCUSSION

The aim of evidence-based clinical guidelines is essentially to improve the medical care of patients and assist physicians with regard to appropriate diagnosis, monitoring and treatment. Table 5 lists several basic advantages and drawbacks of such guidelines. It is noteworthy that pneumologists tended to use the GOLD guidelines while PCPs tended to apply the German national COPD guidelines [[2], [16]].

Table 5 Possible advantages and disadvantages of evidence-based clinical guidelines (modified from 5, 7, 15)

Advantages	Disadvantages
Improved medical care	Incorrect, contradictory or outdated guidelines
Quality standards for clinical practice	“Blind” faith in guidelines
Time-saving orientation and self-checking	Poor acceptance (“cookbook medicine,” time-consuming)
Identification of research gaps	Loss of physician's autonomy
Financial savings	Relevance in medical liability cases?
Legal certainty	Unclear individual benefit, e.g., in elderly people with multiple concomitant diseases

The physicians were aware of the increasing epidemiologic and economic significance of COPD, whereby pneumologists were more likely to view COPD as a major public health problem.

By contrast, the question concerning the diagnostic criteria for COPD revealed significant differences in the response pattern, probably due to imprecision of the term COPD. By definition, COPD encompasses chronic obstructive bronchiolitis, obstructive pulmonary emphysema and chronic bronchitis [[2]]. Although the inclusion of chronic, non-obstructive bronchitis (risk group, grade 0) among chronic obstructive pulmonary diseases is correct according to the GOLD classification, it remains controversial, as study data show it to have no predictive value for the later development of an obstructive component [[13]]. The diagnosis of COPD presupposes a repeated, not fully reversible airway obstruction, the severity of which is best determined by measuring the one-second forced expiratory volume (FEV₁) and calculating the FEV₁/VC ratio (Tiffenau index).

Irrespective of the guideline used, deficits exist with regard to the severity classification of COPD on the basis of spirometric criteria. Thus, more than a third of the PCPs and around 20% of the pneumologists applied incorrect pulmonary function criteria to the classification of moderate and severe COPD. It is noteworthy that most of the PCPs and pneumologists still orient themselves to the old GOLD classification of 2001. One reason for this may be that the current German national guideline is still based on the 2001 GOLD severity classification [[6], [16]]. Where national guidelines exist, regular revision on a yearly basis seems important to provide most up-to-date knowledge for both physicians and patients.

The two most important medical goals for successful COPD therapy are to improve quality of life and prevent exacerbations. Long-acting bronchodilators and in severe forms also inhaled steroids are often prescribed for the treatment of moderate and severe COPD. Combinations of long-acting ß2-agonists and anticholinergics are used especially in severe COPD. Theophylline, as a drug of second choice, is used relatively rarely. Pulmonary rehabilitation measures are prescribed sparingly, even in advanced stages of the disease. This means that COPD patients are referred to pulmonary rehabilitation either very late or not at all. Thus, the positive effect of pulmonary rehabilitation on quality of life and symptoms as well as its potential preventive effects (e.g., cessation of smoking, disease progression) remains under-utilized [[12]].

Contrary to the recommendations of the current national and international guidelines, oral steroids are also used between exacerbations in the long-term treatment of severe COPD. Clinical deterioration, severe COPD with recurrent exacerbations and technical problems with the use of inhaled steroids are cited as the rationale for long-term therapy (> 3 months) with oral steroids. However, no study data are available on the beneficial effect of long-term oral steroid therapy on quality of life, disease progression or mortality [[2], [16]]. On the other hand, oral steroid therapy brings about a dose-dependent increase in the rate and severity of side effects (including steroid-induced myopathy, osteoporosis) and mortality [[10]].

Opinions varied widely among the surveyed pneumologists and PCPs regarding the clinical benefit of further therapeutic interventions. Thus, a prognosis-improving effect has in fact been demonstrated for long-term oxygen therapy in advanced COPD with chronic hypoxemia [[2], [8], [16]]. However, no definitive study data are available on the effects of pulmonary rehabilitation, long-acting bronchodilators and/or inhaled steroids on mortality and disease progression, though this question is currently being investigated in long-term, large-scale, prospective, randomized studies with inhaled steroids and/or long-acting bronchodilators [[1], [11]].

In addition to oxygen therapy, successful cessation of smoking still remains the single most effective measure to favorably influence disease progression and life expectancy. This fact is known to all the physicians, who advocate training measures to help patients quit smoking but rarely implement them in practice. Another aspect relates to the teaching and monitoring of patients in the correct inhalation technique. This is essential for effective treatment and is not carried out by nearly a third of the PCPs.

In this survey, the guideline compliance of PCPs was better than that found in a recently published study in Switzerland [[9]]. With regard to pneumologists, this study is the first to investigate COPD guideline adherence. Interestingly, no relevant differences were found in the response patterns according to practice location (major city vs. surrounding area). As with all surveys, the results of this study can be neither validly checked nor generalized. Moreover, it is unclear to what degree a positive selection bias was present and to what extent the respondents' answers differed from their actual practice and from those of non-respondents. Another limitation of the study is that no chart audit on a subset of responding physicians in each speciality was done to assure that self-reported care of COPD accurately reflects the care actually delivered to patients.

The varied response rates of physicians might suggest that responders are more concerned about COPD care and that the survey underestimates the misknowledge of current guidelines. Accordingly, our data may represent a best-case scenario, because survey respondents may generally be aware of and adhere to COPD guidelines.

The results of the study reveal deficits in outpatient COPD management by PCPs and pneumologists. Pneumologists in general were more knowledgeable about COPD diagnosis, severity classification and therapy benefits and were more likely to check their patients' inhalation technique. However, even among pneumologists deficits were found regarding the spirometric classification of severity based on the GOLD guidelines and the indications for steroids. Measured on the basis of the guideline recommendations, the practical implementation of training measures to help patients quit smoking and the use of pulmonary rehabilitation measures still remains inadequate among both PCPs and pneumologists. These deviations from guideline recommendations might be best addressed through targeted information and continuing-education measures. In addition, it is essential that the content of current COPD guidelines be implemented more consistently into all physician groups concerned.

We wish to thank Nicole Huep, Sandra Mahn and Monika Lengfelder, GSK, for their technical support.