BRIEF REPORT
Impact of the Medical Home Model on the Quality of Primary Care The Belgian Experience Catherine Moureaux, MD, BA,* Julian Perelman, PhD,w Elise Mendes da Costa, PhD,* Isabelle Roch, BA,z Lieven Annemans, PhD,y Isabelle Heymans, BA,8 and Marie-Christine Closon, PhDz
Background: The Belgium medical home (MH) model, which has been garnering support of late, resembles its US counterpart in that it aims at improving the quality of health care while containing costs. Objectives: To compare the quality of care offered by MHs with that offered by traditional individual practices (IPs) in Belgium in terms of the extent of their adherence to clinical practice guidelines in antibiotherapy, cervical-cancer screening, influenza vaccination, and the management of diabetes. Research Design: This is a retrospective study using public insurance claims data. Data consisted of a random sample of patients using the services of MHs and IPs who were previously matched according to sex, age category, location, disability, and socioeconomic status. We applied the McNemar test, the t test, or the Wilcoxon test, depending on the type of variable being compared. Subjects: The final sample comprised 43,678 patients in the year 2004. Measures: On the basis of a review of the literature, we selected 4 themes, corresponding to 25 indicators: antibiotherapy, cervicalcancer screening, influenza vaccination, and the management of diabetes. Results: MHs were more likely than IPs to adhere to evidencebased clinical practice guidelines. They prescribed less and more appropriate antibiotherapy, provided wider influenza-vaccination coverage for target groups, and provided a better follow-up for diabetics than did IPs. In regard to cervical-cancer screening, no significant differences were found. From the *School of Public Health, Universite´ Libre de Bruxelles, Brussels, Belgium; wEscola Nacional de Sau´de Pu´blica, Universidade Nova de Lisboa, Lisbon, Portugal; zFaculte´ de Sante´ Publique, Universite´ catholique de Louvain, Brussels, Belgium; yDepartment of Public Health, Faculty of Medicine, Ghent University, Ghent, Belgium; and 8Fe´de´ration des Maisons Me´dicales, Brussels, Belgium. Supported by Belgian Health Care Knowledge Centre (KCE) I.H is currently the general secretary of the French-speaking Medical Home and Health Groups Federation. C.M. has worked as general practitioner in a medical home. The remaining authors declare no conflict of interest. Reprints: Julian Perelman, PhD, Escola Nacional de Sau´de Pu´blica, Avenida Padre Cruz, Lisbon 1600-560, Portugal. E-mail: [email protected]. Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0025-7079/15/5305-0396
396 | www.lww-medicalcare.com
Conclusions: MHs, as they combine a greater adherence to guidelines and savings in secondary care, are a cost-effective alternative to traditional IPs and therefore should be encouraged. Key Words: medical home, primary care, clinical practice guidelines, quality (Med Care 2015;53: 396–400)
T
he patient-centered medical home (MH) model has prompted increasing interest in recent years in the United States as a way of revaluing primary health care. Several published assessments on the impact of this model are encouraging in that they indicate that it offers a modest improvement over the traditional, individual practice (IP). A systematic review published in 2013 concluded that it had moderate positive effects on prevention, patients’ and physicians’ satisfaction, and efficiency.1 Since then, other studies have confirmed the modest improvements in quality.2–4 The Belgian MH model resembles its US counterpart in several respects1,5: (i) team-based multidisciplinary care; (ii) enhanced access to care, concretized in Belgium by free access at point of care; (iii) a commitment to the practice of evidencebased medicine; (iv) the use of the so-called Global Medical File, which systematically gathers all patient information across care episodes; (v) a sustained physician-patient relationship; (vi) and recognized added value, achieved by means of capitation financing topped up by additional payments related to MHs treating underprivileged patients. In a recent paper, we showed that primary care expenditures were higher at MHs than at IPs but offset by substantial savings in secondary care, drugs, and examinations.6 This pair of facts raised a concern about the quality of care: Had the savings been achieved at the expense of patients’ health? In this paper, we seek to contribute to this debate by comparing MHs and IPs in terms of the extent of their adherence to evidence-based clinical practice guidelines (CPGs).
METHODS Data We used administrative data collected for administrative purposes (mainly invoicing) in 2004 by the 3 main Medical Care
Volume 53, Number 5, May 2015
Medical Care
Volume 53, Number 5, May 2015
Belgian nonprofit sickness funds (representing 88.5% of the total population). These data constituted a random sample of patients consulting at MH and IP, who were previously matched according to sex, 8 age categories, disability, neighborhood of residence, and low-copayment insurance status (a status specific to underprivileged patients). Because these data did not include information regarding the patients’ clinical characteristics, the matching could be performed only on the basis of sociodemographic variables. The final sample comprised 43,678 patients.
Choice of Themes and Indicators The process of choosing themes and indicators consisted of 3 steps. First, we performed a broad review of the literature to determine which themes and indicators that would enable us to evaluate the quality of the primary-care services (the details are provided by Annemans et al7). Second, a list of indicators was identified on the basis of the selected papers. Third, the list of themes and indicators was reviewed by a panel of 17 experts, who verified that they were relevant, that the selected indicators could be measured by means of our data, and they were worthy of deeper investigation. At this point, we had eliminated all but 4 themes: antibiotherapy, cervical-cancer screening, influenza vaccination, and management of diabetes. These themes corresponded to 25 indicators.
Definition of Populations The subpopulations appropriate for the analysis of antibiotherapy, cervical-cancer screening, and influenza vaccination were defined according to age groups to evaluate the adherence to CPGs. The subsample of diabetics was defined on the basis of health care use, the main information available from sickness fund records. We defined this subpopulation in terms of a specificity target instead of a sensibility objective. Therefore, it is safe to assume that patients identified as being diabetics suffer from this pathology. In contrast, it is possible that the least severe untreated diabetics have not been included. The priority given to specificity was to guarantee that the estimates for the diabetics’ follow-up are not biased by the inclusion of nondiabetics. The definition of populations are displayed in Table 1.
Statistical Analysis Three statistical tests were used to compare the health care outcomes of MH patients with those of IP patients: (i) to compare proportions, we used a McNemar test (in the case of samples that were too small for the McNemar test, an exact binomial test); (ii) to compare continuous variables following a Normal distribution, we used a t test for matched cohorts; (iii) to compare continuous variables following a non-Normal distribution, we used a (nonparametric) Wilcoxon test for matched cohorts.
Definition of Quality For each indicator, we chose the CPGs that were available to general practitioners in Belgium at the time of the study. That is, we gave priority to the CPGs published in French and/or Dutch and distributed by the Ministry of Health and the Belgian physicians’ associations. In those Copyright
r
2015 Wolters Kluwer Health, Inc. All rights reserved.
Medical Homes and Quality of Care
TABLE 1. Definition of Populations for the Different Topics Topic Antibiotherapy Cervix cancer screening Influenza vaccination Diabetes
Definition All people aged 25 or above All women aged 25–64 y old as target group, and women older than 64 as “non-target” group Complete population, divided by group according target groups and criteria All people aged 18 or above (except pregnant women) regularly treated with insulin (more than 37.5 DDD over 1 year); and/or having been prescribed oral antidiabetics for at least 12 mo (at least 3 packages); and/or having benefitted from at least 3 HbAIc measures over 1 year
cases in which CPGs did not exist, we chose the international CPGs prevailing in 2004, and which were used as reference points by Belgian general practitioners.
RESULTS The MH and IP samples matched perfectly for age groups, sex, disability, and low-copayment insurance status (Table 2). Given that such a matching was not performed on the subsample of diabetics, we additionally confirmed that in this regard no significant differences existed for this subgroup.
Antibiotherapy The evidence-based standard used in Belgian CPGs8 recommended amoxicillin, which is appropriate for a high proportion of pathologies among the general population. In contrast, there were indications that the combination of amoxicillin-clavulanic acid, the macrolides, and the quinolones corresponded to a small proportion of cases. Finally, the CPGs recommended that any prescription of antibiotherapy be conservative and that its long-term use be avoided. The percentage of patients consuming antibiotics was significantly lower at MHs than at IPs [65.9% vs. 72.2%; odds ratio (OR) = 0.7; P < 0.01], such as the mean volume of prescription among the treated [29.6 vs. 36.5 defined daily dose (DDD)] (Table 3). Patients at MH consumed significantly more amoxicillin, resistant penicillin, tetracycline, and nitrofurantoin. In contrast, MH patients were prescribed less in the way of combinations of amoxicillin and clavulanic acid (27.8 vs. 32.3%; OR = 0.8; P < 0.01), half as much in the way of cephalosporins (9.6% vs. 19.0%; OR = 0.5; P < 0.01), 30% less in the way of macrolides (17.5% vs. 24.9%; OR = 0.6; P < 0.01), and 30% less in the way of quinolones (16.2% vs. 23.9%; OR = 0.6; P < 0.01). According to these findings, the treatment of MH patients adhered more closely to evidence-based medicine than did the treatment of IP patients.
Cervical-cancer Screening The Belgian CPGs recommended that women 25–64 years of age who engaged in sexual intercourse undergo a cervical-cancer screening every 3 years.9 The likelihood of having been screened for cervical cancer during the previous 3 years did not differ significantly between IP and MH patients in either the 25–64 age group or above the 64 age group (Table 4). www.lww-medicalcare.com |
397
Medical Care
Moureaux et al
Volume 53, Number 5, May 2015
TABLE 2. Descriptive Statistics, by Practice Type Medical Home (n = 21,840)
Individual Practice (n = 21,838)
P (v2 test)
All patients (%) Age 1–4* Age 5–24* Age 25–39* Age 40–64* Age 65–74* Age 75–84* Age 85+* Female* Disable* Low-copayment status* Diabetic patients (%)
6.3 34.9 23.0 27.7 5.2 2.6 0.3 52.4 5.7 18.5 n = 700
6.3 34.9 23.0 27.7 5.2 2.6 0.3 52.4 5.7 18.5 n = 607
1.00
Age 1–4 Age 5–24 Age 25–39 Age 40–64 Age 65–74 Age 75–84 Age 85+ Female Disable Low-copayment status
0.0 1.0 5.6 59.6 21.2 11.5 1.0 54.7 15.8 41.2
0.0 1.0 7.6 55.6 23.1 11.6 1.1 55.3 16.4 43.0
0.63
1.00 1.00 1.00
0.83 0.76 0.51
*Matched variable.
Influenza Vaccination The CPGs recommended that all persons 65 years of age or above, diabetic patients, and patients suffering from chronic pulmonary disease be vaccinated against influenza.10 Among those 65 years of age or above, the vaccination coverage was significantly higher at MHs than at IPs (56.2% vs. 44.8%; OR = 1.6; P < 0.01); as it was among those 75 years of age or above (59.1% vs. 50.5%; OR = 1.4; P < 0.01), diabetics (64.0% vs. 41.2%; OR = 2.5; P < 0.01), young patients suffering from asthma (39.1% vs. 3.6%; OR = 17.4; P < 0.01), and patients suffering from asthma and/or chronic obstructive pulmonary disease (COPD) (64.3% vs. 44.7%; OR = 2.2; P < 0.01). Thus, overall coverage, in accordance with evidenced-based medical practice, was significantly greater at MHs than at IPs.
Management of Diabetes The CPGs recommended performing the HbA1c examination once every 3 months11 but apparently this is not the norm in either the MH or the IP setting. Indeed, the
number of diabetics who received at least 3 HbA1c per year was small and did not differ significantly between those at MHs and those at IPs (15.4% at MHs vs. 17.1% at IPs; OR = 0.9; P = 0.40) (Table 5). In contrast, the proportion of diabetics who did not undergo a single HbA1c examination over the 1-year period was significantly higher at IPs (44.3% vs. 30.4%; OR = 1.8; P < 0.01) than at MHs. No significant differences were observed for ophthalmology consultations and lipidogram. In contrast, MH patients were more likely than IP patients to get their blood creatinine measured at least once a year (68.7% vs. 62.9%; OR = 1.3; P = 0.03), in accordance with the CPGs’ recommendation.11
DISCUSSION In a previous paper, we showed that MHs produced substantial savings in secondary care.6 These savings were interpreted as signaling an encouraging shift toward primary care. What remained to be determined was precisely what
TABLE 3. Comparison of Indicators for Antibiotherapy Between Medical Homes and Individual Practices n (%) Medical Home (N = 13,131) Any antibiotherapy Any penicillin Resistant penicillin Amoxicillin Amoxicillin/clavulanic acid Cephalosporin Macrolide Quinolone Tetracycline Cotrimoxazole/trimethoprim Nitrofurantoin Mean prescription volume (DDD)
398 | www.lww-medicalcare.com
8648 6376 586 3809 3654 1258 2303 2130 1626 310 681 3891
Individual Practice (N = 13,131)
(65.9) (48.6) (4.5) (29.0) (27.8) (9.6) (17.5) (16.2) (12.4) (2.4) (5.2) (29.6)
9479 6384 485 3383 4237 2498 3275 3134 1317 397 520 4793
Copyright
r
(72.2) (48.6) (3.7) (25.8) (32.3) (19.0) (24.9) (23.9) (10.0) (3.0) (4.0) (36.5)
P < 0.01 0.92 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 0.81
2015 Wolters Kluwer Health, Inc. All rights reserved.
Medical Care
Volume 53, Number 5, May 2015
Medical Homes and Quality of Care
TABLE 4. Comparison of Indicators for Prevention Between Medical Homes and Individual Practices Medical Home [n (%)] Cervix cancer screening (MH/IP = 5999) Women aged 26–64 At least 1 Pap test in the last 3 y Women aged 65 or above (MH/IP* = 1146) At least 1 Pap test in the last 3 y Influenza vaccination (MH/IP = 1956) Population aged 65 or above At least 3 vaccinations in the 3 last years Population aged 75 or above (MH/IP = 756) At least 3 vaccinations in the 3 last years Diabetic patients (MH = 700; IP = 607) At least 2 vaccinations in the 3 last years Patients with asthma, aged 25–40 (MH = 69; IP = 56) At least 2 vaccinations in the 3 last years Patients with asthma and/or COPD, aged 25 or above (MH = 392; At least 2 vaccinations in the 3 last years
Individual Practice [n (%)]
P
3660 (61.0)
3660 (61.0)
0.98
209 (18.2)
225 (19.6)
0.39
1100 (56.2)
876 (44.8)
< 0.01
447 (59.1)
382 (50.5)
< 0.01
448 (64.0)
250 (41.2)
< 0.01
27 (39.1) IP = 360) 252 (64.3)
2 (3.6)
< 0.01
161 (44.7)
< 0.01
*MH indicates sample size for medical homes; IP, sample size for individual practices.
changes in approach accounted for this improvement and whether MHs had achieved these savings without compromising their primary objective: to improve the quality of health care by means of evidence-based medical practices. The study showed that MHs were more likely to adhere to CPGs grounded in evidence-based medicine. This was the case for antibiotherapy, which is less frequently prescribed at MHs than at IPs but when it is prescribed features the recommended drug more often than this is the case at IPs. In regard to cervical-cancer screening, no differences were observed between MHs and IPs; however, the influenza vaccination was significantly more widespread at MHs among target groups. An insufficient follow-up of diabetics through HbA1c tests was far more common at IPs than at MHs. It is noteworthy that in regard to all of the indicators, MHs performed as well as if not better than IPs. These results are open to interpretation. First, MHs are financed by capitation; however, the effects of capitation on quality of care remain unclear.12,13 Second, as MHs are committed to practicing evidence-based, the fact that they are more nearly in line with CPGs than are IPs comes as no surprise. Finally, the literature clearly shows that the teamwork approach is associated with better clinical outcomes and greater patient satisfaction than is the single-provider approach of IPs.14 The study was limited in 3 respects. First and foremost, some themes that would have contributed to our evaluation
of the quality could not be addressed, such as healthfullifestyle counseling and the treatment of mental health problems. Likewise, we did not evaluate the quality itself of the care provided but rather the adherence to evidence-based CPGs; however, the literature indicates that these are effective in improving care,15 and indicators similar to ours are frequently used in the literature as markers of quality.3 Second, the administrative data are not as fertile a source of information as clinical observational studies, particularly in regard to the patients’ clinical characteristics. However, the data were sufficiently ample to permit us to track a large and diverse sample of patients over the course of 1 year (more than 40,000 observations), whereas this would have been impossible if we had been limited to the data from a single clinical study. It is worth mentioning, however, that MHs are committed to treating more socioeconomically vulnerable population with more severe pathologies than the population that frequents IPs,16 a difference that our data may not have fully captured. A recent study indicates that the lower adherence to CPGs is more likely to prevail among populations facing adverse socioeconomic circumstances.17 In other words, because we were unable to control for the severity of social problems, the performance of MHs relative to that of IPs may be even better than this study indicates. Third, the definition of the subsample of diabetics was based on health care use, and not on clinical information, which our data did not provide. Consequently, the untreated patients were not
TABLE 5. Comparison of Indicators for Diabetes Management Between Medical Homes and Individual Practices Medical Home [n (%)] Diabetes (MH = 700; IP = 607) At least 3 measures of HbA1c per year At least 1 measures of HbA1c per year At least 1 ophthalmology consultation per year Yearly analysis of total cholesterol At least 1 analysis of total cholesterol in the last 3 years One measure of creatine at least once a year
108 487 113 421 675 481
(15.4) (69.6) (16.1) (60.1) (96.4) (68.7)
Individual Practice [n (%)] 104 338 100 355 570 382
(17.1) (55.7) (16.5) (58.4) (93.9) (62.9)
P 0.40 < 0.01 0.87 0.54 0.20 0.03
IP indicates individual practices; MH, medical home.
Copyright
r
2015 Wolters Kluwer Health, Inc. All rights reserved.
www.lww-medicalcare.com |
399
Medical Care
Moureaux et al
represented in our sample. However, a recent paper shows that only 3.6% of all diabetes cases in Belgium go untreated18; the fact that this value is rather low indicates that this bias does not compromise our findings. This study was the first to compare the MH and IP approaches to health care in Belgium in terms of the extent to which each adheres to CPGs. Our findings are encouraging in that they indicate that the Belgian MH model is characterized by a commitment to heeding evidence-based recommendations; the consequent improvement in health outcomes indicates that this commitment is paying off. Between this study’s findings and evidence published elsewhere that MHs are less costly than IPs,6 it is safe to conclude that the MH model is a cost-effective one that therefore ought to be encouraged. REFERENCES 1. Jackson GL, Powers BJ, Chatterjee R, et al. The patient-centered medical home. A systematic review. Ann Intern Med. 2013;158: 169–178. 2. Werner RM, Duggan M, Duey K, et al. The patient-centered medical home: an evaluation of a single private payer demonstration in New Jersey. Med Care. 2013;51:487–493. 3. Friedberg MW, Schneider EC, Rosenthal MB, et al. Association between participation in a multipayer medical home intervention and changes in quality, utilization, and costs of care. JAMA. 2014;311: 815–825. 4. Kern LM, Edwards A, Kaushal R. The patient-centered medical home, electronic health records, and quality of care. Ann Intern Med. 2014;160:741–749. 5. Stange KC, Nutting PA, Miller WL, et al. Defining and measuring the patient-centered medical home. J Gen Intern Med. 2010;25:601–612. 6. Perelman J, Roch I, Heymans I, et al. Medical homes versus individual practice in primary care: impact on health care expenditures. Med Care. 2013;51:682–688. 7. Annemans L, Closon J, Closon M, et al. Comparaison du cou^t et de la qualite´ de deux syste`mes de financement des soins de premie`re ligne en Belgique [Comparison of the cost and quality of two financing systems
400 | www.lww-medicalcare.com
8.
9.
10.
11.
12. 13. 14. 15. 16.
17. 18.
Volume 53, Number 5, May 2015
of first-line primary care in Belgium]. Brussels: Centre Fe´de´ral d’Expertise des Soins de Sante´; 2008. INAMI. Plateforme pour la promotion de la qualite´ Antibiotiques Feedback individuel, Prescription ambulatoire, Ge´ne´ralistes, Donne´es 2006 [Platform for quality promotion Antibiotics Individual feedback, ambulatory prescription, General practititioners, Data 2006]. Brussels: INAMI; 2007. Smeets F, De Deken L, Baeten R, et al. Aanbeveling voor goede medische praktijkvoering: Cervixkankerscreening [Recommendation for good medical practice: screening of cervix cancer]. Huisarts Nu. 2002;31:275–295. Govaerts F, Van De Vyver N, Pilaet A. Aanbeveling voor goede medische praktijkvoering. Preventie van influenza [Recommendation for good medical practice: prevention of influenza]. Huisarts Nu. 2006;34:4–18. Wens J, Sunaert P, Nobels F, et al. Recommandations de Bonne Pratique: Diabete sucre´ de type 2 [Recommendation for good medical practice: type 2 diabetes]. Societe´ Scientifique de Me´decine Ge´ne´rale. 2007;1:72. Akbari A, Mayhew A, Al-Alawi MA, et al. Interventions to improve outpatient referrals from primary care to secondary care. Cochrane Database Syst Rev. 2008;4. Gosden T, Forland F, Kristiansen I, et al. Capitation, salary, fee-forservice and mixed systems of payment: effects on the behaviour of primary care physicians. Cochrane Database Syst Rev. 2000;3. Grumbach K, Bodenheimer T. Can health care teams improve primary care practice? JAMA. 2004;291:1246–1251. Lugtenberg M, Burgers J, Westert G. Effects of evidence-based clinical practice guidelines on quality of care: a systematic review. Quality and Safety in Health Care. 2009;18:385–392. Fe´de´rations des maisons me´dicales et des collectifs de sante´ francophones. Charte de la Fe´de´ration des maisons me´dicales et des collectifs de sante´ francophones [Charter of the medical homes and French-speaking health groups Federation]. Brussels: Fe´de´rations des maisons me´dicales et des collectifs de sante´ francophones; 2006. Mercuri M, Sherbino J, Sedran RJ, et al. When guidelines don’t guide: the effect of patient context on management decisions based on clinical practice guidelines. Acad Med. 2014;1:6. Stone MA, Charpentier G, Doggen K, et al. Quality of care of people with type 2 diabetes in eight European countries: findings from the Guideline Adherence to Enhance Care (GUIDANCE) study. Diabetes Care. 2013;36:2628–2638.
Copyright
r
2015 Wolters Kluwer Health, Inc. All rights reserved.
Impact of the Medical Home Model on the Quality of Primary Care The Belgian Experience Catherine Moureaux, MD, BA,* Julian Perelman, PhD,w Elise Mendes da Costa, PhD,* Isabelle Roch, BA,z Lieven Annemans, PhD,y Isabelle Heymans, BA,8 and Marie-Christine Closon, PhDz
Background: The Belgium medical home (MH) model, which has been garnering support of late, resembles its US counterpart in that it aims at improving the quality of health care while containing costs. Objectives: To compare the quality of care offered by MHs with that offered by traditional individual practices (IPs) in Belgium in terms of the extent of their adherence to clinical practice guidelines in antibiotherapy, cervical-cancer screening, influenza vaccination, and the management of diabetes. Research Design: This is a retrospective study using public insurance claims data. Data consisted of a random sample of patients using the services of MHs and IPs who were previously matched according to sex, age category, location, disability, and socioeconomic status. We applied the McNemar test, the t test, or the Wilcoxon test, depending on the type of variable being compared. Subjects: The final sample comprised 43,678 patients in the year 2004. Measures: On the basis of a review of the literature, we selected 4 themes, corresponding to 25 indicators: antibiotherapy, cervicalcancer screening, influenza vaccination, and the management of diabetes. Results: MHs were more likely than IPs to adhere to evidencebased clinical practice guidelines. They prescribed less and more appropriate antibiotherapy, provided wider influenza-vaccination coverage for target groups, and provided a better follow-up for diabetics than did IPs. In regard to cervical-cancer screening, no significant differences were found. From the *School of Public Health, Universite´ Libre de Bruxelles, Brussels, Belgium; wEscola Nacional de Sau´de Pu´blica, Universidade Nova de Lisboa, Lisbon, Portugal; zFaculte´ de Sante´ Publique, Universite´ catholique de Louvain, Brussels, Belgium; yDepartment of Public Health, Faculty of Medicine, Ghent University, Ghent, Belgium; and 8Fe´de´ration des Maisons Me´dicales, Brussels, Belgium. Supported by Belgian Health Care Knowledge Centre (KCE) I.H is currently the general secretary of the French-speaking Medical Home and Health Groups Federation. C.M. has worked as general practitioner in a medical home. The remaining authors declare no conflict of interest. Reprints: Julian Perelman, PhD, Escola Nacional de Sau´de Pu´blica, Avenida Padre Cruz, Lisbon 1600-560, Portugal. E-mail: [email protected]. Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0025-7079/15/5305-0396
396 | www.lww-medicalcare.com
Conclusions: MHs, as they combine a greater adherence to guidelines and savings in secondary care, are a cost-effective alternative to traditional IPs and therefore should be encouraged. Key Words: medical home, primary care, clinical practice guidelines, quality (Med Care 2015;53: 396–400)
T
he patient-centered medical home (MH) model has prompted increasing interest in recent years in the United States as a way of revaluing primary health care. Several published assessments on the impact of this model are encouraging in that they indicate that it offers a modest improvement over the traditional, individual practice (IP). A systematic review published in 2013 concluded that it had moderate positive effects on prevention, patients’ and physicians’ satisfaction, and efficiency.1 Since then, other studies have confirmed the modest improvements in quality.2–4 The Belgian MH model resembles its US counterpart in several respects1,5: (i) team-based multidisciplinary care; (ii) enhanced access to care, concretized in Belgium by free access at point of care; (iii) a commitment to the practice of evidencebased medicine; (iv) the use of the so-called Global Medical File, which systematically gathers all patient information across care episodes; (v) a sustained physician-patient relationship; (vi) and recognized added value, achieved by means of capitation financing topped up by additional payments related to MHs treating underprivileged patients. In a recent paper, we showed that primary care expenditures were higher at MHs than at IPs but offset by substantial savings in secondary care, drugs, and examinations.6 This pair of facts raised a concern about the quality of care: Had the savings been achieved at the expense of patients’ health? In this paper, we seek to contribute to this debate by comparing MHs and IPs in terms of the extent of their adherence to evidence-based clinical practice guidelines (CPGs).
METHODS Data We used administrative data collected for administrative purposes (mainly invoicing) in 2004 by the 3 main Medical Care
Volume 53, Number 5, May 2015
Medical Care
Volume 53, Number 5, May 2015
Belgian nonprofit sickness funds (representing 88.5% of the total population). These data constituted a random sample of patients consulting at MH and IP, who were previously matched according to sex, 8 age categories, disability, neighborhood of residence, and low-copayment insurance status (a status specific to underprivileged patients). Because these data did not include information regarding the patients’ clinical characteristics, the matching could be performed only on the basis of sociodemographic variables. The final sample comprised 43,678 patients.
Choice of Themes and Indicators The process of choosing themes and indicators consisted of 3 steps. First, we performed a broad review of the literature to determine which themes and indicators that would enable us to evaluate the quality of the primary-care services (the details are provided by Annemans et al7). Second, a list of indicators was identified on the basis of the selected papers. Third, the list of themes and indicators was reviewed by a panel of 17 experts, who verified that they were relevant, that the selected indicators could be measured by means of our data, and they were worthy of deeper investigation. At this point, we had eliminated all but 4 themes: antibiotherapy, cervical-cancer screening, influenza vaccination, and management of diabetes. These themes corresponded to 25 indicators.
Definition of Populations The subpopulations appropriate for the analysis of antibiotherapy, cervical-cancer screening, and influenza vaccination were defined according to age groups to evaluate the adherence to CPGs. The subsample of diabetics was defined on the basis of health care use, the main information available from sickness fund records. We defined this subpopulation in terms of a specificity target instead of a sensibility objective. Therefore, it is safe to assume that patients identified as being diabetics suffer from this pathology. In contrast, it is possible that the least severe untreated diabetics have not been included. The priority given to specificity was to guarantee that the estimates for the diabetics’ follow-up are not biased by the inclusion of nondiabetics. The definition of populations are displayed in Table 1.
Statistical Analysis Three statistical tests were used to compare the health care outcomes of MH patients with those of IP patients: (i) to compare proportions, we used a McNemar test (in the case of samples that were too small for the McNemar test, an exact binomial test); (ii) to compare continuous variables following a Normal distribution, we used a t test for matched cohorts; (iii) to compare continuous variables following a non-Normal distribution, we used a (nonparametric) Wilcoxon test for matched cohorts.
Definition of Quality For each indicator, we chose the CPGs that were available to general practitioners in Belgium at the time of the study. That is, we gave priority to the CPGs published in French and/or Dutch and distributed by the Ministry of Health and the Belgian physicians’ associations. In those Copyright
r
2015 Wolters Kluwer Health, Inc. All rights reserved.
Medical Homes and Quality of Care
TABLE 1. Definition of Populations for the Different Topics Topic Antibiotherapy Cervix cancer screening Influenza vaccination Diabetes
Definition All people aged 25 or above All women aged 25–64 y old as target group, and women older than 64 as “non-target” group Complete population, divided by group according target groups and criteria All people aged 18 or above (except pregnant women) regularly treated with insulin (more than 37.5 DDD over 1 year); and/or having been prescribed oral antidiabetics for at least 12 mo (at least 3 packages); and/or having benefitted from at least 3 HbAIc measures over 1 year
cases in which CPGs did not exist, we chose the international CPGs prevailing in 2004, and which were used as reference points by Belgian general practitioners.
RESULTS The MH and IP samples matched perfectly for age groups, sex, disability, and low-copayment insurance status (Table 2). Given that such a matching was not performed on the subsample of diabetics, we additionally confirmed that in this regard no significant differences existed for this subgroup.
Antibiotherapy The evidence-based standard used in Belgian CPGs8 recommended amoxicillin, which is appropriate for a high proportion of pathologies among the general population. In contrast, there were indications that the combination of amoxicillin-clavulanic acid, the macrolides, and the quinolones corresponded to a small proportion of cases. Finally, the CPGs recommended that any prescription of antibiotherapy be conservative and that its long-term use be avoided. The percentage of patients consuming antibiotics was significantly lower at MHs than at IPs [65.9% vs. 72.2%; odds ratio (OR) = 0.7; P < 0.01], such as the mean volume of prescription among the treated [29.6 vs. 36.5 defined daily dose (DDD)] (Table 3). Patients at MH consumed significantly more amoxicillin, resistant penicillin, tetracycline, and nitrofurantoin. In contrast, MH patients were prescribed less in the way of combinations of amoxicillin and clavulanic acid (27.8 vs. 32.3%; OR = 0.8; P < 0.01), half as much in the way of cephalosporins (9.6% vs. 19.0%; OR = 0.5; P < 0.01), 30% less in the way of macrolides (17.5% vs. 24.9%; OR = 0.6; P < 0.01), and 30% less in the way of quinolones (16.2% vs. 23.9%; OR = 0.6; P < 0.01). According to these findings, the treatment of MH patients adhered more closely to evidence-based medicine than did the treatment of IP patients.
Cervical-cancer Screening The Belgian CPGs recommended that women 25–64 years of age who engaged in sexual intercourse undergo a cervical-cancer screening every 3 years.9 The likelihood of having been screened for cervical cancer during the previous 3 years did not differ significantly between IP and MH patients in either the 25–64 age group or above the 64 age group (Table 4). www.lww-medicalcare.com |
397
Medical Care
Moureaux et al
Volume 53, Number 5, May 2015
TABLE 2. Descriptive Statistics, by Practice Type Medical Home (n = 21,840)
Individual Practice (n = 21,838)
P (v2 test)
All patients (%) Age 1–4* Age 5–24* Age 25–39* Age 40–64* Age 65–74* Age 75–84* Age 85+* Female* Disable* Low-copayment status* Diabetic patients (%)
6.3 34.9 23.0 27.7 5.2 2.6 0.3 52.4 5.7 18.5 n = 700
6.3 34.9 23.0 27.7 5.2 2.6 0.3 52.4 5.7 18.5 n = 607
1.00
Age 1–4 Age 5–24 Age 25–39 Age 40–64 Age 65–74 Age 75–84 Age 85+ Female Disable Low-copayment status
0.0 1.0 5.6 59.6 21.2 11.5 1.0 54.7 15.8 41.2
0.0 1.0 7.6 55.6 23.1 11.6 1.1 55.3 16.4 43.0
0.63
1.00 1.00 1.00
0.83 0.76 0.51
*Matched variable.
Influenza Vaccination The CPGs recommended that all persons 65 years of age or above, diabetic patients, and patients suffering from chronic pulmonary disease be vaccinated against influenza.10 Among those 65 years of age or above, the vaccination coverage was significantly higher at MHs than at IPs (56.2% vs. 44.8%; OR = 1.6; P < 0.01); as it was among those 75 years of age or above (59.1% vs. 50.5%; OR = 1.4; P < 0.01), diabetics (64.0% vs. 41.2%; OR = 2.5; P < 0.01), young patients suffering from asthma (39.1% vs. 3.6%; OR = 17.4; P < 0.01), and patients suffering from asthma and/or chronic obstructive pulmonary disease (COPD) (64.3% vs. 44.7%; OR = 2.2; P < 0.01). Thus, overall coverage, in accordance with evidenced-based medical practice, was significantly greater at MHs than at IPs.
Management of Diabetes The CPGs recommended performing the HbA1c examination once every 3 months11 but apparently this is not the norm in either the MH or the IP setting. Indeed, the
number of diabetics who received at least 3 HbA1c per year was small and did not differ significantly between those at MHs and those at IPs (15.4% at MHs vs. 17.1% at IPs; OR = 0.9; P = 0.40) (Table 5). In contrast, the proportion of diabetics who did not undergo a single HbA1c examination over the 1-year period was significantly higher at IPs (44.3% vs. 30.4%; OR = 1.8; P < 0.01) than at MHs. No significant differences were observed for ophthalmology consultations and lipidogram. In contrast, MH patients were more likely than IP patients to get their blood creatinine measured at least once a year (68.7% vs. 62.9%; OR = 1.3; P = 0.03), in accordance with the CPGs’ recommendation.11
DISCUSSION In a previous paper, we showed that MHs produced substantial savings in secondary care.6 These savings were interpreted as signaling an encouraging shift toward primary care. What remained to be determined was precisely what
TABLE 3. Comparison of Indicators for Antibiotherapy Between Medical Homes and Individual Practices n (%) Medical Home (N = 13,131) Any antibiotherapy Any penicillin Resistant penicillin Amoxicillin Amoxicillin/clavulanic acid Cephalosporin Macrolide Quinolone Tetracycline Cotrimoxazole/trimethoprim Nitrofurantoin Mean prescription volume (DDD)
398 | www.lww-medicalcare.com
8648 6376 586 3809 3654 1258 2303 2130 1626 310 681 3891
Individual Practice (N = 13,131)
(65.9) (48.6) (4.5) (29.0) (27.8) (9.6) (17.5) (16.2) (12.4) (2.4) (5.2) (29.6)
9479 6384 485 3383 4237 2498 3275 3134 1317 397 520 4793
Copyright
r
(72.2) (48.6) (3.7) (25.8) (32.3) (19.0) (24.9) (23.9) (10.0) (3.0) (4.0) (36.5)
P < 0.01 0.92 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 0.81
2015 Wolters Kluwer Health, Inc. All rights reserved.
Medical Care
Volume 53, Number 5, May 2015
Medical Homes and Quality of Care
TABLE 4. Comparison of Indicators for Prevention Between Medical Homes and Individual Practices Medical Home [n (%)] Cervix cancer screening (MH/IP = 5999) Women aged 26–64 At least 1 Pap test in the last 3 y Women aged 65 or above (MH/IP* = 1146) At least 1 Pap test in the last 3 y Influenza vaccination (MH/IP = 1956) Population aged 65 or above At least 3 vaccinations in the 3 last years Population aged 75 or above (MH/IP = 756) At least 3 vaccinations in the 3 last years Diabetic patients (MH = 700; IP = 607) At least 2 vaccinations in the 3 last years Patients with asthma, aged 25–40 (MH = 69; IP = 56) At least 2 vaccinations in the 3 last years Patients with asthma and/or COPD, aged 25 or above (MH = 392; At least 2 vaccinations in the 3 last years
Individual Practice [n (%)]
P
3660 (61.0)
3660 (61.0)
0.98
209 (18.2)
225 (19.6)
0.39
1100 (56.2)
876 (44.8)
< 0.01
447 (59.1)
382 (50.5)
< 0.01
448 (64.0)
250 (41.2)
< 0.01
27 (39.1) IP = 360) 252 (64.3)
2 (3.6)
< 0.01
161 (44.7)
< 0.01
*MH indicates sample size for medical homes; IP, sample size for individual practices.
changes in approach accounted for this improvement and whether MHs had achieved these savings without compromising their primary objective: to improve the quality of health care by means of evidence-based medical practices. The study showed that MHs were more likely to adhere to CPGs grounded in evidence-based medicine. This was the case for antibiotherapy, which is less frequently prescribed at MHs than at IPs but when it is prescribed features the recommended drug more often than this is the case at IPs. In regard to cervical-cancer screening, no differences were observed between MHs and IPs; however, the influenza vaccination was significantly more widespread at MHs among target groups. An insufficient follow-up of diabetics through HbA1c tests was far more common at IPs than at MHs. It is noteworthy that in regard to all of the indicators, MHs performed as well as if not better than IPs. These results are open to interpretation. First, MHs are financed by capitation; however, the effects of capitation on quality of care remain unclear.12,13 Second, as MHs are committed to practicing evidence-based, the fact that they are more nearly in line with CPGs than are IPs comes as no surprise. Finally, the literature clearly shows that the teamwork approach is associated with better clinical outcomes and greater patient satisfaction than is the single-provider approach of IPs.14 The study was limited in 3 respects. First and foremost, some themes that would have contributed to our evaluation
of the quality could not be addressed, such as healthfullifestyle counseling and the treatment of mental health problems. Likewise, we did not evaluate the quality itself of the care provided but rather the adherence to evidence-based CPGs; however, the literature indicates that these are effective in improving care,15 and indicators similar to ours are frequently used in the literature as markers of quality.3 Second, the administrative data are not as fertile a source of information as clinical observational studies, particularly in regard to the patients’ clinical characteristics. However, the data were sufficiently ample to permit us to track a large and diverse sample of patients over the course of 1 year (more than 40,000 observations), whereas this would have been impossible if we had been limited to the data from a single clinical study. It is worth mentioning, however, that MHs are committed to treating more socioeconomically vulnerable population with more severe pathologies than the population that frequents IPs,16 a difference that our data may not have fully captured. A recent study indicates that the lower adherence to CPGs is more likely to prevail among populations facing adverse socioeconomic circumstances.17 In other words, because we were unable to control for the severity of social problems, the performance of MHs relative to that of IPs may be even better than this study indicates. Third, the definition of the subsample of diabetics was based on health care use, and not on clinical information, which our data did not provide. Consequently, the untreated patients were not
TABLE 5. Comparison of Indicators for Diabetes Management Between Medical Homes and Individual Practices Medical Home [n (%)] Diabetes (MH = 700; IP = 607) At least 3 measures of HbA1c per year At least 1 measures of HbA1c per year At least 1 ophthalmology consultation per year Yearly analysis of total cholesterol At least 1 analysis of total cholesterol in the last 3 years One measure of creatine at least once a year
108 487 113 421 675 481
(15.4) (69.6) (16.1) (60.1) (96.4) (68.7)
Individual Practice [n (%)] 104 338 100 355 570 382
(17.1) (55.7) (16.5) (58.4) (93.9) (62.9)
P 0.40 < 0.01 0.87 0.54 0.20 0.03
IP indicates individual practices; MH, medical home.
Copyright
r
2015 Wolters Kluwer Health, Inc. All rights reserved.
www.lww-medicalcare.com |
399
Medical Care
Moureaux et al
represented in our sample. However, a recent paper shows that only 3.6% of all diabetes cases in Belgium go untreated18; the fact that this value is rather low indicates that this bias does not compromise our findings. This study was the first to compare the MH and IP approaches to health care in Belgium in terms of the extent to which each adheres to CPGs. Our findings are encouraging in that they indicate that the Belgian MH model is characterized by a commitment to heeding evidence-based recommendations; the consequent improvement in health outcomes indicates that this commitment is paying off. Between this study’s findings and evidence published elsewhere that MHs are less costly than IPs,6 it is safe to conclude that the MH model is a cost-effective one that therefore ought to be encouraged. REFERENCES 1. Jackson GL, Powers BJ, Chatterjee R, et al. The patient-centered medical home. A systematic review. Ann Intern Med. 2013;158: 169–178. 2. Werner RM, Duggan M, Duey K, et al. The patient-centered medical home: an evaluation of a single private payer demonstration in New Jersey. Med Care. 2013;51:487–493. 3. Friedberg MW, Schneider EC, Rosenthal MB, et al. Association between participation in a multipayer medical home intervention and changes in quality, utilization, and costs of care. JAMA. 2014;311: 815–825. 4. Kern LM, Edwards A, Kaushal R. The patient-centered medical home, electronic health records, and quality of care. Ann Intern Med. 2014;160:741–749. 5. Stange KC, Nutting PA, Miller WL, et al. Defining and measuring the patient-centered medical home. J Gen Intern Med. 2010;25:601–612. 6. Perelman J, Roch I, Heymans I, et al. Medical homes versus individual practice in primary care: impact on health care expenditures. Med Care. 2013;51:682–688. 7. Annemans L, Closon J, Closon M, et al. Comparaison du cou^t et de la qualite´ de deux syste`mes de financement des soins de premie`re ligne en Belgique [Comparison of the cost and quality of two financing systems
400 | www.lww-medicalcare.com
8.
9.
10.
11.
12. 13. 14. 15. 16.
17. 18.
Volume 53, Number 5, May 2015
of first-line primary care in Belgium]. Brussels: Centre Fe´de´ral d’Expertise des Soins de Sante´; 2008. INAMI. Plateforme pour la promotion de la qualite´ Antibiotiques Feedback individuel, Prescription ambulatoire, Ge´ne´ralistes, Donne´es 2006 [Platform for quality promotion Antibiotics Individual feedback, ambulatory prescription, General practititioners, Data 2006]. Brussels: INAMI; 2007. Smeets F, De Deken L, Baeten R, et al. Aanbeveling voor goede medische praktijkvoering: Cervixkankerscreening [Recommendation for good medical practice: screening of cervix cancer]. Huisarts Nu. 2002;31:275–295. Govaerts F, Van De Vyver N, Pilaet A. Aanbeveling voor goede medische praktijkvoering. Preventie van influenza [Recommendation for good medical practice: prevention of influenza]. Huisarts Nu. 2006;34:4–18. Wens J, Sunaert P, Nobels F, et al. Recommandations de Bonne Pratique: Diabete sucre´ de type 2 [Recommendation for good medical practice: type 2 diabetes]. Societe´ Scientifique de Me´decine Ge´ne´rale. 2007;1:72. Akbari A, Mayhew A, Al-Alawi MA, et al. Interventions to improve outpatient referrals from primary care to secondary care. Cochrane Database Syst Rev. 2008;4. Gosden T, Forland F, Kristiansen I, et al. Capitation, salary, fee-forservice and mixed systems of payment: effects on the behaviour of primary care physicians. Cochrane Database Syst Rev. 2000;3. Grumbach K, Bodenheimer T. Can health care teams improve primary care practice? JAMA. 2004;291:1246–1251. Lugtenberg M, Burgers J, Westert G. Effects of evidence-based clinical practice guidelines on quality of care: a systematic review. Quality and Safety in Health Care. 2009;18:385–392. Fe´de´rations des maisons me´dicales et des collectifs de sante´ francophones. Charte de la Fe´de´ration des maisons me´dicales et des collectifs de sante´ francophones [Charter of the medical homes and French-speaking health groups Federation]. Brussels: Fe´de´rations des maisons me´dicales et des collectifs de sante´ francophones; 2006. Mercuri M, Sherbino J, Sedran RJ, et al. When guidelines don’t guide: the effect of patient context on management decisions based on clinical practice guidelines. Acad Med. 2014;1:6. Stone MA, Charpentier G, Doggen K, et al. Quality of care of people with type 2 diabetes in eight European countries: findings from the Guideline Adherence to Enhance Care (GUIDANCE) study. Diabetes Care. 2013;36:2628–2638.
Copyright
r
2015 Wolters Kluwer Health, Inc. All rights reserved.
