|Home | About | Journals | Submit | Contact Us | Français|
To investigate the effect of a primary care model for COPD on process of care and patient outcome.
Controlled study with delayed intervention in control group.
The GP delegates tasks to a COPD support service (CSS) and a practice nurse. The CSS offers logistic support to the practice through a patient register and recall system for annual history-taking and lung function measurement. It also forms the link with the chest physician for diagnostic and therapeutic advice. The practice nurse's most important tasks are education and counselling.
A total of 44 practices (n =22 for intervention and n =22 for control group) and 260 of their patients ≥40 years with obstructive lung diseases.
Within the intervention group planned visits increased from 16% to 44% and from 19% to 25% in the control condition (difference between groups p =0.014). Annual lung function measurement rose from 17% to 67% in the intervention and from 11% to 18% in the control group (difference between groups p =0.001). Compared with control, more but not statistically significant smokers received periodic advice to quit smoking (p =0.16). At baseline 41% of the intervention group were using their inhalers correctly and this increased to 54% after two years; it decreased in the control group from 47 to 29% (difference between groups p =0.002). The percentage of patients without exacerbation did not change significantly compared with the control condition. The percentage of the intervention group not needing emergency medication rose from 79% to 84% but decreased in the controls from 81 to 76% (difference between groups p =0.08).
Combining different disciplines in one model has a positive effect on compliance with recommendations for monitoring patients, and improves the care process and some patient outcomes.
By far the majority of patients with COPD are treated in general practice. To that end, international and national guidelines have been developed [1–4]. Recommendations include using lung function measurement in diagnostics and monitoring, checking symptoms periodically and if necessary adjusting medication, and educating patients to take responsibility for the daily management of their disease. However, care is often still not being given according to the guidelines , . In particular lung function measurements , periodic check-ups , and supervision of inhalation techniques  have proved to be difficult. To implement guidelines, care must meet stringent requirements and should involve various disciplines, from non-medical staff to medical specialists. This means new disciplines are needed in general practice and/or specific disciplines outside general practice should be called in for task delegation and consultation . Various interventions have been tested over the last few years. It appears that a recall system can be successfully organized on a scale that goes beyond the individual general practice and can be run by someone without a medical background . Nurses can fulfil a key role in clinics by providing patient education and counselling [12–15]. Delegation of medical tasks and the presence of a practice nurse are significant associated with spirometry utilization . A combination of spirometry and smoking cessation advice by a nurse appears to increase smoking cessation rates in smokers with mild COPD . According to Bodenheimer et al.  the possibility of consultation with a specialist without a full referral contributes fundamentally to the care of people with a chronic disease.
We have developed a care model that fits into such an approach to chronic disease management, one in which various disciplines (general practitioner – (GP), practice nurse, logistic COPD Support Service (CSS) and chest physician (see Figure 1)) are integrated. All these participating elements were financed by the regular health insurance system. The GP can delegate tasks to the nurse and the CSS. The chest physician can support the diagnostic and therapeutic decision-making without actually seeing the patient. We assumed that the management of chronic obstructive lung diseases could be improved by the introduction of this model. The model was focused on the≥40 age group with chronic obstructive lung diseases covering asthma as well as COPD, as the difference between COPD and asthma is not always clearly established in the primary care population. We performed a study to test the effects of the care model for COPD and/or asthma on the process of care and on patient outcomes.
The effect of the care model on COPD and/or asthma was examined in a controlled study with delayed intervention in the control group. With an interval of two years, pre-test and post-test measurements were performed. The care model was introduced in a region in the south of the Netherlands in 2002 with a running-in period of over a year. During that period the practice nurses were trained and the CCS (a logistic support service, linked to the regional primary care laboratory with a specialized lung nurse and some administrative staff) was set up. GPs qualified for a practice nurse if they had working space for a nurse and an electronic patient register. Due to limited funding, not all interested practices could start at the same time. Based on regional distribution criteria (division between sub-regions and between urban–rural) the first cohort of practices was selected to start with the care model in 2002. These practices formed the intervention group (n=22); the practices on the waiting list formed the control group (n=22). By October 2003, 11 practice nurses had been appointed. Before the nurse started, a random sample of patients per practice was drawn: patients≥40 years with a documented lung condition and using inhalation medicines (137 in the intervention group and 123 in the control group).
In each practice, the nurse (without knowing who had been selected in the random sample) made a survey of all patients≥40 years with chronic lung obstruction on the basis of diagnostic data and medication use. The CSS called up all these patients for extensive history-taking and lung function measurement. The results were sent to a chest physician for assessment, diagnosis (or confirming or adjusting an earlier diagnosis), and advice on treatment. The CSS maintained a register of patients qualifying for annual history-taking and lung function measurement. Patients visited their GP to discuss the results and determine whether the medication was still adequate. They also visited the practice nurse who checked their inhalation technique, and gave education and counselling (smoking cessation). Patients who, according to the GP, met the criteria for referral were referred to the chest physician.
In the control group the patients received the usual care, which generally meant that they were seen only when they consulted the GP about their symptoms.
To study whether patients received care according to the guidelines, we collected data on planned consultations, periodic lung function measurements, and smoking cessation advice. To measure the effect of the care model on patient outcomes, we collected data on smoking status, inhalation technique, exacerbations, and emergency medication. Furthermore, some general characteristics from GPs and patients were noted.
In more detail:
All data were collected by means of questionnaires completed by the GP and the patients , except inhalation technique and emergency medication. Inhalation technique was checked pre-test and post-test by the same laboratory assistant. Data on emergency medication were obtained from community pharmacists.
As primary outcome variable we chose the correct use by the patient of the inhaler because it has been shown that change in organization of care integrating non-physicians can influence this patient outcome . An earlier study showed that 60% of COPD patients used their inhalers correctly . With the intervention, we expected an increase of 20%. Based on an alpha of 5% and a beta of 80% a random sample survey of 39 general practices with 5 patients each was needed (195 patients), taking into account clustering of patients per GP. With an expected dropout of 10%, the total number of patients needed was 215.
Differences between the intervention and control group were tested with mixed logic model repeated measures (Glimmix procedure SAS V8.2). In all tests, corrections were made for the random/cluster effect caused by patient and GP.
The 44 general practices approached were all included; there were no dropouts. At baseline the practices in the control group were comparable with those in the intervention group regarding number of GPs, population size, and average number of asthma/COPD patients (Table II). None of the practices had an active recall system offering planned care to COPD and/or asthma patients. The intervention practices were more often located in a city (>80000 inhabitants) and were less often single-handed than the controls.
Data of 260 patients were collected at baseline. The intervention group was comparable with the controls regarding age, gender, and dyspnoea score (Table II). During the intervention 94 patients (37% of the intervention and 36% of the control group) dropped out (stopped inhalation therapy, moved, died, were referred to a chest physician, or no longer wanted to participate, Figure 2), resulting in 87 in the intervention and 79 in the control group for the final measurement. Analysis showed that patients who dropped out did not differ in gender, age, and dyspnoea score from patients who underwent a second measurement. That is why we included all the patients in the effect measurement (with more patients in the pre-test than in the post-test).
After two years of intervention, the percentage of patients included in planned care at the general practice rose from 16% to 44% (Table III). The difference in change compared with the control group is statistically significant. The percentage undergoing periodical lung function measurements rose from 17% to 67% (to 75% if only patients involved in planned care are counted). All smokers were periodically advised to quit.
The percentage of non-smokers rose 11% in the intervention and 4% in the control group; the difference in change was not statistically significant. Regarding inhalation technique, the percentage of patients handling their inhalers correctly rose from 41% to 54% in the intervention group while it decreased in the control group from 47% to 29%. The percentage of patients without exacerbation in the previous three months rose from 79% to 81% in the intervention group; it decreased in the controls from 77% to 69%. The difference in change was not statistically significant.
We also noticed a difference in change regarding emergency medication. The percentage of patients not needing emergency medication rose from 79% to 84% in the intervention group, while it decreased in the control group; this difference was not statistically significant.
A model for integrated primary COPD and/or asthma management was evaluated in this study. The model proved to have a positive effect on planned care and periodic lung function measurement. A positive effect on patient outcomes was also found. The percentage of patients who used their inhalers correctly rose, but the gain in preventing deterioration was even greater, as seen in the control group where the inhalation technique was not checked periodically (we hypothesized that if you assess inhalation technique at a random moment in a cross-sectional population, you also include people who have recently started on medication and have received instructions on how to use the inhaler and have a good technique. If you follow that same population, few people appear to retain the good technique.) The data on patients in our study correspond to those of another Dutch study GPs (70% with a dyspnoea score≤2) . At baseline we found fewer patients with a correct inhalation technique than in a comparable study among a Dutch population (45% vs. 72% with the correct technique). The high score there may have been due to extra attention to inhalation technique in a previous study by the same researchers, as they suggest themselves . The difference between the intervention and control group in our study may be substantial but still half of all users do not handle their inhalers correctly, meaning it is unclear whether they inhale the correct dose of medication. Further studies are thus needed to find out whether this can be improved by shortening the intervals between inhaler checks – for example, a check at every prescription renewal, because research shows that mistakes occur shortly after the instructions are given, arguing in favour of short cyclic check . The treatment of acute COPD exacerbations is shifting to general practitioners . Patient recognition of exacerbations and prompt treatment improves exacerbation recovery, reduces risks of hospitalization, and is associated with a better health-related quality of life . From this perspective, the question is whether we should have expected less or maybe even more emergency drug use as positive effect of the care model. The decrease we found is not significantly different from the control group, but we believe it is a positive effect because we also saw a decrease in self-reported exacerbations.
The number of patients willing to take part in the study was relatively low (intervention: 68%, control group: 65%) and the dropout rate was very high. This can be considered a weakness of the study (for data collection patients had to visit a laboratory twice to check the inhalation technique and fill in the questionnaire) but not a weakness of the care model. In fact 88% of all patients being treated by the GP were included in the care model .
Although the number of patients with both measurements was lower than the calculated number needed in the power analyses, we do not think that our study is under-powered. In the repeated measurement analysis (PROC MIXED, SAS) all patients are included. This means that data on patients with only one measurement were also analysed. We did not study the cost-effectiveness of the model, but we would like to make a few points here. A great deal of the efforts (and thus also the costs) in the intervention group were put into surveying the target group. These efforts will always be needed if the GP is going to provide planned care for patients with asthma or COPD, and therefore should not be accounted to this specific model. The same applies to setting up the call-up system. On the other hand, paper consultations by chest physicians are model-specific. Consultation in this way is cheap, has proved to be valid , and increases the number of patients who can be treated in primary care.
We conclude that this study has shown that combining various disciplines in an integrated model as described here improves care processes and patient outcomes in primary care for COPD and/or asthma. The care model is especially interesting in those settings in which chronic disease management is general practice based.
This study was funded by the Netherlands Asthma Foundation and the Health Insurance Companies CZ and VGZ.