Empirical Literature on Asthma Care

Empirical Literature on Asthma Care This brief critically considers the empirical literature on asthma care. Emphasis is on UK studies although research from the USA (and other countries) is also considered. It is argued that both environmental and genetic factors are implicated in asthma onset, based on epidemiological evidence. Deficits in care provision persist: these gaps in care may be attributable to a wide range of modifiable factors, including unsatisfactory health professional (GP, nurses) input, limited use of care plans, and patient unawareness. Overall, however, conclusive inferences about asthma care provision are hampered by: A preponderance of retrospective/correlational studies, and a paucity of randomised control trials, which demonstrate causality; A paucity of research on particular gaps in asthma care; Failure to account for third-variable moderator effects. The Office for National Statistics (2004) publishes comprehensive statistics on asthma-related mortality, morbidity, treatment, and care, collapsed by demographic categories. Data is collected from the General Practice Research Database (GPRD). Issues addressed include mortality, prevalence, time trends, patients consulting general practice, incidence of acute asthma, and hospital inpatient admissions. Research suggests that health care providers often fail to agree on the precise criteria for diagnosing asthma, whether mild or severe (e.g. Buford, 2005). Severe asthma is often defined based on pulmonary function measurements, such as forced expiratory volume in 1 second, and hospitalisation. However, neither of these indicators reliably predicts asthma severity (Eisner et al, 2005). Eisner et al (2005) evaluated the efficacy of a method for identifying a cohort of adults with severe asthma based on recent admissions to an intensive care unit (ICU) for asthma. Four hundred adults with severe asthma enrolled at seventeen Northern Carolina hospitals were surveyed. A control group of patients hospitalised without ICU unit admission was also recruited. The study examined whether admission to an ICU unit is in itself a reliable indicator of asthma severity. Asthma patients with a recent ICU admission generated higher asthma scores (based on the frequency of current asthma symptoms, use of steroids and other medications, and history of hospitalisations/intubations), and poorer quality of life, were more likely to have been hospitalised, visited an asthma specialist in the previous twelve months, been in an asthma-related emergency department, and received inhaled corticosteroids in the past year. Data analysis controlled key background variables (e.g. demographic factors), increasing confidence in the reliability of the findings. However, this study was based on quasi-experimental design and hence may be confounded by sampling bias. Trends in annual rates of primary care consultations, mortality, and hospital visits/admissions were monitored for children under 5 years and 5-14 year olds. For children aged For 5-14 year olds, weekly general practice visits rose in the early 1990s (circa 70/100,000 in 1990), showed a fluctuating pattern through the mid 1990s, but has declined steadily since 1997 (about 50/100,000 by 2000). The number of patients treated annually for asthma has risen slowly but steadily, although this increase seemed to level out by the mid/late1990s. Both mortality rates have dropped steadily since the early 1990s, from about 14 million in 1990 to circa 2 million by 2000. Annual hospital admissions has also fallen steadily, from just under 30/10,000 in 1990 to about 15/10,000 by 2000. These patterns suggest an increase in self-management (e.g. action plans) that obviates the need to visit a general practice, and that asthma care overall is having the desired effect on mortality. The prevalence of wheezing and asthma in children has generally increased during the last 40 years. Although there is a paucity of reliable national statistics, data is available from specific parts of the UK, notably Leicester, Sheffield, and Aberdeen (see Figure 1). The prevalence of wheezing increased from 12% (1990) to 26% (1998) in Leicester, and from 17% (1991) to 19% (1999) in Sheffield. The prevalence of asthma showed a similar pattern in both cities, rising from 11% (1990) to 18% (1998) in Leicester, and from 18% (1991) to 30% (1999) in Sheffield. Wheezing incidence rates for Aberdeen increased from 10% (1964), to 20% (1989), 25% (1994), and 28% (1991). Data from national birth cohorts suggests a sharp increase in the average weekly GP consultations for hay fever/allergic rhinitis from 1991 to 1992. The rates rose from circa 13/100,000 (0-4 year olds) and 40/100,000 (5-14 years olds) in 1991 to about 25/100,000 (0-4 year olds) and 76/100,000 (5-14 year olds) as 1992 approached. Trends subsequently dropped off slightly but then started to show an increase again around 1998. By the year 2000 the figures were roughly 20/100,000 (0-4 year olds), and 56/100,000 (5-14 year olds). Data from a nationally representative sample of schools across the country suggests that the prevalence of asthma was fairly even across different regions. However, Data for England suggests a higher prevalence outside big cities. The greatest proportions of wheezing was found in the South West, while the highest proportion of asthma cases was found in East Anglia and Oxford (see Figure 2). In a recent Annual Report, Asthma UK (2003/2004) noted that one child in 10 has asthma and a child is admitted to hospital every 18 minutes due to an asthma attack. Over 600 copies of Asthma in the Under Fives are downloaded from the UK Asthma website monthly and on average every classroom in the UK has at least 3 children with asthma. The impact of acute asthma can be debilitating. Around 5.2 million people in Britain are presently being treated for asthma, and asthma prevention/care costs the NHS on average almost  £900 (i.e.  £889) million per year. GPs across the country treat over 14,000 new episodes of asthma each week, and UK Asthma met almost 25,000 requests for health promotion documents and other materials. About 40% of workers who have asthma find that working actually exacerbates their asthma, and 1 in 5 asthmatic people feel excluded from areas of the workplace in which people smoke. Over 12.7 million working days in the UK are lost as a result of asthma, and it is estimated that the annual cost of asthma to the economy is  £2.3 billion. Asthma UK also states that 82% of people who are asthmatic find that passive smoking triggers their asthma, and 19% of people with asthma indicate that their medical condition makes it difficult for them to play with children in their family. One in 3 children has had their routine daily activities disrupted due to asthma and 39% of asthmatic people are badly affected by traffic fumes (which stop them exercising). About 500,000 people have asthma that is very difficult to control. In 2003/2004 over 90 researchers worked on Asthma UK-funded projects and, Asthma UK spent  £2.5 million on asthma-related research. The group funded/is funding 63 research projects. These statistics paint a rather bleak picture of asthma prevalence, incidence, and the effects on people’s lives. Numerous epidemiological studies have been published that address the etiology of asthma in population groups (International Archives of Allergy Immunology, 2000; Kitch et al, 2000; Schweigert et al, 2000; Tan, 2001; Court et al, 2002; Smyth, 2002; Weissman, 2002; Tan et al, 2003; Wenzel, 2003; Gibson Powell, 2004; Barnes, 2005; Pinto Almeida, 2005). Barnes (2005) considered evidence on the role of genetic factors in resistance to atopic asthma, Studies which focus on the role of genetic factors in resistance to tropical/parasitic diseases (e.g. malaria) overlap with genetic associations found for asthma. It was concluded that genetic factors might be implicated in the development of allergic illnesses. Pregnancy is thought to increase the probability of asthma attacks in about 4% of all pregnant women. Beckmann (2006) assessed eighteen pregnant women with asthma. The study was based on a longitudinal design. Participants were recruited from local prenatal clinics and private enterprises, and enrolled during the first trimester. Patients kept a daily log recording peak expiratory flow data until delivery. Three peak-flow assessments were recorded after which the best value was entered into the log. Asthma was diagnosed by a health professional. Participants were also required to record asthma symptoms, exacerbations, medications, and cigarette use. To increase participation, subjects were reminded by telephone to complete their log. Data analysis showed that peak expiratory flow (PEF) was variable as a function of particular trimesters. Peak air flow was highest during the second trimester, with a statistically reliable difference between the second and third trimester. Unfortunately, the small sample size limits the generalisability of the findings. However, the study was based on a longitudinal design, allowing tentative causal inferences. Schweigert et al (2000) reviewed the literature on the role of industrial enzymes in occupational asthma and allergy. Enzymes used by detergent manufacturing companies (e.g. amylases, cellulases) are toxicologically benign, with mild irritation effects on the body. However, these enzymes do affected asthma and allergy. Thus, the industry is required to adhere to exposure guidelines for these enzymes. Kitch et al (2000) considered literature on the histopathology of late onset of asthma (i.e. onset in adulthood), and whether allergic exposure and sensitivity have the same impact on asthma development in adulthood as they do in children. Epidemiological studies suggest that the prevalence of asthma in older adults aged 65years or more is between 4% and 8%. The illness appears to be more common in women, especially those with a long history of smoking, and with respiratory symptoms (e.g. cough, wheeze, shortness of breath). Asthma in adulthood often developed before the age of 40, with maximum incidence occurring around early childhood. Beyond the age of 20 years the incidence of asthma tends to remain stable through young, middle-aged, and older adulthood. Death rates in adults are generally lower than figures for children; â€Å"Mortality rates attributable to asthma among those aged between 55 and 59 years of age and 60 and 64 years of age were 2.8 and 4.2 respectively, per 100,000 people, the highest rates among all age groups† (p.387). However, as adults get older asthma is less and less likely to be identified as the main cause of death due to the increased incidence of other pathology. Epidemiological research in Japan highlights a link with air pollution (International Archives of Allergy Immunology, 2000). The prevalence of asthma among kindergarten and elementary school children has increased steadily since the early 1960s, rising from 0.5-1.2% between 1960 and 1969, to 1.2-4.5% (1970-1979), 1.7%-6.8% (1980-1989), and 3.9-8.2% (1990 onwards). By contrast, data indicates little or no change in asthma prevalence amongst adults. Figures range from 1.2% in 1950-1959 to 1.2-4.0% (1960-1969), 0.9-5.0% (1970-1979), 0.5-3.1% (1980-1989) the 1960s to 1.6-2.9% (1990 onwards) (see Figure 3). Asthma in Japanese children is more common amongst boys than girls although this gender difference has diminished noticeably since the 1960s. Asthma usually appears in infancy or early in childhood but has been known to begin across all age groups. Inherited (genetic) dispositions to allergies have been implicated in the onset of asthma. There is normally a strong correlation between asthma onset and a family history of asthma. Overall, asthma-related mortality in Japan has decreased since the mid 1990s. Delays in seeking treatment and rapid exacerbation of symptoms have been strongly implicated in asthma mortality. Unfortunately, this article offers little information about the designs of studies reviewed. Inferences regarding the possible causes of asthma morbidity and mortality may be inconclusive if much of the evidence is derived from cohort studies, rather than case control studies that more effectively eliminate alternative causes. The premenstrual period in women has been implicated in asthma exacerbation. Tan (2001) reviews epidemiological literature suggesting that female sex-steroid hormones may be significant in understanding the premenstrual-asthma link, albeit the available evidence is tenuous. The luteal phase of the menstrual cycle is associated with airway inflammation and hyper-responsiveness, and hence may explain asthma exacerbation during the premenstrual phase. However, this increase in asthma severity can still be treated effectively using the normal drugs. Studies suggest that premenstrual asthma affects the rate of hospital admissions – the majority of adults admitted are women, indicating that hormonal factors play an important role. Other evidence suggested that emergency presentations increased before ovulation. It is suggested that oral contraceptive pills or gonadotrophin releasing hormone analogues may be especially effective treatments. However, premenstrual asthma was rarely associated with serious mortality. Unfortunately, most of the studies reviewed were retrospective and questionnaire based, and hence subject to response bias. There was a paucity of randomised control trials, or pseudo experiments that may permit causal inferences. Court et al (2002) considered the distinction between atopic (extrinsic) asthma, common in younger people, and non-atopic (intrinsic) asthma, found mostly in older groups. Additionally, they also considered whether identification of asthma cases in epidemiological research should be based on a doctor’s diagnosis or self-reported asthma symptoms. Nearly 25,000 people in England were surveyed. Data was collected regarding whether participants had experienced wheezing in the past 12 months and/or had been diagnosed as asthmatic by a doctor. People with atopic asthma were more likely to have experienced wheeze and been diagnosed as asthmatic in the past, compared with the non-atopic group. Logistic regression analysis showed that gender, social class, smoking status, living in an urban/rural area, and house dust mice (HDM), were all risk factors for the presence of wheeze both with (age not significant) and without (urban/rural area not significant) a diagnosis of asthma. Wheeze/asthma was more prevalent in women, younger people, lower social classes, previous/current smoking, living in an urban area, and greater HDM IgE levels. Smoking status, social class, and age were all risk factors for wheeze in both atopic and non-atopic cases. Gender was also a risk factor for atopic subjects, and urban living for non-atopics. Other research has considered the epidemiology of severe or ‘refractory’ asthma, which is rather less well understood compared with milder forms of asthma. Wenzel (2003) reviews evidence indicating that severe asthma (defined as asthmatics requiring continuous high-dose inhaled corticosteroids or oral corticosteroids for over half of the preceding year) may account for circa ≠¤ 5% of asthma cases. Data from a large Australian-based study, which has followed a large cohort of asthmatics for over three decades, implicates childhood pulmonary problems with reduced lung function in adulthood. Data suggests that over two-thirds of severe asthmatics were afflicted with asthma in childhood. Other risk factors implicated include genetic mutations (in the IL-4 gene and IL-4 receptor), and environmental factors (e.g. allergen, tobacco exposure, house dust mite, cockroach and alternaria exposures), respiratory infections (e.g. pathogens like chlamydia), obesity, gastroesophageal reflux disease, increased body mass index, lack of adherence to corticosteroid regimes, and poor physiological response to medication. Physiological factors are also implicated, notably structural changes in airway reactivity, inflammation of the peripheral regions of the lungs. Steroids are the main form of treatment. Tan et al (2003) demonstrated the role of respiratory infection in patients with severe (i.e. near fatal) asthma, acute exacerbations, or chronic obstructive pulmonary illness (COPD). Participants had all been diagnosed as asthmatic by a physician and were undergoing treatment. All showed evidence of forced expiratory volume in 1 second (FEV1) increase of 200mL. COPD patients were suffering from chronic cough and dyspnea, with a predicted FEV1% 50%, with no ÃŽ ²-agonist reversibility. Near fatal cases were patients undergoing ventilatory support in the intensive care unit of a hospital (National University Hospital and Alexandria Hospital, Singapore) as a result of a severe exacerbation. Acute asthma subjects were characterised by non-improvement following administration of ÃŽ ²-agonists, and/or severe exacerbation judging from clinical/blood data. Analysis showed that near-fatal cases were the least likely to have the influenza A + influenza B virus, but the most prone to have adenovirus and picornavirus, compared with the other two groups (see Figure 4). This suggests that viral infection may be a risk factor for severe asthma. However, due to sampling size/bias (n= 68), and failure to control for key background variables (e.g. asthma history, smoking history, prior medication use, and outpatient spirometry), the findings can be considered tentative. Smyth (2002) reviewed epidemiological studies on asthma in the UK, and worldwide. The number of new asthma cases seen by GPs has increased noticeably since the mid 1970s. Nevertheless, asthma incidence has tended to decrease since the early 1990s, consistent with data from the Office for National Statistics (2004). By the year 2000 circa 60-70, 40-50, 20-25 new cases (per 100,000 of a given age group) were reported amongst, respectively, preschool children, 5-14 year olds, and people older than 15 years. Significant ethnic differences have been reported, with high asthma prevalence in Afro-Caribbean children. Since 1962, the number of preschool children hospitalised for asthma rose steadily, then peaked in the late 80s and early 90s, and has begun to decline since. The hospitalisation rates in 1989 were 90/10,000 (preschool children), 30/10,000 (5-14 year olds), and 10/10,000 (15 years or older). By comparison the rates for 1999 were 60/10,000, 20/10,000, and 10/10,000 respectively ( see Figure 5). The British Thoracic Society identifies specific benchmarks or ‘best practice’ which health professionals are required to meet when caring for asthma patients (BTS, 2004). These recommendations are mostly based on scientific evidence from RCTs, epidemiological studies (cohort and case-control), meta-analytic reviews, and other good quality research. The recommendations related specifically to the following topics: Diagnosis and assessment in children and adults (e.g. key symptoms, recording criteria which justified diagnosis of asthma); Pharmacological management (e.g. use of drugs [inhaled steroids, ÃŽ ²2 agonist] to control symptoms, prevent exacerbation, eliminating side effects, employing a ‘stepwise’ protocol for treatment); Use of inhaler devices (technique and training for patients, agonist delivery, inhaled steroids, CFC vs. HFA propellant inhalers, suggestions on prescribing devices); Non-pharmacological management (e.g. breast feeding and modified milk formulae for primary prevention, and allergen avoidance for secondary prevention, alternative medicines); Management of acute asthma (initial assessment, clinical features, chest x-rays, oxygen, steroid treatment, referral to intensive care) Asthma in pregnancy (drug therapy, management during labour, drug treatment in breastfeeding mothers); Organisation and delivery of care (e.g. access to primary care delivered by trained clinicians, regular reviews of people with asthma, audit tools for monitoring patient care after diagnosis); Patient education (e.g. action plans, self-management, compliance with treatment regimes). Overall, despite these guidelines, recent research suggests that patients’ treatment needs are not being met. For example, Hyland and Elisabeth (2004) report data on the unmet needs of patients. Focus groups were organised between parents, patients, and clinicians. Patients and parents reported various needs that weren’t been met including frequent exacerbations, and a preference for less complex drug regimens (i.e. with fewer drugs). Many individuals had worries regarding treatment and experienced asthma symptoms 3 or more days per week. As Levy (2004) suggests, there is a need for health professionals to address these concerns, especially in relation to the BTS guidelines. Levy, a GP and Research Fellow in Community Health, identified current deficiencies in the care of asthma victims. These comprised: Higher than expected exacerbations (42/1000 patients per year); Under-diagnosis: more patients presenting for treatment with uncontrollable asthma, who had not been diagnosed previously; Deficiencies in treatment uptake: many patients fail to collect their prescriptions; Many patients with symptoms delay presenting for treatment, until their medical situation becomes critical; Health care professionals are failing to assess patients objectively (PEF, oximetry), both pre- and post-treatment; Failure to adhere to national guidelines for the care of acute asthma (e.g. not enough oral steroids and ÃŽ ²-agonists are prescribed for patients presenting with asthma attacks. Considerable variations across GPs, NHS Trusts, clinics, and other sources of care provision: patient follow-up appointments range from a few days to six months, in direct violation of standards set by the British Thoracic Society (BTS, 2004). Levy suggests various strategies for improving asthma care including diagnosis criteria (e.g. â€Å"any patient with recurring or respiratory symptoms [cough, wheeze, or shortness of breath], or who has been prescribed anti-asthma treatment should be considered to have asthma† (p.44)), use of computerised templates, having systems or triggers in place for recalling patients (e.g. patients requesting more medication, or who have been seen out of hours), introducing more effective protocols for monitoring and informing asthma patients (e.g. using a checklist to ascertain various key information on patients status, such as effects of asthma on patients life, recent exacerbations), providing written self-management plans (e.g. how to detect uncontrolled asthma, using PFM charts), and having an agreed procedure for managing acute asthma attacks (e.g. selecting a low threshold for using oral steroids). Currently there is a lack of research testing the value of these recommendations on asthma health outcomes. However various strategies are continually being implemented in various parts of the country to improve the quality of asthma care. For example, Holt (2004) describes the effects of implementing the RAISE initiative, launched by the National Respiratory Training Centre, in a primary care setting. This scheme is designed to raise awareness of existing variations in standards of care, improve standards of care through education, support, and feedback, increase awareness and understanding of respiratory disease, use asthma as platform to demonstrate the value of shared experiences across different agencies/professionals, and augment the profile of primary care settings as the main source of asthma care and innovation. The RAISE led to various improvements, such as: The use of ‘active’ and ‘inactive’ asthma registers, to distinguish patients who currently have asthma symptoms from those who don’t. Introduction of computerised templates to improve accuracy and reliability of data recording during consultations (e.g. progressing sequentially from assessment of symptoms, to peak flow, inhaler, and advice stages). Use of symptom questionnaires (e.g. handed out with repeat prescriptions) that help patients with well-managed asthma decide whether they can opt for a telephone consultation, rather than taking the trouble to visit the practice for a face-to-face consultation. Haggerty (2005) identifies several factors paramount to effective care and management of asthma in UK patients. These comprise adequate patient education about the nature of asthma (e.g. number of asthma episodes, use of quick relief medicines, long term symptoms, restrictions on daily activities, and emergency visits), use of asthma action plans, and customised treatment plans (to achieve early control), and addressing patients own concerns and perception. Treatment for asthma is usually in the form of regular inhaled corticosteroids (ICS), oral corticosteroids (OCS), and ÃŽ ² agonists. These treatments are usually administered by a health professional when symptoms manifest and/or become severe. However, since asthma can often exacerbate rapidly, before an individual can seek medical help, it is vital that asthma patients receive the necessary care from health professionals, and also self-management skills. GPs and nurses play a critical role. Griffiths et al (2004) conducted a randomised control trial to assess the effect of a specialist nurse intervention on the frequency of unscheduled asthma care in an inner city multiethnic clinic in London. The role of specialist nurses in asthma care has been uncertain. Interventions in which specialist nurses educate patients about asthma, after hospital attendance with acute asthma, were shown to have inconsistent effects on unscheduled care. However, outreach initiatives to educate medical staff had shown no effect. Thus, an intervention was designed that combined patient education with educational outreach for doctors and practice nurses. It was suspected that such an integrated approach would benefit ethnic minority groups, especially given their higher hospital admission rates and reduced access to care during asthma exacerbation. The key research question was whether specialist nurses could improve health outcomes in ethnic minority groups. Outcome variables were the percentage of patients receiving unscheduled treatment for acute asthma during a 12 month period, and time to first unscheduled attendance with acute asthma. The study was based on 44 practices in two east London boroughs. Participants comprised over 300 patients (aged 6 to 60) who were admitted to or attending the hospital, or the out of hours GP service with acute asthma. Half the sample were classified as South Asians, 34% were Caucasian, while 16% were Caucasian. The intervention was based on a liaison model. Practices were assigned to either a treatment or control condition. Practices randomised to the treatment condition ran a nurse led clinic involving liaison with GPs and practice nurses, incorporating education, raising the profile of guidelines for the management of acute asthma, and providing on-going clinical support. In practice these practices received two one-hour visits from a specialist nurse who discussed guidelines for managing patients with acute asthma. Discussions were based on relevant empirical evidence. A computer template was provided to elicit patient information on various treatment issues, such as inhaler technique and peak expiratory flow, and offer self-management advice. By contrast, control practices received a visit promoting standard asthma care guidelines. Data analysis showed that the intervention lengthened the time to first attendance (median 194 days for intervention practices, and 126 days for control practices), and also reduced the proportion of patients presenting with acute asthma (58% treatment practices versus 68% in control practices (see Figure 6). These effects were not moderated by individual differences in ethnicity, albeit Caucasians seemed to benefit more from the intervention compared with minority ethnic groups. O’Connor (2006) noted that asthma care in the UK remains below the required standards. The majority of the 69,000 hospital admissions and circa 1400 deaths annually are attributable to poor patient adherence to treatment regimens. Nurses, it is argued, play an important role in promoting adherence. Additionally, use of a new inhaled corticosteroids – circlesonide – may also help increase adherence. Circlesonide is much easier to use than more established asthma drugs (e.g. it has a once-daily dosing). Evidence is reviewed suggesting that peak expiratory flow remains stable when patients are given circlesonide compared with a placebo. Tsuyuki et al (2005) assessed the quality of asthma care delivered by community-based GPs in Alberta, Canada. They reviewed clinical charts for over 3000 patients from 45 primary care GPs. Of this number 20% had ever visited an emergency department or hospital, 25% had evidence that a spirometry had been performed, 55% showed no evidence of having received any asthma education, 68% were prescribed an inhaled corticosteroid within the past 6 months, while a very small minority (2%) had received a written action plan. Figure 6 shows percentage of participants receiving medication. Sixty-eight percent were prescribed an inhaled corticosteroid, 11% were given an oral corticosteroid, and 80% received a short acting ÃŽ ²-agonist, while 8% were prescribed a long acting ÃŽ ²-agonist. Participants with an emergency room/hospital event were (marginally) more likely to be prescribed medication (no group differences in use of short acting ÃŽ ²-agonists). Regarding pulmonary testing, 25% had evidence of a pulmonary function test (not peak flow), 46% had peak flow monitored, 34% showed no evidence of pulmonary function tests, while 26% had an x-ray. Again individuals with an emergency room/hospital event were more likely to be tested (see Figure 7). Data about education received by patients was also evaluated. Twenty-two percent received information about environmental triggers, 20% on inhaler use, 10% on how to perform a home PEF test, 2% on written action plans, while 55% received no education at all. Those with an emergency room/hospital event were more likely to receive education. Receiving asthma education, use of spirometry, and prescription of inhaled corticosteroids, were all predicted by number of asthma-related clinic visits (4 or more) and having an emergency room/hospital event. Additionally, asthma education was predicted by cormorbidities, and absence of documentation regarding asthma triggers, while use of spirometry was predicted by being a non-smoker, and symptoms or triggers. Finally, use of inhaled steroids was predicted by symptoms. Overall, this study highlights numerous gaps in the care provided by GPs, partly echoing criticisms of GPs in the UK (Levy, 2004). For example, Levy (2004) cited ‘under