Interventions to improve the management of asthma in primary care settings
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
The objectives of this review are to determine the effectiveness of various interventions to improve the management of asthma in primary care settings.
Background
Description of the condition
Asthma is a major chronic illness of childhood that also has a significant impact on the adult population. Pediatric asthma affects 4 million U.S. children, with a prevalence of 4‐9% (NHLBI 2002), and it has been estimated that greater than 5% of American adults have asthma (GINA 2002). The International Study of Asthma and Allergies in Childhood (ISAAC 1998a) produced prevalence rates for symptoms of asthma in children ranging from 1.6% to 36.8%, with the highest rates in the UK, Australia and New Zealand followed by the Americas (ISAAC 1998b). Internationally there are fewer data about the prevalence of adult asthma though data from Australia, Belgium and Finland show rates ranging from 2‐16% (GINA 2002). The burden of asthma on children is substantial: children with asthma have a threefold greater risk of school absence than children without asthma and almost thirty percent of children with asthma have some limitation of activity, compared to five percent of children without asthma. From Australian studies, 4% of all age groups have moderate or severe asthma that requires regular medications(Peat 1994). In the U.S., over 10 million adults over the age of 18 experience over 100 million days of restricted activity annually due to asthma(Mannino 1998). 10.1 million school days are missed to asthma(Fowler 1992).
Studies documenting changes in asthma prevalence over at least nine years have observed increasing prevalence. The prevalence of adult asthma in the US has increased in recent decades as measured by diagnostic criteria and the use in acute care services for asthma(NHLBI 2002). Increased asthma morbidity may be related to increased severity of the disease, undertreatment of patients with anti‐inflammatory therapy, over‐reliance on bronchodilators and delay in seeking medical help during exacerbation. It is clear that certain subpopulations show disproportionately high mortality from asthma, for example inner‐city dwelling African Americans (Mannino 1998) in the United States and Maori in New Zealand (Sears 1987). Poverty in developed countries appears to be a risk factor for increased morbidity. Internationally, the highest prevalence rates for asthma in children are found in Australia, New Zealand and England, with large differences among different ethnic populations. There is insufficient data to determine whether these differences result from responses to the environment, to industrialization or to difference allergen loads.
Description of the intervention
The Global Initiative for Asthma (GINA 2002), in its Practical Guide to Asthma Management (http://www.ginasthma.com/), provides an evidence‐based management strategy that addresses severity assessment, pharmacotherapy, monitoring, environmental control and the patient‐provider partnership (GINA 2002). Despite the existence of effective therapies, including inhaled anti‐inflammatory and long‐acting bronchodilators, current practice falls short of guidelines and consequently, inadequate asthma control continues to be the norm. Although evidence based guidelines have been established, there is significant variation among practitioners in treating this condition, indicating a significant gap between what we know is optimal care and current practice (Finkelstein 2000, Finkelstein 2002, Lozano 2003).
How the intervention might work
We propose to undertake a review of interventions to improve the management of asthma in primary care. A variety of approaches have been taken to improve asthma care including; patient self‐management support, provider directed education and behavior change, and organizational approaches. Considerable work has been done in reviewing patient self‐management support interventions. For example, Gibson found that information only patient education programs did not appear to improve outcomes (Gibson 2001). Toelle found insufficient evidence about the effectiveness of written management plans for adults and children (Toelle 2001). Reviews of more complex self‐management interventions have shown positive results: self‐management support education for children directed at prevention and attack management showed demonstrably improved outcomes (Wolf 2002). Positive outcomes were found in studies of asthma education for adults that included self‐monitoring of either peak flow or symptoms, when combined with regular medical review and a written action plan (Gibson 2002). Another review of multifaceted adult interventions showed equivalent effects for control of asthma management by regular medical review of inhaled corticosteroid use vs. control by self‐adjustment using written self management plan (Powell 2002). To complement these reviews, we will focus our review on provider‐directed interventions and organizational approaches, only including patient‐oriented interventions when they are part of a multi‐focus study.
Why it is important to do this review
Interventions that focus on changing provider decision‐making and behavior, including well conducted trials and systematic reviews of provider level interventions, have found many to be ineffective as stand‐alone measures (Davis 1995, Davis 1999). Changes in the organization of care in order to better support provider and patient behavior, creating multi‐level interventions, have shown more success, but are not well understood (Bodenheimer 2002a, Bodenheimer 2002b). Identifying and understanding effective interventions and their implementation in primary care systems would provide useful information in bridging the gap between evidence and practice. To do so, we propose a systematic review of literature to identify such successful interventions and their implementation strategies.
Objectives
The objectives of this review are to determine the effectiveness of various interventions to improve the management of asthma in primary care settings.
Methods
Criteria for considering studies for this review
Types of studies
Randomized or quasi‐randomized controlled trials, controlled clinical trials, controlled before and after studies, and interrupted time series studies.
Types of participants
Health professionals including primary care physicians, nurses, pharmacists and other allied health care professionals working in primary health care settings. Primary health care is defined as 'integrated, easy to access, health care services by clinicians who are accountable for addressing a large majority of personal health care needs, developing a sustained and continuous relationship with patients, and practicing in the context of family and community' (Vanselow 1995).
Children and adults with asthma or reactive airways disease (RAD) receiving care in primary health care settings. We will not include studies addressing bronchiolitis in very young children, nor those addressing emphysema or chronic obstructive pulmonary disease in older adults
Types of interventions
Professional, financial and organizational intervention strategies to improve asthma care and patient directed interventions involving primary care providers or care teams.
Care delivered in community settings such as schools or community agencies will be included only if there is strong linkage to the primary care provider or the intervention is provided by a clinician that substantially replaces primary care, including treatment plan with appropriately prescribed medication regimen.
Types of outcome measures
Health Professional Outcomes/Process Measures
Guideline recommended provider practices
severity at time of visit
prescribing inhaled anti‐inflammatory medications for persistent disease
Patient Outcomes
Symptom Burden
markers of disease control
symptom days/scores
Health Status
functional health status
quality of life
school/work days lost
Utilization of Healthcare
ED visits
sick day/acute care visits
hospitalizations
Patient behavior
medication adherence
Economic Variables
cost effectiveness
Search methods for identification of studies
Electronic searches
For electronic searches, we will use the EPOC criteria with the addition of asthma specific terms in developing the search strategy. We will combine the methodological terms from the existing EPOC criteria with additional MeSH headings and text words to narrow the scope of the search to asthma and RAD care in primary practice. Electronic searches will use the Cochrane libraries Central/CCTR database and pertinent Cochrane registries.
Searching other resources
Other sources in the literature identified by EPOC collaborators will be searched upon their recommendations. These might include the Resource Base in Continuing Medical Education, the database of the Airways Group, the Consumers and Communications Group, the Literature Database on Quality Improvement, and bibliographies maintained by reviewers working in pertinent areas.
Data collection and analysis
Potentially relevant studies will be determined by review of the abstracts of search results. Two members of the review team will review studies for inclusion, undertake quality appraisal using EPOC quality criteria and abstract results (including assessment of level of baseline compliance). One reviewer (JS) will complete all other aspects of data abstraction.
Data Analysis
Because of the expected diversity of the interventions and populations, the review will not use formal meta‐analysis methods to pool results of the studies. Instead, we will present the data in tabular form. For studies reporting more than one outcome, we will summarize the findings for one outcome per category shown above, as available. If more than one outcome is reported per category, preference will be given to validated measures.
The primary analysis will be a comparison of the mean proportion (per provider) of times outcome was improved (as determined by the investigators) for each patient or episode of care. However, in practice, important differences in study participants, interventions and outcomes among the included trials and missing data may restrict the application of this analysis for published trials.
For each study, we will report the main results in natural units. For dichotomous variables, we will report risk difference and relative risk (with 95% confidence intervals and/or p‐values where possible). For continuous variables, we will report post‐intervention raw mean differences and standardized mean differences where possible (with 95% confidence intervals and/or p‐values where possible). We will also analyze relative effects of continuous data qualitatively by two reviewers looking at the distribution of effects for continuous outcomes relative to that of the dichotomous outcomes. We will categorize the size of effect relatively as small, medium or large.
We will adhere to the convention of reporting outcomes as unfavorable events, which means that a risk ratio less than unity signifies a reduction in unfavorable events.
For cluster RCTs and CBAs with potential unit of analysis errors (i.e. the unit of allocation is the practice and the unit of analysis is the patient), we will attempt to account for this by recalculating results using the appropriate unit of analysis providing sufficient data are present. If this is not possible, we will report the effect size without confidence intervals or p‐values.
For Interrupted Time Series design studies, we will abstract the difference in slope and the difference in level pre‐ to post‐ intervention. If the differences are not available in the primary reports, we will attempt reanalysis using data from graphs or tables(Ramsay 2003).
We will consider the following potential sources of heterogeneity to explain variation in the results of the included studies:
Study design
Study quality
Baseline compliance
Type of intervention
Intensity of intervention
Population differences ‐ adult vs pediatric studies
We will analyze these potential differences by preparing tables and visual displays to explore the size of observed effects relative to these variables. In particular, we will explore whether larger effects are associated with poorer study quality, lower baseline non‐compliance, more intense or multifaceted interventions or population differences such as age.
We will compare the results for studies of specific types of interventions included in this review to systematic reviews of similar interventions across a range of clinical problems.
