- Email: [email protected]
Family physician ethnicity influences quality of diabetes care for Chinese but not South Asian patients
p r i m a r y c a r e d i a b e t e s 9 ( 2 0 1 5 ) 432–438
Contents lists available at ScienceDirect
Primary Care Diabetes journal homepage: http://www.elsevier.com/locate/pcd
Original research
Family physician ethnicity influences quality of diabetes care for Chinese but not South Asian patients Baiju R. Shah a,b,c,∗ , Jeremiah Hwee a , Sonia S. Anand d , Peter C. Austin a,b , Douglas G. Manuel a,e , Janet E. Hux a,1 a
Institute for Clinical Evaluative Sciences, Toronto, ON, Canada University of Toronto, Toronto, ON, Canada c Department of Medicine, Sunnybrook Health Sciences Centre, Toronto, ON, Canada d McMaster University, Hamilton, ON, Canada e University of Ottawa, Ottawa, ON, Canada b
a r t i c l e
i n f o
a b s t r a c t
Article history:
Aims: To determine whether sharing the same ethnicity as their family physician influenced
Received 9 July 2014
the quality of diabetes care for Chinese and South Asian patients in Ontario, Canada.
Received in revised form
Methods: We conducted two related studies: a population-based cohort study of Chinese
3 October 2014
and South Asian patients with incident diabetes using health care administrative data
Accepted 10 February 2015
(n = 49,484), and a cross-sectional study of Chinese and South Asian patients with estab-
Available online 8 March 2015
lished diabetes using data collected directly from their family physicians’ clinical records (n = 416). In both studies, quality of care measures were compared between patients whose
Keywords:
family physicians were or were not from the same ethnic group.
Ethnicity
Results: In the cohort study, Chinese patients whose family physicians were also Chinese
Chinese
were more likely to have a diabetes-related family physician visit and appropriate HbA1c
South Asian
and cholesterol testing. In the cross-sectional study, they were more likely to have foot
Quality of care
examinations, to have microalbuminuria testing, and to achieve recommended treatment
Family physicians
targets for HbA1c and for LDL-cholesterol. In contrast, for South Asian patients, most quality measures in either study did not differ by physician ethnicity. Conclusions: Having a family physician from the same ethnic group was associated with better quality of diabetes care for Chinese but not for South Asian patients. © 2015 Primary Care Diabetes Europe. Published by Elsevier Ltd. All rights reserved.
∗
1
Corresponding author at: G106 – 2075 Bayview Avenue, Toronto, ON, Canada M4N 3M5. Tel.: +1 416 480 4706; fax: +1 416 480 6048. E-mail address: [email protected] (B.R. Shah). Present address: Canadian Diabetes Association, Toronto, Ontario, Canada.
http://dx.doi.org/10.1016/j.pcd.2015.02.002 1751-9918/© 2015 Primary Care Diabetes Europe. Published by Elsevier Ltd. All rights reserved.
p r i m a r y c a r e d i a b e t e s 9 ( 2 0 1 5 ) 432–438
1.
Introduction
Disparities in chronic disease care for ethnic minority populations are well documented [1–3], but the modifiers of these disparities are uncertain. For example, the ethnicity of their primary care physician may influence the quality of care received by minority patients [4]. Physicians from the same ethnic group may be able to overcome language barriers to care, and may have a better understanding of the patient’s dietary practices or cultural norms. American studies have shown that minority patients with physicians report greater satisfaction and fewer unmet care needs with physicians from the same ethnic group [5,6]. This literature has also suggested better access to and utilization of care, and improved quality of care [7–9]. However, a study specifically of patients with diabetes found that physician ethnicity had no impact on glycemic, blood pressure or lipid control for African-American and Hispanic patients [10]. There are no studies on the impact of physician ethnicity for minority patients from other countries, and limited data for other ethnic groups. Diabetes, an archetypal and increasingly prevalent chronic disease [11], leads to reduced quality of life and premature mortality [12,13]. Research in many jurisdictions has found that quality of care at a population level falls short of recommended targets [14–16]. Ethnic minorities may be particularly vulnerable [17–19], although in our jurisdiction, minorities receive similar quality of care to the general population, albeit still below recommended goals [20]. We conducted two related studies with an objective to determine whether sharing the same ethnic origin as their family physician influenced the quality of diabetes care for Chinese and South Asian patients in Ontario, Canada.
2.
Methods
2.1.
Design, data sources and patient selection
databases and across time through unique but anonymous numbers. Using the ODD, all individuals in Ontario newly diagnosed with diabetes between January 1, 2000 and December 31, 2008 were identified. Inclusion criteria were age ≥20 years, eligibility for provincial health insurance for ≥2 years (to ensure that new immigrants with prevalent diabetes were not inadvertently included as incident cases due to a lack of prior data), and Chinese or South Asian ethnicity. Because Canadian health care data do not contain ethnic identifiers, we identified Chinese and South Asian patients by identifying all those whose earliest known surnames (i.e., before name changes through marriage) matched two lists of surnames validated in this population to have excellent positive predictive values when compared to self-reported ethnicity (91.9% for Chinese and 89.3% for South Asian) [22]. In Ontario, virtually all primary care is delivered by family physicians. Each patient was assigned to their regular family physician based on ambulatory billing claims [23]; those patients who could not be assigned to a family physician were excluded. Although population-based health administrative data are comprehensive for examining health service utilization, they lack the detailed clinical information needed to more broadly assess quality of care. To overcome this limitation, we also conducted a cross-sectional study of clinical quality of care measures by reviewing patients’ clinical records in their family physicians’ offices. We enrolled randomly selected family physicians practicing in neighborhoods where ≥25% of the population reported either Chinese or South Asian ethnicity in the 2006 Canadian census. (These neighborhoods are all in Toronto or its suburbs.) Enrolment was stratified by selfreported physician ethnicity. In each participating physician’s practice, we randomly selected up to 10 patients who were age ≥18 years, had type 2 diabetes for ≥2 years, and had Chinese or South Asian ethnicity (based on recorded ethnicity in the clinical record, or reported by the physician). Details of the abstraction are presented elsewhere [20].
2.2. We conducted these studies in Ontario, Canada’s most populous province, using different study designs, data sources, patient populations and outcome measures. First, we conducted a population-based cohort study. The study used health care databases from the government-funded health insurance program of the Ontario Ministry of Health and Long-Term Care, which provides coverage to all permanent residents of Ontario. The data used in the study included demographic information for all Ontario residents, demographic and practice information for all licensed physicians in Ontario, claims data from physicians and clinical laboratories for fee-for-service reimbursement, and abstracts of all hospitalizations and emergency department visits. We also used the Ontario Diabetes Database (ODD), a registry derived from administrative data that identifies all people with physician-diagnosed non-gestational diabetes in Ontario. When validated against chart review, the ODD was found to have a sensitivity of 86% and a specificity of 97% [21]. Although the database does not distinguish between types of diabetes, the vast majority of cases would have type 2 diabetes. Individual patients and physicians are linked between all the
433
Exposure and quality of care measures
In the cohort study, physicians were identified as Chinese or South Asian based on their surnames as noted above. To improve sensitivity in identifying ethnic minority physicians, we also included as Chinese or South Asian those who had graduated from a medical school in (respectively) China, Taiwan or Hong Kong; or India, Pakistan, Bangladesh or Sri Lanka. This additional surrogate information for ethnicity was available through the physician registration data. In the crosssectional study, physician ethnicity was self-reported. Patients in the cohort study were followed using the claims data for 2 years after diabetes diagnosis to determine the following process measures of quality of care, derived from recommendations in national diabetes clinical practice guidelines [24]: at least one family physician claim for an ambulatory diabetes-related care visit, at least one ophthalmologist or optometrist claim for a retinal screening visit, at least four laboratory claims for HbA1c tests, and at least two laboratory claims for cholesterol tests. The clinical quality of care measures evaluated in the cross-sectional study were glycemic control (most recent HbA1c ≤7.0% [53 mmol/mol]),
434
p r i m a r y c a r e d i a b e t e s 9 ( 2 0 1 5 ) 432–438
714,948 people aged ≥20 years diagnosed with diabetes in Ontario, 2000 to 2008
27,444 Chinese adults with newly-diagnosed diabetes
33,103 South Asian adults with newly-diagnosed diabetes
3,399 with health insurance coverage for <2 years
7,234 with health insurance coverage for <2 years
180 with no family physician
250 with no family physician
23,865 Chinese patients included in the cohort study 17,381 with Chinese physician
6,484 with nonChinese physician
25,619 South Asian patients included in the cohort study 11,434 with South Asian physician
14,185 with nonSouth Asian physician
Fig. 1 – Patients included in the cohort study.
blood pressure control (most recent ≤130/80), cholesterol control (most recent LDL-cholesterol ≤2.0 mmol/L), at least one microalbuminuria test documented in the preceding year, and at least one foot examination documented in the preceding year.
3.
Analyses
In both studies, family physician ethnicity was analyzed separately for Chinese patients (Chinese family physicians versus others) and South Asian patients (South Asian family physicians versus others). Baseline characteristics were compared using t-tests or chi-square tests. For each quality of care measure, we estimated logistic regression models to determine the independent, adjusted effect of physician ethnicity. The models were estimated using generalized estimating equation methods to account for the clustering of patients within physicians. Models adjusted for patient age, sex, socioeconomic status (defined based on neighborhood median household income, divided into quintiles), rural residence (based on the patient’s home postal code), prior history of coronary artery disease, heart failure or stroke (based on hospitalization records from the previous 5 years), and hypertension (using a validated administrative data-derived algorithm) [25]. In the cohort study, the models were additionally adjusted for the following physician variables: age, sex, number of years in practice, and whether or not the physician was a Canadian medical graduate. The studies were approved by the research ethics board of Sunnybrook Health Sciences Centre, Toronto.
4.
Results
4.1.
Patient and physician characteristics
In the cohort study, there were 714,948 people aged ≥20 years diagnosed with diabetes in Ontario, 60,547 of whom were
Chinese or South Asian. After excluding 10,633 (17.6%) who had health insurance coverage for <2 years and 430 (0.7%) who had no family physician, the final cohort included 23,865 Chinese patients and 25,619 South Asian patients (Fig. 1). Seventy-three percent of Chinese patients had a Chinese physician, while only 45 percent of South Asian patients had a South Asian physician. At baseline, both Chinese and South Asian patients with physicians of the same ethnicity were older, had lower socioeconomic status and were less likely to live in rural areas than those with discordant physicians (Table 1). For the cross-sectional study, we recruited 26 family physicians practicing in neighborhoods with large South Asian populations (13 of whom were themselves South Asian), and 19 physicians in neighborhoods with large Chinese populations (15 of whom were themselves Chinese). We were unable to find more non-Chinese physicians practicing in these neighborhoods who had sufficient numbers of Chinese patients with diabetes in their practice. The baseline characteristics of the randomly selected eligible patients who were included in the cross-sectional study are shown in Table 1.
4.2.
Quality of care measures
Quality of care is presented in Table 2, with adjusted odds ratios showing the effect of family physician ethnicity independent of baseline differences. Among Chinese patients, most quality of care measures were better in those who had a Chinese family physician. In the cohort study, they were more likely to have a diabetes-related family physician visit, appropriate HbA1c testing and appropriate cholesterol testing. In the cross-sectional study, they were more likely to have had microalbuminuria testing and foot examination. They were also markedly more likely to have met targets for glycemic control (64.3% vs 35.3%, OR 2.87, 95% CI 1.20–6.87) and cholesterol control (57.1% vs 29.4%, OR 2.77, 95% CI 1.32–5.82). However, they were marginally less likely to receive a retinal screening examination. In contrast,
435
p r i m a r y c a r e d i a b e t e s 9 ( 2 0 1 5 ) 432–438
Table 1 – Baseline characteristics of Chinese and South Asian patients in the cohort and cross-sectional studies, stratified by the ethnicity of their family physician. Characteristic
Chinese patients Chinese physician
Cohort study Patients N Age, mean ± SD Sex Male Female Socioeconomic status Lowest 2 3 4 Highest Rural residence Previous CAD Previous heart failure Previous stroke Previous hypertension Physicians N Year of birth ≤1949 1950–1959 1960–1969 ≥1970 Sex Male Female Year of starting practice ≤1974 1975–1984 1985–1994 ≥1995 Canadian graduate Cross-sectional study Patients N Age, mean ± SD Sex Male Female Socioeconomic status Lowest 2 3 4 Highest Previous CAD, heart failure or stroke Previous hypertension a
Non-Chinese physician
17,381 57.3 ± 14.3
6484 55.2 ± 14.4
9314 (53.6%) 8067 (46.4%)
3440 (53.1%) 3044 (46.9%)
3576 (20.6%) 4482 (25.8%) 3511 (20.2%) 3349 (19.3%) 2440 (14.0%) 29 (0.2%) 216 (1.2%) 105 (0.6%) 121 (0.7%) 8228 (47.3%)
1360 (21.0%) 1387 (21.4%) 1269 (19.6%) 1236 (19.1%) 1215 (18.7%) 93 (1.4%) 110 (1.7%) 57 (0.9%) 50 (0.8%) 3003 (46.3%)
South Asian patients P
<0.001 0.5
South Asian physician
Non-South Asian physician
11,434 51.8 ± 13.7
14,185 51.3 ± 13.3
6222 (54.4%) 5212 (45.6%)
7997 (56.4%) 6188 (43.6%)
3370 (29.5%) 3172 (27.7%) 2557 (22.4%) 1514 (13.2%) 809 (7.1%) 20 (0.2%) 295 (2.6%) 70 (0.6%) 48 (0.4%) 4764 (41.7%)
3929 (27.7%) 3321 (23.4%) 3016 (21.3%) 2180 (15.4%) 1720 (12.1%) 74 (0.5%) 438 (3.1%) 98 (0.7%) 83 (0.6%) 5782 (40.8%)
<0.001
619
<0.001 0.007 0.02 0.5 0.2
2539 832 (32.8%) 890 (35.1%) 646 (25.4%) 171 (6.7%)
470 (75.9%) 149 (24.1%)
1698 (66.9%) 841 (33.1%)
148 (23.9%) 218 (35.2%) 188 (30.4%) 65 (10.5%) 436 (70.4%)
750 (29.5%) 828 (32.6%) 716 (28.2%) 245 (9.6%) 1808 (71.2%)
473
3452
150 (31.7%) 133 (28.1%) 125 (26.4%) 65 (13.7%)
1092 (31.6%) 1117 (32.4%) 940 (27.2%) 303 (8.8%)
264 (55.8%) 209 (44.2%)
245 (71.0%) 1001 (29.0%)
142 (30.0%) 137 (29.0%) 122 (25.8%) 72 (15.2%) 118 (24.9%)
981 (28.4%) 1074 (31.1%) 997 (28.9%) 400 (11.6%) 2462 (71.3%)
129 60.8 ± 10.9
117 55.5 ± 9.6
<0.001
0.051
148 66.3 ± 13.3
22 61.1 ± 13.1
82 (55.4%) 66 (44.6%)
13 (59.1%) 9 (40.9%)
18 (12.2%) 40 (27.0%) 33 (22.3%) 31 (20.9%) 26 (17.6%) 7 (4.7%) 112 (75.7%)
7 (31.8%) 5 (22.7%)
0.08 0.7
0.07
66 (51.2%) 63 (48.8%)
78 (66.7%) 39 (33.3%)
32 (24.8%) 25 (19.4%) 27 (20.9%) 30 (23.3%) 15 (11.6%) 11 (8.5%) 90 (69.8%)
31 (26.5%) 35 (29.9%) 30 (25.6%) 11 (9.4%) 10 (8.5%) 9 (7.7%) 73 (62.4%)
0.1
a a a a
0.9 0.8
16 (72.7%)
<0.001 0.02 0.4 0.06 0.1
0.004
<0.001
0.7
<0.001 0.002
<0.001
0.01 200 (32.3%) 181 (29.2%) 187 (30.2%) 51 (8.2%)
P
<0.001
0.003 0.01
0.04
0.8 0.2
Cell sizes <5 are not reported to protect confidentiality of patients.
South Asian patients who had South Asian family physicians were less likely to have appropriate HbA1c testing or a retinal screening examination in the cohort study. Other measures of quality of care in the cohort study, and all the measures examined in the cross-sectional study, were not associated with family physician ethnicity for South Asian patients.
5.
Discussion
In both the cohort and cross-sectional studies, Chinese patients with diabetes whose family physicians were also Chinese received better quality of care and were markedly more likely to achieve glycemic and cholesterol targets. In contrast,
436
Odds ratios are adjusted for patient age, sex, socioeconomic status, rural residence, prior history of CAD, heart failure or stroke, and hypertension. In the cohort study, odds ratios were additionally adjusted for physician age, sex, years in practice, and graduation from a Canadian vs foreign medical school. a
0.91 0.62 0.28 0.53 0.13 0.97 (0.53–1.77) 0.85 (0.45–1.60) 0.70 (0.37–1.34) 0.73 (0.27–1.96) 0.49 (0.20–1.24) 47 (42.7%) 72 (64.9%) 44 (47.3%) 74 (63.2%) 49 (41.9%) 55 (46.2%) 76 (63.3%) 41 (37.6%) 71 (55.0%) 34 (26.4%) 0.02 0.84 0.007 0.02 0.001 2.87 (1.20–6.87) 1.09 (0.45–2.68) 2.77 (1.32–5.82) 3.91 (1.29–11.82) 2.65 (1.46–4.82) 6 (35.3%) 13 (61.9%) 5 (29.4%) 10 (45.5%) 10 (45.5%) 92 (64.3%) 94 (64.4%) 72 (57.1%) 112 (75.7%) 102 (68.9%) Cross-sectional study HbA1c ≤ 7.0% (53 mmol/mol) Blood pressure ≤130/80 LDL-cholesterol ≤2.0 mmol/L Microalbuminuria test Foot examination
0.77 0.002 0.01 0.99 1.03 (0.86–1.23) 0.87 (0.80–0.95) 0.85 (0.74–0.97) 1.00 (0.89–1.13) 12,741 (87.9%) 7625 (53.8%) 6214 (43.8%) 8710 (61.4%) 10,157 (88.8%) 5765 (50.4%) 4286 (37.5%) 6941 (60.7%) <0.001 0.02 0.03 0.008 1.33 (1.14–1.56) 0.91 (0.83–0.99) 1.15 (1.02–1.29) 1.19 (1.05–1.36) 5532 (85.3%) 3610 (55.7%) 2639 (40.7%) 3746 (57.8%) 15,439 (88.8%) 9417 (54.2%) 7610 (43.8%) 10,919 (62.8%) Cohort study Family physician visit Retinal screening ≥4 HbA1c tests ≥2 cholesterol tests
Adjusted ORa (95% CI) Non-South Asian physician Chinese physician
Non-Chinese physician
Adjusted ORa (95% CI)
P
South Asian physician
South Asian patients Chinese patients Quality of care measure
Table 2 – Quality of care measures for Chinese and South Asian patients with diabetes, stratified by the ethnicity of their family physician.
P
p r i m a r y c a r e d i a b e t e s 9 ( 2 0 1 5 ) 432–438
among South Asian patients, having a physician from the same ethnic group had no influence on most quality of care measures in either study. Of note, the study did not explore the mechanistic reasons why quality of care may have differed by physician ethnicity, and clearly both patient and provider factors will influence care. The discordance in results between the two ethnic groups suggests that a shared ethnic origin itself does not drive quality of care for minority patients, but instead ethnic origin may be a proxy for other factors that, when shared between patient and physician, lead to better care. In Ontario, the Chinese population is more homogeneous than the South Asian population with respect to country of origin, mother tongue, religious background and diet; sharing these factors with their family physicians may contribute more directly to quality of care for minority patients than sharing their ethnic origin itself. For example, the predominant mother tongue for Chinese people in Ontario is Cantonese (and, to a much lesser extent, Mandarin), whereas common South Asian languages in Ontario include Punjabi, Urdu, Tamil, Gujarati and Hindi [26]. This heterogeneity among South Asians means that although South Asian physicians will share the same ethnic group with South Asian patients, they will not necessarily share the same language or culture. These studies have several strengths to highlight. First, they examined physician ethnicity in Chinese and South Asian populations, two minority groups that have not been studied previously. Second, the cohort study used health care administrative databases that covered an entire population, so there was no selection bias or loss to follow-up. Third, the singlepayer universal health care system in Ontario ensured that no differences in insurance coverage or health care access could impact quality of care. Finally, all participating physicians in the cross-sectional study were drawn from the same neighborhoods with large minority populations, limiting the likelihood that geographic or access factors could have contributed to differences in care. However, there are some limitations to note. First, we identified ethnic minority patients in the cohort study using surnames lists that have been validated within this population [22]. However, surnames that were not unique to these ethnic origins were excluded from the lists in order to maximize positive predictive value, and therefore the lists had lower sensitivity. While this low sensitivity may make the studied populations less generalizable to the entire Chinese or South Asian population of Ontario, the studies’ internal validity remains strong. We also used these surnames lists to identify ethnic minority physicians. The low sensitivity could result in the misclassification of ethnically concordant physicians as discordant, which would tend to narrow any observed differences, thereby making the observed statistically significant results even more striking. Second, we did not have any information on generation status or time since immigration, which might influence patient–physician relationships and therefore quality of care. However, given the mean age of the patients in these studies (in their 50s in the cohort study, and up to their 60s in the cross-sectional study), we anticipate that virtually all are first-generation immigrants since the Asian population of Ontario in the 1950s and earlier was tiny. Third, other confounding factors may contribute to the observed relationships. For example, if Chinese patients who
p r i m a r y c a r e d i a b e t e s 9 ( 2 0 1 5 ) 432–438
are more engaged in their health care or more adherent to treatment recommendations preferentially seek out Chinese family physicians, the observed association with quality of care may not be due to the family physician’s ethnicity per se. Fourth, we had no information on other health care providers working in family physicians’ offices, who may have been from the patients’ ethnic group even if the physicians themselves were not. However, in the context of the Canadian health care system, most primary health care services would be delivered by the assigned family physician him or herself, rather than by anyone else in the practice. Fifth, the cross-sectional study specifically recruited physicians from neighborhoods with large minority populations; hence, the findings from this study are not generalizable to the entire Chinese or South Asian population. However, the findings mirror those of the cohort study, which was population-based using unselected patients, and which is therefore highly generalizable. Finally, our recruitment of non-Chinese physicians practicing in Chinese neighborhoods in the cross-sectional study fell below our targets, which could have impacted the study’s power. However, despite this, most of the findings for Chinese patients in this study remained statistically significant. The previous literature on the impact of physician ethnicity on quality of chronic disease care for minority patients has demonstrated mixed results. African-Americans with HIV received protease inhibitors sooner if they had AfricanAmerican physicians [8], whereas quality of hypertension care for African-Americans was not influenced by physician ethnicity [27]. Alcohol interventions were more effective for Hispanic patients with Hispanic physicians [28]. Among patients with diabetes specifically, there were virtually no differences by physician ethnicity in the proportion of patients achieving glycemic, blood pressure or lipid control among African-American and Hispanic patients in California [10]. Sharing their physician’s ethnicity increased adherence to cardiovascular disease medications for African-American patients with diabetes, but not Hispanic or Asian patients [9]. However, a small study found that the arrival of a Russian-trained internist improved glycemic, blood pressure and cholesterol control for ethnically Russian patients with diabetes in Denver, Colorado [29]. There is no previous literature examining the impact of physician ethnicity on the care of Chinese or South Asian patients, and no studies from outside the United States. There are many hypothetical mechanisms through which having a physician from the same ethnic group might improve care for minority patients [30]. Physicians from the same ethnic group may have a greater familiarity with the patient’s dietary practices, alternative medicine preferences, and behavioral-cultural norms. They may be more likely to be able to communicate with patients in their language, and may have a greater understanding of patients’ health literacy. Furthermore, minority patients themselves may have greater confidence and trust in physicians who share their ethnic background, and so may be more likely to be adherent to their recommendations. However, these hypothetical factors do not appear to influence care for South Asian patients, as their care did not differ by physician ethnicity. In summary, in both the cohort and cross-sectional studies, Chinese-Canadian patients received better quality of diabetes care if they were treated by a family physician who was also
437
of Chinese ethnicity. They were much more likely to achieve glycemic and cholesterol control targets, and processes of care such as appropriate laboratory testing were also more common. However, physician ethnicity had little influence on quality of diabetes care for South Asian-Canadian patients. Thus, sharing ethnicity with their physician – by itself – is not predictive of better diabetes care for minority patients, since the relationship was inconsistent between Chinese and South Asian populations. Instead, concordance on factors associated with ethnicity (such as language, religion, country of origin, time since immigration, or other more specific measures of culture) may instead drive these relationships, and ought to be investigated in the future to better understand strategies to improve care for minority patients.
Conflict of interest The authors declare that they have no conflicts of interest.
Acknowledgements We would like to thank Karen Cauch-Dudek, Charles Victor and Ryan Ng for their assistance with the acquisition of data and the analyses. The studies were funded by an operating grant from the Heart and Stroke Foundation of Canada, grant number NA6330. BRS receives salary support from the Canadian Institutes of Health Research, and was previously supported by the Canadian Diabetes Association. The Institute for Clinical Evaluative Sciences (ICES) is a non-profit research institute funded by the Ontario Ministry of Health and LongTerm Care (MOHLTC). The opinions, results and conclusions reported in these studies are those of the authors and are independent from the funding sources. No endorsement by the funders, ICES or the MOHLTC is intended or should be inferred.
references
[1] E.C. Schneider, A.M. Zaslavsky, A.M. Epstein, Racial disparities in the quality of care for enrollees in Medicare managed care, JAMA 287 (2002) 1288–1294. [2] B.D. Smedley, A.Y. Stith, A.R. Nelson, Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care, National Academies Press, Washington, DC, 2003. [3] A.N. Trivedi, A.M. Zaslavsky, E.C. Schneider, et al., Trends in the quality of care and racial disparities in Medicare managed care, N. Engl. J. Med. 353 (2005) 692–700. [4] S.H. Meghani, J.M. Brooks, T. Gipson-Jones, et al., Patient–provider race-concordance: does it matter in improving minority patients’ health outcomes? Ethnic. Health 14 (2009) 107–130. [5] T.A. LaVeist, A. Nuru-Jeter, Is doctor–patient race concordance associated with greater satisfaction with care? J. Health Soc. Behav. 43 (2002) 296–306. [6] S. Saha, M. Komaromy, T.D. Koepsell, et al., Patient–physician racial concordance and the perceived quality and use of health care, Arch. Intern. Med. 159 (1999) 997–1004. [7] T.A. LaVeist, A. Nuru-Jeter, K.E. Jones, The association of doctor–patient race concordance with health service utilization, J. Public Health Policy 24 (2003) 312–323.
438
p r i m a r y c a r e d i a b e t e s 9 ( 2 0 1 5 ) 432–438
[8] W.D. King, M.D. Wong, M.F. Shapiro, et al., Does racial concordance between HIV-positive patients and their physicians affect the time to receipt of protease inhibitors? J. Gen. Intern. Med. 19 (2004) 1146–1153. [9] A.H. Traylor, J.A. Schmittdiel, C.S. Uratsu, et al., Adherence to cardiovascular disease medications: does patient–provider race/ethnicity and language concordance matter? J. Gen. Intern. Med. 25 (2010) 1172–1177. [10] A.H. Traylor, U. Subramanian, C.S. Uratsu, et al., Patient race/ethnicity and patient–physician race/ethnicity concordance in the management of cardiovascular disease risk factors for patients with diabetes, Diabetes Care 33 (2010) 520–525. [11] Y.J. Cheng, G. Imperatore, L.S. Geiss, et al., Secular changes in the age-specific prevalence of diabetes among U.S. adults: 1988–2010, Diabetes Care 36 (2013) 2690–2696. [12] A.L. Stewart, S. Greenfield, R.D. Hays, et al., Functional status and well-being of patients with chronic conditions. Results from the Medical Outcomes Study, JAMA 262 (1989) 907–913. [13] E.W. Gregg, Q. Gu, Y.J. Cheng, et al., Mortality trends in men and women with diabetes, 1971 to 2000, Ann. Intern. Med. 147 (2007) 149–155. [14] J.B. Saaddine, B. Cadwell, E.W. Gregg, et al., Improvements in diabetes processes of care and intermediate outcomes: United States, 1988–2002, Ann. Intern. Med. 144 (2006) 465–474. [15] M.K. Ali, K. McKeever Bullard, J.B. Saaddine, et al., Achievement of goals in US diabetes care, 1999–2010, N. Engl. J. Med. 368 (2013) 1613–1624. [16] B.R. Shah, J.E. Hux, A. Laupacis, et al., Deficiencies in the quality of diabetes care: comparing specialist with generalist care misses the point, J. Gen. Intern. Med. 22 (2007) 275–279. [17] A.F. Brown, E.W. Gregg, M.R. Stevens, et al., Race, ethnicity, socioeconomic position, and quality of care for adults with diabetes enrolled in managed care: the Translating Research Into Action for Diabetes (TRIAD) study, Diabetes Care 28 (2005) 2864–2870. [18] M.H. Chin, J.X. Zhang, K. Merrell, Diabetes in the African-American Medicare population: morbidity, quality of care, and resource utilization, Diabetes Care 21 (1998) 1090–1095. [19] J.M. McWilliams, E. Meara, A.M. Zaslavsky, et al., Differences in control of cardiovascular disease and diabetes by race,
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
ethnicity, and education: U.S. trends from 1999 to 2006 and effects of Medicare coverage, Ann. Intern. Med. 150 (2009) 505–515. B.R. Shah, K. Cauch-Dudek, S.S. Anand, et al., Absence of disparities in the quality of primary diabetes care for South Asians and Chinese in an urban Canadian setting, Diabetes Care 35 (2012) 794–796. J.E. Hux, F. Ivis, V. Flintoft, et al., Diabetes in Ontario: determination of prevalence and incidence using a validated administrative data algorithm, Diabetes Care 25 (2002) 512–516. B.R. Shah, M. Chiu, S. Amin, et al., Surname lists to identify South Asian and Chinese ethnicity from secondary data in Ontario, Canada: a validation study, BMC Med. Res. Methodol. 10 (2010) 42. T. Kiran, J.C. Victor, A. Kopp, et al., The relationship between financial incentives and quality of diabetes care in Ontario, Canada, Diabetes Care 35 (2012) 1038–1046. Canadian Diabetes Association Clinical Practice Guidelines Expert Committee, Canadian Diabetes Association 2008 clinical practice guidelines for the prevention and management of diabetes in Canada, Can. J. Diabetes 32 (2008) S1–S201. K. Tu, N.R.C. Campbell, X.-L. Chen, et al., Accuracy of administrative databases in identifying patients with hypertension, Open Med. 1 (2007) 18–26. Statistics Canada Language Highlight Tables, 2006 Census. 97-555-XWE2006002, Statistics Canada, Ottawa, 2007, Ref Type: Report. T.R. Konrad, D.L. Howard, L.J. Edwards, et al., Physician–patient racial concordance, continuity of care, and patterns of care for hypertension, Am. J. Public Health 95 (2005) 2186–2190. C. Field, R. Caetano, The role of ethnic matching between patient and provider on the effectiveness of brief alcohol interventions with Hispanics, Alcohol. Clin. Exp. Res. 34 (2010) 262–271. P.S. Mehler, R.A. Lundgren, I. Pines, et al., A community study of language concordance in Russian patients with diabetes, Ethnic. Dis. 14 (2004) 584–588. M. van Ryn, Research on the provider contribution to race/ethnicity disparities in medical care, Med. Care 40 (2002) I140–I151.
Contents lists available at ScienceDirect
Primary Care Diabetes journal homepage: http://www.elsevier.com/locate/pcd
Original research
Family physician ethnicity influences quality of diabetes care for Chinese but not South Asian patients Baiju R. Shah a,b,c,∗ , Jeremiah Hwee a , Sonia S. Anand d , Peter C. Austin a,b , Douglas G. Manuel a,e , Janet E. Hux a,1 a
Institute for Clinical Evaluative Sciences, Toronto, ON, Canada University of Toronto, Toronto, ON, Canada c Department of Medicine, Sunnybrook Health Sciences Centre, Toronto, ON, Canada d McMaster University, Hamilton, ON, Canada e University of Ottawa, Ottawa, ON, Canada b
a r t i c l e
i n f o
a b s t r a c t
Article history:
Aims: To determine whether sharing the same ethnicity as their family physician influenced
Received 9 July 2014
the quality of diabetes care for Chinese and South Asian patients in Ontario, Canada.
Received in revised form
Methods: We conducted two related studies: a population-based cohort study of Chinese
3 October 2014
and South Asian patients with incident diabetes using health care administrative data
Accepted 10 February 2015
(n = 49,484), and a cross-sectional study of Chinese and South Asian patients with estab-
Available online 8 March 2015
lished diabetes using data collected directly from their family physicians’ clinical records (n = 416). In both studies, quality of care measures were compared between patients whose
Keywords:
family physicians were or were not from the same ethnic group.
Ethnicity
Results: In the cohort study, Chinese patients whose family physicians were also Chinese
Chinese
were more likely to have a diabetes-related family physician visit and appropriate HbA1c
South Asian
and cholesterol testing. In the cross-sectional study, they were more likely to have foot
Quality of care
examinations, to have microalbuminuria testing, and to achieve recommended treatment
Family physicians
targets for HbA1c and for LDL-cholesterol. In contrast, for South Asian patients, most quality measures in either study did not differ by physician ethnicity. Conclusions: Having a family physician from the same ethnic group was associated with better quality of diabetes care for Chinese but not for South Asian patients. © 2015 Primary Care Diabetes Europe. Published by Elsevier Ltd. All rights reserved.
∗
1
Corresponding author at: G106 – 2075 Bayview Avenue, Toronto, ON, Canada M4N 3M5. Tel.: +1 416 480 4706; fax: +1 416 480 6048. E-mail address: [email protected] (B.R. Shah). Present address: Canadian Diabetes Association, Toronto, Ontario, Canada.
http://dx.doi.org/10.1016/j.pcd.2015.02.002 1751-9918/© 2015 Primary Care Diabetes Europe. Published by Elsevier Ltd. All rights reserved.
p r i m a r y c a r e d i a b e t e s 9 ( 2 0 1 5 ) 432–438
1.
Introduction
Disparities in chronic disease care for ethnic minority populations are well documented [1–3], but the modifiers of these disparities are uncertain. For example, the ethnicity of their primary care physician may influence the quality of care received by minority patients [4]. Physicians from the same ethnic group may be able to overcome language barriers to care, and may have a better understanding of the patient’s dietary practices or cultural norms. American studies have shown that minority patients with physicians report greater satisfaction and fewer unmet care needs with physicians from the same ethnic group [5,6]. This literature has also suggested better access to and utilization of care, and improved quality of care [7–9]. However, a study specifically of patients with diabetes found that physician ethnicity had no impact on glycemic, blood pressure or lipid control for African-American and Hispanic patients [10]. There are no studies on the impact of physician ethnicity for minority patients from other countries, and limited data for other ethnic groups. Diabetes, an archetypal and increasingly prevalent chronic disease [11], leads to reduced quality of life and premature mortality [12,13]. Research in many jurisdictions has found that quality of care at a population level falls short of recommended targets [14–16]. Ethnic minorities may be particularly vulnerable [17–19], although in our jurisdiction, minorities receive similar quality of care to the general population, albeit still below recommended goals [20]. We conducted two related studies with an objective to determine whether sharing the same ethnic origin as their family physician influenced the quality of diabetes care for Chinese and South Asian patients in Ontario, Canada.
2.
Methods
2.1.
Design, data sources and patient selection
databases and across time through unique but anonymous numbers. Using the ODD, all individuals in Ontario newly diagnosed with diabetes between January 1, 2000 and December 31, 2008 were identified. Inclusion criteria were age ≥20 years, eligibility for provincial health insurance for ≥2 years (to ensure that new immigrants with prevalent diabetes were not inadvertently included as incident cases due to a lack of prior data), and Chinese or South Asian ethnicity. Because Canadian health care data do not contain ethnic identifiers, we identified Chinese and South Asian patients by identifying all those whose earliest known surnames (i.e., before name changes through marriage) matched two lists of surnames validated in this population to have excellent positive predictive values when compared to self-reported ethnicity (91.9% for Chinese and 89.3% for South Asian) [22]. In Ontario, virtually all primary care is delivered by family physicians. Each patient was assigned to their regular family physician based on ambulatory billing claims [23]; those patients who could not be assigned to a family physician were excluded. Although population-based health administrative data are comprehensive for examining health service utilization, they lack the detailed clinical information needed to more broadly assess quality of care. To overcome this limitation, we also conducted a cross-sectional study of clinical quality of care measures by reviewing patients’ clinical records in their family physicians’ offices. We enrolled randomly selected family physicians practicing in neighborhoods where ≥25% of the population reported either Chinese or South Asian ethnicity in the 2006 Canadian census. (These neighborhoods are all in Toronto or its suburbs.) Enrolment was stratified by selfreported physician ethnicity. In each participating physician’s practice, we randomly selected up to 10 patients who were age ≥18 years, had type 2 diabetes for ≥2 years, and had Chinese or South Asian ethnicity (based on recorded ethnicity in the clinical record, or reported by the physician). Details of the abstraction are presented elsewhere [20].
2.2. We conducted these studies in Ontario, Canada’s most populous province, using different study designs, data sources, patient populations and outcome measures. First, we conducted a population-based cohort study. The study used health care databases from the government-funded health insurance program of the Ontario Ministry of Health and Long-Term Care, which provides coverage to all permanent residents of Ontario. The data used in the study included demographic information for all Ontario residents, demographic and practice information for all licensed physicians in Ontario, claims data from physicians and clinical laboratories for fee-for-service reimbursement, and abstracts of all hospitalizations and emergency department visits. We also used the Ontario Diabetes Database (ODD), a registry derived from administrative data that identifies all people with physician-diagnosed non-gestational diabetes in Ontario. When validated against chart review, the ODD was found to have a sensitivity of 86% and a specificity of 97% [21]. Although the database does not distinguish between types of diabetes, the vast majority of cases would have type 2 diabetes. Individual patients and physicians are linked between all the
433
Exposure and quality of care measures
In the cohort study, physicians were identified as Chinese or South Asian based on their surnames as noted above. To improve sensitivity in identifying ethnic minority physicians, we also included as Chinese or South Asian those who had graduated from a medical school in (respectively) China, Taiwan or Hong Kong; or India, Pakistan, Bangladesh or Sri Lanka. This additional surrogate information for ethnicity was available through the physician registration data. In the crosssectional study, physician ethnicity was self-reported. Patients in the cohort study were followed using the claims data for 2 years after diabetes diagnosis to determine the following process measures of quality of care, derived from recommendations in national diabetes clinical practice guidelines [24]: at least one family physician claim for an ambulatory diabetes-related care visit, at least one ophthalmologist or optometrist claim for a retinal screening visit, at least four laboratory claims for HbA1c tests, and at least two laboratory claims for cholesterol tests. The clinical quality of care measures evaluated in the cross-sectional study were glycemic control (most recent HbA1c ≤7.0% [53 mmol/mol]),
434
p r i m a r y c a r e d i a b e t e s 9 ( 2 0 1 5 ) 432–438
714,948 people aged ≥20 years diagnosed with diabetes in Ontario, 2000 to 2008
27,444 Chinese adults with newly-diagnosed diabetes
33,103 South Asian adults with newly-diagnosed diabetes
3,399 with health insurance coverage for <2 years
7,234 with health insurance coverage for <2 years
180 with no family physician
250 with no family physician
23,865 Chinese patients included in the cohort study 17,381 with Chinese physician
6,484 with nonChinese physician
25,619 South Asian patients included in the cohort study 11,434 with South Asian physician
14,185 with nonSouth Asian physician
Fig. 1 – Patients included in the cohort study.
blood pressure control (most recent ≤130/80), cholesterol control (most recent LDL-cholesterol ≤2.0 mmol/L), at least one microalbuminuria test documented in the preceding year, and at least one foot examination documented in the preceding year.
3.
Analyses
In both studies, family physician ethnicity was analyzed separately for Chinese patients (Chinese family physicians versus others) and South Asian patients (South Asian family physicians versus others). Baseline characteristics were compared using t-tests or chi-square tests. For each quality of care measure, we estimated logistic regression models to determine the independent, adjusted effect of physician ethnicity. The models were estimated using generalized estimating equation methods to account for the clustering of patients within physicians. Models adjusted for patient age, sex, socioeconomic status (defined based on neighborhood median household income, divided into quintiles), rural residence (based on the patient’s home postal code), prior history of coronary artery disease, heart failure or stroke (based on hospitalization records from the previous 5 years), and hypertension (using a validated administrative data-derived algorithm) [25]. In the cohort study, the models were additionally adjusted for the following physician variables: age, sex, number of years in practice, and whether or not the physician was a Canadian medical graduate. The studies were approved by the research ethics board of Sunnybrook Health Sciences Centre, Toronto.
4.
Results
4.1.
Patient and physician characteristics
In the cohort study, there were 714,948 people aged ≥20 years diagnosed with diabetes in Ontario, 60,547 of whom were
Chinese or South Asian. After excluding 10,633 (17.6%) who had health insurance coverage for <2 years and 430 (0.7%) who had no family physician, the final cohort included 23,865 Chinese patients and 25,619 South Asian patients (Fig. 1). Seventy-three percent of Chinese patients had a Chinese physician, while only 45 percent of South Asian patients had a South Asian physician. At baseline, both Chinese and South Asian patients with physicians of the same ethnicity were older, had lower socioeconomic status and were less likely to live in rural areas than those with discordant physicians (Table 1). For the cross-sectional study, we recruited 26 family physicians practicing in neighborhoods with large South Asian populations (13 of whom were themselves South Asian), and 19 physicians in neighborhoods with large Chinese populations (15 of whom were themselves Chinese). We were unable to find more non-Chinese physicians practicing in these neighborhoods who had sufficient numbers of Chinese patients with diabetes in their practice. The baseline characteristics of the randomly selected eligible patients who were included in the cross-sectional study are shown in Table 1.
4.2.
Quality of care measures
Quality of care is presented in Table 2, with adjusted odds ratios showing the effect of family physician ethnicity independent of baseline differences. Among Chinese patients, most quality of care measures were better in those who had a Chinese family physician. In the cohort study, they were more likely to have a diabetes-related family physician visit, appropriate HbA1c testing and appropriate cholesterol testing. In the cross-sectional study, they were more likely to have had microalbuminuria testing and foot examination. They were also markedly more likely to have met targets for glycemic control (64.3% vs 35.3%, OR 2.87, 95% CI 1.20–6.87) and cholesterol control (57.1% vs 29.4%, OR 2.77, 95% CI 1.32–5.82). However, they were marginally less likely to receive a retinal screening examination. In contrast,
435
p r i m a r y c a r e d i a b e t e s 9 ( 2 0 1 5 ) 432–438
Table 1 – Baseline characteristics of Chinese and South Asian patients in the cohort and cross-sectional studies, stratified by the ethnicity of their family physician. Characteristic
Chinese patients Chinese physician
Cohort study Patients N Age, mean ± SD Sex Male Female Socioeconomic status Lowest 2 3 4 Highest Rural residence Previous CAD Previous heart failure Previous stroke Previous hypertension Physicians N Year of birth ≤1949 1950–1959 1960–1969 ≥1970 Sex Male Female Year of starting practice ≤1974 1975–1984 1985–1994 ≥1995 Canadian graduate Cross-sectional study Patients N Age, mean ± SD Sex Male Female Socioeconomic status Lowest 2 3 4 Highest Previous CAD, heart failure or stroke Previous hypertension a
Non-Chinese physician
17,381 57.3 ± 14.3
6484 55.2 ± 14.4
9314 (53.6%) 8067 (46.4%)
3440 (53.1%) 3044 (46.9%)
3576 (20.6%) 4482 (25.8%) 3511 (20.2%) 3349 (19.3%) 2440 (14.0%) 29 (0.2%) 216 (1.2%) 105 (0.6%) 121 (0.7%) 8228 (47.3%)
1360 (21.0%) 1387 (21.4%) 1269 (19.6%) 1236 (19.1%) 1215 (18.7%) 93 (1.4%) 110 (1.7%) 57 (0.9%) 50 (0.8%) 3003 (46.3%)
South Asian patients P
<0.001 0.5
South Asian physician
Non-South Asian physician
11,434 51.8 ± 13.7
14,185 51.3 ± 13.3
6222 (54.4%) 5212 (45.6%)
7997 (56.4%) 6188 (43.6%)
3370 (29.5%) 3172 (27.7%) 2557 (22.4%) 1514 (13.2%) 809 (7.1%) 20 (0.2%) 295 (2.6%) 70 (0.6%) 48 (0.4%) 4764 (41.7%)
3929 (27.7%) 3321 (23.4%) 3016 (21.3%) 2180 (15.4%) 1720 (12.1%) 74 (0.5%) 438 (3.1%) 98 (0.7%) 83 (0.6%) 5782 (40.8%)
<0.001
619
<0.001 0.007 0.02 0.5 0.2
2539 832 (32.8%) 890 (35.1%) 646 (25.4%) 171 (6.7%)
470 (75.9%) 149 (24.1%)
1698 (66.9%) 841 (33.1%)
148 (23.9%) 218 (35.2%) 188 (30.4%) 65 (10.5%) 436 (70.4%)
750 (29.5%) 828 (32.6%) 716 (28.2%) 245 (9.6%) 1808 (71.2%)
473
3452
150 (31.7%) 133 (28.1%) 125 (26.4%) 65 (13.7%)
1092 (31.6%) 1117 (32.4%) 940 (27.2%) 303 (8.8%)
264 (55.8%) 209 (44.2%)
245 (71.0%) 1001 (29.0%)
142 (30.0%) 137 (29.0%) 122 (25.8%) 72 (15.2%) 118 (24.9%)
981 (28.4%) 1074 (31.1%) 997 (28.9%) 400 (11.6%) 2462 (71.3%)
129 60.8 ± 10.9
117 55.5 ± 9.6
<0.001
0.051
148 66.3 ± 13.3
22 61.1 ± 13.1
82 (55.4%) 66 (44.6%)
13 (59.1%) 9 (40.9%)
18 (12.2%) 40 (27.0%) 33 (22.3%) 31 (20.9%) 26 (17.6%) 7 (4.7%) 112 (75.7%)
7 (31.8%) 5 (22.7%)
0.08 0.7
0.07
66 (51.2%) 63 (48.8%)
78 (66.7%) 39 (33.3%)
32 (24.8%) 25 (19.4%) 27 (20.9%) 30 (23.3%) 15 (11.6%) 11 (8.5%) 90 (69.8%)
31 (26.5%) 35 (29.9%) 30 (25.6%) 11 (9.4%) 10 (8.5%) 9 (7.7%) 73 (62.4%)
0.1
a a a a
0.9 0.8
16 (72.7%)
<0.001 0.02 0.4 0.06 0.1
0.004
<0.001
0.7
<0.001 0.002
<0.001
0.01 200 (32.3%) 181 (29.2%) 187 (30.2%) 51 (8.2%)
P
<0.001
0.003 0.01
0.04
0.8 0.2
Cell sizes <5 are not reported to protect confidentiality of patients.
South Asian patients who had South Asian family physicians were less likely to have appropriate HbA1c testing or a retinal screening examination in the cohort study. Other measures of quality of care in the cohort study, and all the measures examined in the cross-sectional study, were not associated with family physician ethnicity for South Asian patients.
5.
Discussion
In both the cohort and cross-sectional studies, Chinese patients with diabetes whose family physicians were also Chinese received better quality of care and were markedly more likely to achieve glycemic and cholesterol targets. In contrast,
436
Odds ratios are adjusted for patient age, sex, socioeconomic status, rural residence, prior history of CAD, heart failure or stroke, and hypertension. In the cohort study, odds ratios were additionally adjusted for physician age, sex, years in practice, and graduation from a Canadian vs foreign medical school. a
0.91 0.62 0.28 0.53 0.13 0.97 (0.53–1.77) 0.85 (0.45–1.60) 0.70 (0.37–1.34) 0.73 (0.27–1.96) 0.49 (0.20–1.24) 47 (42.7%) 72 (64.9%) 44 (47.3%) 74 (63.2%) 49 (41.9%) 55 (46.2%) 76 (63.3%) 41 (37.6%) 71 (55.0%) 34 (26.4%) 0.02 0.84 0.007 0.02 0.001 2.87 (1.20–6.87) 1.09 (0.45–2.68) 2.77 (1.32–5.82) 3.91 (1.29–11.82) 2.65 (1.46–4.82) 6 (35.3%) 13 (61.9%) 5 (29.4%) 10 (45.5%) 10 (45.5%) 92 (64.3%) 94 (64.4%) 72 (57.1%) 112 (75.7%) 102 (68.9%) Cross-sectional study HbA1c ≤ 7.0% (53 mmol/mol) Blood pressure ≤130/80 LDL-cholesterol ≤2.0 mmol/L Microalbuminuria test Foot examination
0.77 0.002 0.01 0.99 1.03 (0.86–1.23) 0.87 (0.80–0.95) 0.85 (0.74–0.97) 1.00 (0.89–1.13) 12,741 (87.9%) 7625 (53.8%) 6214 (43.8%) 8710 (61.4%) 10,157 (88.8%) 5765 (50.4%) 4286 (37.5%) 6941 (60.7%) <0.001 0.02 0.03 0.008 1.33 (1.14–1.56) 0.91 (0.83–0.99) 1.15 (1.02–1.29) 1.19 (1.05–1.36) 5532 (85.3%) 3610 (55.7%) 2639 (40.7%) 3746 (57.8%) 15,439 (88.8%) 9417 (54.2%) 7610 (43.8%) 10,919 (62.8%) Cohort study Family physician visit Retinal screening ≥4 HbA1c tests ≥2 cholesterol tests
Adjusted ORa (95% CI) Non-South Asian physician Chinese physician
Non-Chinese physician
Adjusted ORa (95% CI)
P
South Asian physician
South Asian patients Chinese patients Quality of care measure
Table 2 – Quality of care measures for Chinese and South Asian patients with diabetes, stratified by the ethnicity of their family physician.
P
p r i m a r y c a r e d i a b e t e s 9 ( 2 0 1 5 ) 432–438
among South Asian patients, having a physician from the same ethnic group had no influence on most quality of care measures in either study. Of note, the study did not explore the mechanistic reasons why quality of care may have differed by physician ethnicity, and clearly both patient and provider factors will influence care. The discordance in results between the two ethnic groups suggests that a shared ethnic origin itself does not drive quality of care for minority patients, but instead ethnic origin may be a proxy for other factors that, when shared between patient and physician, lead to better care. In Ontario, the Chinese population is more homogeneous than the South Asian population with respect to country of origin, mother tongue, religious background and diet; sharing these factors with their family physicians may contribute more directly to quality of care for minority patients than sharing their ethnic origin itself. For example, the predominant mother tongue for Chinese people in Ontario is Cantonese (and, to a much lesser extent, Mandarin), whereas common South Asian languages in Ontario include Punjabi, Urdu, Tamil, Gujarati and Hindi [26]. This heterogeneity among South Asians means that although South Asian physicians will share the same ethnic group with South Asian patients, they will not necessarily share the same language or culture. These studies have several strengths to highlight. First, they examined physician ethnicity in Chinese and South Asian populations, two minority groups that have not been studied previously. Second, the cohort study used health care administrative databases that covered an entire population, so there was no selection bias or loss to follow-up. Third, the singlepayer universal health care system in Ontario ensured that no differences in insurance coverage or health care access could impact quality of care. Finally, all participating physicians in the cross-sectional study were drawn from the same neighborhoods with large minority populations, limiting the likelihood that geographic or access factors could have contributed to differences in care. However, there are some limitations to note. First, we identified ethnic minority patients in the cohort study using surnames lists that have been validated within this population [22]. However, surnames that were not unique to these ethnic origins were excluded from the lists in order to maximize positive predictive value, and therefore the lists had lower sensitivity. While this low sensitivity may make the studied populations less generalizable to the entire Chinese or South Asian population of Ontario, the studies’ internal validity remains strong. We also used these surnames lists to identify ethnic minority physicians. The low sensitivity could result in the misclassification of ethnically concordant physicians as discordant, which would tend to narrow any observed differences, thereby making the observed statistically significant results even more striking. Second, we did not have any information on generation status or time since immigration, which might influence patient–physician relationships and therefore quality of care. However, given the mean age of the patients in these studies (in their 50s in the cohort study, and up to their 60s in the cross-sectional study), we anticipate that virtually all are first-generation immigrants since the Asian population of Ontario in the 1950s and earlier was tiny. Third, other confounding factors may contribute to the observed relationships. For example, if Chinese patients who
p r i m a r y c a r e d i a b e t e s 9 ( 2 0 1 5 ) 432–438
are more engaged in their health care or more adherent to treatment recommendations preferentially seek out Chinese family physicians, the observed association with quality of care may not be due to the family physician’s ethnicity per se. Fourth, we had no information on other health care providers working in family physicians’ offices, who may have been from the patients’ ethnic group even if the physicians themselves were not. However, in the context of the Canadian health care system, most primary health care services would be delivered by the assigned family physician him or herself, rather than by anyone else in the practice. Fifth, the cross-sectional study specifically recruited physicians from neighborhoods with large minority populations; hence, the findings from this study are not generalizable to the entire Chinese or South Asian population. However, the findings mirror those of the cohort study, which was population-based using unselected patients, and which is therefore highly generalizable. Finally, our recruitment of non-Chinese physicians practicing in Chinese neighborhoods in the cross-sectional study fell below our targets, which could have impacted the study’s power. However, despite this, most of the findings for Chinese patients in this study remained statistically significant. The previous literature on the impact of physician ethnicity on quality of chronic disease care for minority patients has demonstrated mixed results. African-Americans with HIV received protease inhibitors sooner if they had AfricanAmerican physicians [8], whereas quality of hypertension care for African-Americans was not influenced by physician ethnicity [27]. Alcohol interventions were more effective for Hispanic patients with Hispanic physicians [28]. Among patients with diabetes specifically, there were virtually no differences by physician ethnicity in the proportion of patients achieving glycemic, blood pressure or lipid control among African-American and Hispanic patients in California [10]. Sharing their physician’s ethnicity increased adherence to cardiovascular disease medications for African-American patients with diabetes, but not Hispanic or Asian patients [9]. However, a small study found that the arrival of a Russian-trained internist improved glycemic, blood pressure and cholesterol control for ethnically Russian patients with diabetes in Denver, Colorado [29]. There is no previous literature examining the impact of physician ethnicity on the care of Chinese or South Asian patients, and no studies from outside the United States. There are many hypothetical mechanisms through which having a physician from the same ethnic group might improve care for minority patients [30]. Physicians from the same ethnic group may have a greater familiarity with the patient’s dietary practices, alternative medicine preferences, and behavioral-cultural norms. They may be more likely to be able to communicate with patients in their language, and may have a greater understanding of patients’ health literacy. Furthermore, minority patients themselves may have greater confidence and trust in physicians who share their ethnic background, and so may be more likely to be adherent to their recommendations. However, these hypothetical factors do not appear to influence care for South Asian patients, as their care did not differ by physician ethnicity. In summary, in both the cohort and cross-sectional studies, Chinese-Canadian patients received better quality of diabetes care if they were treated by a family physician who was also
437
of Chinese ethnicity. They were much more likely to achieve glycemic and cholesterol control targets, and processes of care such as appropriate laboratory testing were also more common. However, physician ethnicity had little influence on quality of diabetes care for South Asian-Canadian patients. Thus, sharing ethnicity with their physician – by itself – is not predictive of better diabetes care for minority patients, since the relationship was inconsistent between Chinese and South Asian populations. Instead, concordance on factors associated with ethnicity (such as language, religion, country of origin, time since immigration, or other more specific measures of culture) may instead drive these relationships, and ought to be investigated in the future to better understand strategies to improve care for minority patients.
Conflict of interest The authors declare that they have no conflicts of interest.
Acknowledgements We would like to thank Karen Cauch-Dudek, Charles Victor and Ryan Ng for their assistance with the acquisition of data and the analyses. The studies were funded by an operating grant from the Heart and Stroke Foundation of Canada, grant number NA6330. BRS receives salary support from the Canadian Institutes of Health Research, and was previously supported by the Canadian Diabetes Association. The Institute for Clinical Evaluative Sciences (ICES) is a non-profit research institute funded by the Ontario Ministry of Health and LongTerm Care (MOHLTC). The opinions, results and conclusions reported in these studies are those of the authors and are independent from the funding sources. No endorsement by the funders, ICES or the MOHLTC is intended or should be inferred.
references
[1] E.C. Schneider, A.M. Zaslavsky, A.M. Epstein, Racial disparities in the quality of care for enrollees in Medicare managed care, JAMA 287 (2002) 1288–1294. [2] B.D. Smedley, A.Y. Stith, A.R. Nelson, Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care, National Academies Press, Washington, DC, 2003. [3] A.N. Trivedi, A.M. Zaslavsky, E.C. Schneider, et al., Trends in the quality of care and racial disparities in Medicare managed care, N. Engl. J. Med. 353 (2005) 692–700. [4] S.H. Meghani, J.M. Brooks, T. Gipson-Jones, et al., Patient–provider race-concordance: does it matter in improving minority patients’ health outcomes? Ethnic. Health 14 (2009) 107–130. [5] T.A. LaVeist, A. Nuru-Jeter, Is doctor–patient race concordance associated with greater satisfaction with care? J. Health Soc. Behav. 43 (2002) 296–306. [6] S. Saha, M. Komaromy, T.D. Koepsell, et al., Patient–physician racial concordance and the perceived quality and use of health care, Arch. Intern. Med. 159 (1999) 997–1004. [7] T.A. LaVeist, A. Nuru-Jeter, K.E. Jones, The association of doctor–patient race concordance with health service utilization, J. Public Health Policy 24 (2003) 312–323.
438
p r i m a r y c a r e d i a b e t e s 9 ( 2 0 1 5 ) 432–438
[8] W.D. King, M.D. Wong, M.F. Shapiro, et al., Does racial concordance between HIV-positive patients and their physicians affect the time to receipt of protease inhibitors? J. Gen. Intern. Med. 19 (2004) 1146–1153. [9] A.H. Traylor, J.A. Schmittdiel, C.S. Uratsu, et al., Adherence to cardiovascular disease medications: does patient–provider race/ethnicity and language concordance matter? J. Gen. Intern. Med. 25 (2010) 1172–1177. [10] A.H. Traylor, U. Subramanian, C.S. Uratsu, et al., Patient race/ethnicity and patient–physician race/ethnicity concordance in the management of cardiovascular disease risk factors for patients with diabetes, Diabetes Care 33 (2010) 520–525. [11] Y.J. Cheng, G. Imperatore, L.S. Geiss, et al., Secular changes in the age-specific prevalence of diabetes among U.S. adults: 1988–2010, Diabetes Care 36 (2013) 2690–2696. [12] A.L. Stewart, S. Greenfield, R.D. Hays, et al., Functional status and well-being of patients with chronic conditions. Results from the Medical Outcomes Study, JAMA 262 (1989) 907–913. [13] E.W. Gregg, Q. Gu, Y.J. Cheng, et al., Mortality trends in men and women with diabetes, 1971 to 2000, Ann. Intern. Med. 147 (2007) 149–155. [14] J.B. Saaddine, B. Cadwell, E.W. Gregg, et al., Improvements in diabetes processes of care and intermediate outcomes: United States, 1988–2002, Ann. Intern. Med. 144 (2006) 465–474. [15] M.K. Ali, K. McKeever Bullard, J.B. Saaddine, et al., Achievement of goals in US diabetes care, 1999–2010, N. Engl. J. Med. 368 (2013) 1613–1624. [16] B.R. Shah, J.E. Hux, A. Laupacis, et al., Deficiencies in the quality of diabetes care: comparing specialist with generalist care misses the point, J. Gen. Intern. Med. 22 (2007) 275–279. [17] A.F. Brown, E.W. Gregg, M.R. Stevens, et al., Race, ethnicity, socioeconomic position, and quality of care for adults with diabetes enrolled in managed care: the Translating Research Into Action for Diabetes (TRIAD) study, Diabetes Care 28 (2005) 2864–2870. [18] M.H. Chin, J.X. Zhang, K. Merrell, Diabetes in the African-American Medicare population: morbidity, quality of care, and resource utilization, Diabetes Care 21 (1998) 1090–1095. [19] J.M. McWilliams, E. Meara, A.M. Zaslavsky, et al., Differences in control of cardiovascular disease and diabetes by race,
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
ethnicity, and education: U.S. trends from 1999 to 2006 and effects of Medicare coverage, Ann. Intern. Med. 150 (2009) 505–515. B.R. Shah, K. Cauch-Dudek, S.S. Anand, et al., Absence of disparities in the quality of primary diabetes care for South Asians and Chinese in an urban Canadian setting, Diabetes Care 35 (2012) 794–796. J.E. Hux, F. Ivis, V. Flintoft, et al., Diabetes in Ontario: determination of prevalence and incidence using a validated administrative data algorithm, Diabetes Care 25 (2002) 512–516. B.R. Shah, M. Chiu, S. Amin, et al., Surname lists to identify South Asian and Chinese ethnicity from secondary data in Ontario, Canada: a validation study, BMC Med. Res. Methodol. 10 (2010) 42. T. Kiran, J.C. Victor, A. Kopp, et al., The relationship between financial incentives and quality of diabetes care in Ontario, Canada, Diabetes Care 35 (2012) 1038–1046. Canadian Diabetes Association Clinical Practice Guidelines Expert Committee, Canadian Diabetes Association 2008 clinical practice guidelines for the prevention and management of diabetes in Canada, Can. J. Diabetes 32 (2008) S1–S201. K. Tu, N.R.C. Campbell, X.-L. Chen, et al., Accuracy of administrative databases in identifying patients with hypertension, Open Med. 1 (2007) 18–26. Statistics Canada Language Highlight Tables, 2006 Census. 97-555-XWE2006002, Statistics Canada, Ottawa, 2007, Ref Type: Report. T.R. Konrad, D.L. Howard, L.J. Edwards, et al., Physician–patient racial concordance, continuity of care, and patterns of care for hypertension, Am. J. Public Health 95 (2005) 2186–2190. C. Field, R. Caetano, The role of ethnic matching between patient and provider on the effectiveness of brief alcohol interventions with Hispanics, Alcohol. Clin. Exp. Res. 34 (2010) 262–271. P.S. Mehler, R.A. Lundgren, I. Pines, et al., A community study of language concordance in Russian patients with diabetes, Ethnic. Dis. 14 (2004) 584–588. M. van Ryn, Research on the provider contribution to race/ethnicity disparities in medical care, Med. Care 40 (2002) I140–I151.