Practice patterns in the management of chronic obstructive pulmonary disease in primary practice : The CAGE study

1Respiratory Epidemiology and Clinical Research Unit, Montreal Chest Institute, McGill University Health Centre, Montreal, Quebec; 2Department of Medicine, McMaster University; 3FigP Software Inc; 4Centre for Evaluation of Medicines, St Joseph’s Healthcare, Hamilton; 5University of Toronto, Toronto, Ontario; 6St Joseph’s Hospital, La Malbaie, Quebec; 7York Central Hospital, Richmond Hill, Ontario; 8Queen Elizabeth II Health Sciences Centre, Dalhousie University, Halifax, Nova Scotia; 9Centre hospitalier affilié universitaire de Québec, Quebec, Quebec; 10Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton; 11Boehringer Ingelheim Canada Ltd, Burlington, Ontario; 12Pfizer Canada Inc, Kirkland, Quebec Correspondence: Dr Jean Bourbeau, Montreal Chest Institute, 3650 rue St Urbain, Montreal, Quebec H2X 2P4. Telephone 514-934-1934 ext 32185, fax 514-843-2083, e-mail jean.bourbeau@mcgill.ca J Bourbeau, RJ Sebaldt, A Day, et al. Practice patterns in the management of chronic obstructive pulmonary disease in primary practice: The CAGE study. Can Respir J 2008;15(1):13-19.

Practice patterns in the management of chronic obstructive pulmonary disease in primary practice: The CAGE study Les modes de pratique pour la prise en charge de la maladie pulmonaire obstructive chronique en première ligne : L'étude CAGE

ORIGINAL ARTICLE
B etween 20% and 25% of Canadians smoke cigarettes regu- larly, and among current or former smokers, chronic obstructive pulmonary disease (COPD) is the most common cause of significant morbidity (1).A principal symptom of COPD is dyspnea, leading to a reduced capacity for physical activity.Recurrent acute exacerbations involving respiratory tract infections are frequent, often leading to hospital emergency room visits and admissions (2).An impaired quality of life generally results.
COPD is often unsuspected and undiagnosed (3,4).The 2003 Canadian Thoracic Society (CTS) guidelines (2) attempted to simplify and improve both the diagnosis and management of COPD.The primary treatment goals for patients with symptomatic COPD are improvement of dyspnea and exercise tolerance, and reduction in frequency and severity of exacerbations.Achieving these goals has a positive impact on the patient's quality of life and use of health services.The COPD guidelines (2) were one of the first to include pharmacological treatment recommendations according to disease severity (asymptomatic, mild, moderate and severe) derived from the Medical Research Council (MRC) dyspnea scale (grades 1 to 5).Pharmacological treatment, especially with long-acting bronchodilators (LABDs) (5-7), and pulmonary rehabilitation with exercise training (8) are most effective in improving symptoms of dyspnea, exercise tolerance and overall health status, while only smoking cessation has been proven to prevent disease progression (9).Vaccination, LABDs and inhaled corticosteroids (ICSs) reduce acute exacerbations or respiratory complications (7,10).The majority of COPD patients (80% in Canada) receive care in primary practice (3).While the impact of COPD can be reduced with appropriate management (2), little is known about current practice patterns for COPD patients in primary care.Two recent Canadian surveys of primary care (3,11) revealed some deficiencies in COPD prescribing patterns, access to treatment and diagnosis.The aim of the present prospective study was to evaluate the pharmacological treatment, diagnosis and nonpharmacological management of COPD in primary care practice in Quebec and Ontario.

Study design
The objective of the COPD Care Gap Evaluation (CAGE) study was to document current COPD management in routine primary clinical care and to examine the appropriateness of care in relation to the 2003 CTS evidence-based guidelines.The present prospective, cross-sectional study was divided into sequential phases to allow implementation of physician feedback and educational interventions after each phase.This report presents results for baseline evaluation of COPD care from the first phase.There was no patient follow-up after enrolment, because patient care, rather than patient outcome, was assessed.Primary care physicians from Quebec and Ontario recruited up to eight consecutive COPD patients in each of four sequential eight-week phases.The study was approved by an independent ethics review board.

Physician recruitment
Participating physicians were recruited from all primary care physicians in each province who prescribed respiratory medications (ie, bronchodilators, ICSs or combinations) and whose practice locations were community-based.

Patient enrolment
Each participating physician enrolled successive patients who satisfied selection criteria.Included patients all had a smoking history of 10 pack-years or more, were at least 45 years of age, were seen in the physician's office for routine care, and had a working diagnosis of symptomatic COPD either by clinical judgment or by spirometric evaluation.Patients also had to receive a prescription (new or renewal) for COPD during the enrolment visit and provide informed consent.Patients were excluded if they had had a diagnosis of asthma before the age of 40 years; if they had an acute exacerbation while visiting their physician that was severe enough to require emergency referral, hospitalization or prescription of prednisone; or if they were part of a therapeutic trial.

Data collection
Case report forms were completed by physicians during patient interviews.The data collected were designed to capture individual patient factors that allowed assessment of the appropriateness of care without imposing any guideline or scoring system on the physicians.Collected data included patient demographics, smoking history, COPD diagnosis and history, relevant comorbidities, COPD medications (what patients were receiving for treatment when they left the clinic visit), nonpharmacological care provided within 12 months and current adverse events associated with therapy.Current COPD status, including disease severity, was assessed formally, according to the CTS classification (2) based on the MRC dyspnea scale, and informally, by physician and by patient.A separate survey collected physician characteristics.Data were transferred anonymously to a data centre.

Data analysis
Physicians' reported data were analyzed for consistency with the recommended care for levels of disease severity and stability as defined by the 2003 CTS guidelines (Table 1) (2).All patients were classified into four groups of disease severity by symptoms according to the CTS classification based on the MRC dyspnea scale: asymptomatic (MRC grade 1), mild (MRC grade 2), moderate (MRC grade 3 or 4) or severe (MRC grade 5).Patients were also separately grouped by disease stability (on day of visit) according to whether their visit was principally for an acute exacerbation (unstable) or for any other reason (stable).
Distributions of disease severity were cross-tabulated according to four categories (asymptomatic, mild, moderate and severe) to describe the agreement between the CTS classification based on the MRC dyspnea scale and the informal physician assessment of severity.A weighted kappa coefficient was calculated to assess the agreement between these two measures.A similar analysis was performed comparing the CTS classifications based on the MRC dyspnea scale with the patients' self-assessment of disease severity.Finally, differences in the appropriateness of pharmacological treatment were assessed between patients with and without a spirometric confirmation of their COPD diagnosis.Analyses were performed in SAS version 9.1 (SAS Institute Inc, USA).
Pharmacological treatment data were assessed with respect to disease severity and stability (Table 1), and patients in the formal 'asymptomatic' group, for whom no recommendations exist, were excluded.Diagnosis and nonpharmacological management (including smoking cessation counselling, influenza vaccination and referral to pulmonary rehabilitation) were evaluated by groups of disease severity irrespective of disease stability on the day of visit.These two groupings reflect the structure of the recommendations.Results are reported for the whole patient group, and separately for Quebec and Ontario.

Physician characteristics
Of 4954 primary care physicians invited to participate in the study, 185 (3.7%) completed a brief training in the study procedures and 161 (3.2%; 44 in Quebec, 117 in Ontario) enrolled at least one patient.A survey of 119 participants (74%) indicated that the physicians had graduated an average of 23 years previously (Table 2).Physicians in Quebec reported seeing more COPD patients annually and were less likely to be in solo practice than physicians in Ontario.

Patient characteristics
Data from 1155 patients (331 in Quebec, 824 in Ontario) were collected during phase 1 (baseline), which ran from May 2005 to July 2005.Of these, 65 patients were excluded due to missing or implausible data and 1090 patients were included in the analyses (Figure 1).
Patients were principally male (60%).On average, they smoked for 40.1 years, and 40% were currently smoking.Eighty-two per cent had one or more comorbidity (Table 3), and 73% had had COPD for longer than five years.Most patients had stable COPD (91%) on the date of their visit as suggested by the CTS classification: 14% were asymptomatic (MRC grade 1), 32% were mild (MRC grade 2), 41% were moderate (MRC grade 3 or 4) and 4% were severe (MRC grade 5).This grouping was used for analyses of pharmacological treatment.For analyses of diagnosis and nonpharmacological management, patients were classified as asymptomatic (14%), mild (36%), moderate (46%) or severe (4%) regardless of current stability (Figure 1).
Physicians' informal classifications of severity on the day of the patients' visits were mild in 33%, moderate in 52% and severe in 15%, while patients' corresponding self-classifications were 45%, 44% and 11%, respectively.Because a diagnosis of symptomatic COPD was an inclusion criteria, all patients classified as asymptomatic from the MRC dyspnea scale (14%) were considered as misclassified by the physicians' and patients' assessments.Physicians' informal assessment of severity coincided with the CTS classification based on the MRC dyspnea scale in 47% of patients; the corresponding figure for patients' self-assessment was 48%.While there was some agreement (weighted kappa = 0.28, 95% CI 0.24 to 0.32; P<0.0001) between the physicians' and CTS' classifications of severity, the level of agreement is only classified as 'fair' on the Landis scale (12).With a weighted kappa coefficient of 0.27, the level of agreement between patients' self-assessment and CTS classifications is similar.Distributions of assessments of disease severity indicate that 43% of patients were rated more severely by their physicians than by the CTS classification and 36% of patients rated themselves more severely than the CTS classification.
Comparison of physician-reported prescriptions with the 2003 CTS guidelines revealed that 34% of COPD patients with mild, moderate, severe and unstable COPD (as classified by the MRC dyspnea scale) received recommended pharmacological treatment (Figure 2).Physicians prescribed recommended care most often to patients with unstable COPD (55%), followed by those with moderate and stable COPD (45%), severe and stable COPD (42%), and mild and stable COPD (14%).
Recommended pharmacological treatment was observed in 39% of patients with a documented spirometric confirmation of their COPD diagnosis, compared with 32% of those without spirometric confirmation.

Diagnosis and nonpharmacological management
Diagnosis and nonpharmacological management were assessed in patients grouped by disease severity, regardless of current stability.Spirometry had been performed at some point in their history in 56% of patients overall (48% in Quebec, 59% in Ontario) (Figure 3), as recommended.It was used more frequently among patients with severe COPD (75%) than among those with moderate (58%) and mild (53%) COPD.Spirometry had been performed in 79%, 61% and 52% of patients informally rated by their physicians as severe, moderate and mild, respectively.
More than 90% of currently smoking patients were counselled on smoking cessation (Figure 4).Annual influenza vaccination was provided to 80% of patients.Referral to pulmonary rehabilitation was offered to only 9% of patients with moderate or severe disease, for whom it is recommended.

DISCUSSION
The present study demonstrates that a substantial care gap exists relative to recommended practices in the primary care management of COPD patients in Canada's two largest provinces.While a majority of patients had received the recommended smoking cessation counselling (95%) and influenza   (13).
The Canadian component of the international Confronting COPD study, a survey of 401 patients (80% with mild to moderate COPD) (3,14) performed approximately four years before the CAGE study, reported that 37% of patients received ICSs, closely followed by SABDs (33%) and LABAs (9%), suggesting an overuse of ICSs.These figures are lower than those reported in the CAGE study, (67%, 81% and 74%, respectively) and may suggest a discrepancy between patientreported and physician-reported (CAGE) prescribed medications or may reflect temporal differences between the eras of the two studies.Underuse of efficacious therapies in high-risk patients has been demonstrated in the CAGE study, as has ICS overuse in mild disease.The absence of a gradient between disease severity and use of LABDs may reflect a lack of recognition of the value of these drugs in COPD management in clinical practice.This is reinforced by the fact that physicians tended to overestimate disease severity compared with the CTS classification based on the MRC dyspnea scale.On the other hand, the substantial use of ICSs suggests that their perceived benefits in COPD are exaggerated and that they are used irrespective of a diagnosis of asthma or COPD.Undertreated patients with severe COPD may risk losing their autonomy and becoming hospitalized, while unnecessary use of drugs is not risk free.In terms of public health impact, both misuses represent missed opportunities.If they can be resolved, there will be improvements in the clinical outcomes of COPD patients in Canada.
The CAGE study has revealed that only 56% of patients had undergone spirometry, while the 2003 and 2007 CTS guidelines (2,13) recommend it to confirm a diagnosis of COPD.A postbronchodilator forced expiratory volume in 1 s of less than 80% of the predicted value and forced expiratory volume in 1 s/forced vital capacity less than 0.70 are both required for COPD to be diagnosed (2).Misdiagnosis of COPD may prevent COPD patients from receiving therapy that can reduce symptoms and future risk, while causing others to receive unnecessary treatment.A recent Canadian survey, the COPD national report card (11), found that access to spirometry was relatively high nationally (74%), but few family physicians were using it (35%) or considered themselves able to interpret the results (19%).
The present study suggests that nonpharmacological management, more specifically smoking cessation counselling and influenza vaccination, was provided to almost all patients.This is an improvement over a 2003 Canadian patient survey performed four years earlier (3).However, only 9% of moderate and severe patients in our study were referred for pulmonary rehabilitation -known to be of great value for patients in whom respiratory symptoms are associated with reduced functional capacity.Currently, there are only 24 pulmonary rehabilitation facilities in Ontario and 10 in Quebec, and it is estimated that 1.2% of the Canadian COPD population is being managed by them (11).
Several differences were observed between the provinces in the present study, which may reflect attitudes toward diagnosis, treatment or access to care.The use of LABDs in all groups of patients was much lower in Quebec than Ontario, likely reflecting the restricted listing of LABDs in Quebec during the study.The present study also suggests that the diagnostic evaluation of COPD by spirometry is somewhat less common in Quebec than Ontario, consistent with the use of spirometry reported in the COPD national report card (11).Variations between the provinces may be related to access to spirometry, because fewer physicians in Quebec reported access to this diagnostic tool in the report card (11).
Limitations of the present study include the participation of a self-selected group of family physicians (3% of those contacted) with large COPD practices, physician self-reporting and absence of randomization of sequential patient selection by physicians.However, these features would likely have biased the study toward an underestimation of treatment gaps.Guidelines were used as a reference standard, but these may not always be applicable to individual patients.Adherence to treatment guidelines was evaluated based on CTS classification of disease severity -on which the COPD guidelines are based but which may not necessarily correspond to informal physician classification of disease severity.Finally, the study was not designed to assess reasons for differences between recommended and current practices.
The present study indicates that a minority of COPD patients seen in primary care in Quebec and Ontario receive currently recommended pharmacological treatment and referral to pulmonary rehabilitation; approximately one-half of the patients do not have their diagnosis confirmed by spirometry, while a majority receive recommended smoking cessation counselling and influenza vaccination.Analyses of follow-up phases of the CAGE study will provide some insights into the impact of educational interventions on physician self-reported care of COPD patients.Necessary changes in our health care system must include better ways to transfer new knowledge to clinical practice and to give access to interventions proven to maintain health.If these gaps can be closed, we will gain improvements in health care for our patients, as well as reduce hospitalizations.

ACKNOWLEDGEMENTS:
The authors wish to thank all 161 primary care physicians who participated in this study.This study was made possible by a grant from Pfizer Canada Inc and Boehringer Ingelheim Canada Ltd.This article was prepared with the assistance of BioMedCom Consultants Inc, Montréal, Canada.
COPD care in primary practice: The CAGE study Can Respir J Vol 15 No 1 January/February 2008 15

TABLE 1
Recommended care for chronic obstructive pulmonary disease patients* Patients were classified based on the 2003 Canadian Thoracic Society chronic obstructive pulmonary disease guidelines (2); † Mild disease is defined as patients with grade 2 dyspnea on the Medical Research Council scale, moderate disease as patients with grade 3 or 4, and severe disease as patients with grade 5; ‡ A prescription for prednisone, while appropriate, constituted an exacerbation severe enough to exclude patients from the present study.ICS Inhaled corticosteroid; LAAC Longacting anticholinergic; LABA Long-acting beta 2 -agonist; LABD Long-acting bronchodilator; SABA Short-acting beta 2 -agonist; SABD Short-acting bronchodilator

TABLE 2
Characteristics of surveyed physicians by province* Figure 1) Flow of patients included at various stages of the analyses.COPD Chronic obstructive pulmonary disease; MRC Medical Research Council *Based on a subset of participating physicians (n=119 of 161).COPD Chronic obstructive pulmonary disease bronchodilators (SABDs), in agreement with the 2003 CTS guidelines.A majority of mild, stable COPD patients (63%) received an ICS, although it is not recommended for this severity group (Table 4).A single LABD was prescribed for 39% of moderate, stable COPD patients, in accordance with 2003 CTS Bourbeau et al Can Respir J Vol 15 No 1 January/February 2008 16