The European Journal of Public Health Advance Access originally published online on April 30, 2007
The European Journal of Public Health 2007 17(6):657-663; doi:10.1093/eurpub/ckm044
Miscellaneous |
Adjusted Clinical Groups use as a measure of the referrals efficiency from primary care to specialized in Spain
Antoni Sicras-Mainar1, Josep Serrat-Tarrés2, Ruth Navarro-Artieda3, Rosa Llausí-Sellés4, Ignasi Ruano-Ruano4 and Josep Antón González-Ares1
1 Planning Management, Badalona Serveis Assistencials, SA, Badalona (Barcelona), Spain
2 Public Health Management, Badalona Serveis Assistencials, SA, Badalona (Barcelona), Spain
3 Medical Documentation Service, Hospital Germans Trías i Pujol, Badalona (Barcelona), Spain
4 Planning Management, Barcelona's Sanitary Region, CatSalut, Barcelona, Spain
Correspondence: Dr Antoni Sicras-Mainar, Badalona Serveis Assistencials, SA, Planning Management, Calle Gaietà Soler, 6–8. CP: 08911, Badalona, Barcelona (Spain), tel: +34 93 507 26 84, fax: +34 93 389 32 86, e-mail: asicras{at}bsa.gs
Received October 14, 2006, accepted April 2, 2007
 |
Abstract |
|---|
 Top Abstract MethodologyResultsDiscussionAcknowledgementsReferences |
|---|
Background: To determine the referral rate (RR) per centre, its main causes and the adjusted efficiency indexes, through the retrospective implementation of the Adjusted Clinical Groups (ACG) in a Spanish primary care setting. Methods: Design descriptive–retrospective study. Attended patients by five primary care teams (PCT) during the year 2004 were included. General parameters, age, gender, dependent (visits and episodes), and comorbidity of each patient relative to each ACG are used as measures. The RR was defined as the quotient between the number of referrals and the visits made. Efficiency Index (EI) was established dividing the observed by the expected referrals obtained by indirect standardization. Statistical significance P < 0.05. Results: Studied patients 81 335 (use: 76.9%), 5.0 ± 3.6 episodes and 7.9 ± 7.8 visits/patient/year. Percentage of visits with a referral, adjusted for morbidity burden, was 7.5% (CI: 7.3–7.7); age: 48.3 ± 22.7 years (women: 55.9%), P = 0.000. The average of referrals was of 59.6 per 100 attended patients/year (P = 0.000). Visits and episodes explain 34.1–68.1%, respectively (P = 0.000), the explanatory power of the classification's variability was of 23.6% (P = 0.0001). EI per centre were: 0.95 (CI: 0.82–1.08); 0.78 (CI: 0.63–0.93); 0.88 (CI: 0.73–1.03); 1.15 (CI: 1.03–1.27) and 1.08 (CI: 0.95–1.21), P = 0.034 (family practice); and 0.83 (CI: 0.70–0.96); 0.83 (CI: 0.68–0.98); 0.84 (CI: 0.70–0.98); 1.24 (CI: 1.12–1.36) and 1.16 (CI: 1.03–1.29), P = 0.041 (paediatrics), respectively. Conclusions: Adjusted morbidity by ACG explains an important part of the referrals variability. The study results must be interpreted cautiously even after adjustment by age, gender and morbidity. Should the results be confirmed, it would allow an improvement in the measurement of referrals for clinical management in the PCT.
Keywords: adjusted clinical groups, primary care, referrals
Patient classification systems in ambulatory care, and particularly in primary care, have not been used in a generalized manner even in the US, their main place of origin.1,2 Those that sort the set of patients according to similar resource consumption can provide better comparisons of subject diversity in demographic groups.3,4 Following the current trend of dividing financing, purchase and services provision, more precise instruments of evaluation and measurement of care activity performed are required.5
The Adjusted Clinical Groups (ACG) is a system of grouping diagnoses, which classifies people according to the burden of disease that they present during a period of time. It was developed by Starfield and Weiner6,7 (Johns Hopkins University) and its aim is to measure the degree of illness in patient populations, on the basis of comorbidity levels in resource consumption. Along with Hierarchical Coexisting Conditions (HCC) or Clinical Risk Conditions (CRC), they currently constitute one of the possible methodologies for risk adjustment, which can be used to evaluate in a more precise and equitable manner, the financing of government health plans (capitation payment for providing groups) or to assess the efficiency in the utilization of health services.3–5,8
The act of referral consists of a total or partial transference of patient's care, from primary care teams (PCT) to specialized care, being the second cause of variable resource consumption in operating account of centres, after pharmaceutical prescription.5–9 Therefore, we find ourselves facing an effect of great magnitude in our healthcare system, where the current model encourages continuum care, but always with efficiency criteria; despite this, part of these referrals could be interpreted by the experts as low capacity to solve by the PCT.10,11 In the available bibliography, there are numerous evidences relative to variability in clinical practice of referral rates (RR) in Spain (between 1 and 69/100 visits).10–15 Demographic studies prove the existence of several related factors: sociodemographic characteristics of the population (morbidity pattern, rural or urban areas, expectations and user pressure, etc.), of the healthcare system itself (service offer, geographic accessibility, technological access, degree of coordination between the different healthcare levels, adequacy of referral protocols, etc.) or physician-dependent factors (training and experience, clinical performance, culture of specialized services, etc.), to highlight some examples.10–16 In general, the ACG implementation has been justified in several studies, although a better knowledge is required to reinforce their practical implementation.2,5,9,17–19 In our land, there are barely evidences, which measure or explain the referrals made to specialists depending on some risk adjustment (morbidity burden) of the patients.15
The aim of the study was to determine the RR per centre, from primary care to the specialized one, its main causes and specially to obtain the adjusted efficiency indexes (case-mix), through the retrospective implementation of the ACG, in a Spanish demographic setting.
|
Methodology |
|---|
|
TopAbstractMethodologyResultsDiscussionAcknowledgementsReferences |
|---|
We conducted a retrospective transversal study, from medical records of patients followed in ambulatory treatment and under conditions of usual clinical practice. The study population consists of individuals of both genders of the five PCT, managed by Badalona Serveis Assistencials SA. The population assigned to its zone of influence being of 105 778 inhabitants, 15.8% of them is older than 64 years. Most of the assigned population is urban, with medium-low socioeconomic level, and industrial prevalence. The kind of organization of the PCT is reformed in nature, with management of public entitlement and provision of private services (concerted with the Catalan Service of Health) and business model. Moreover, the company is endowed with resources of personnel, training policy, organizational model and services portfolio, similar to most PCT in Catalonia (Spain); with a decentralized management model and unique structural services.
All attended patients by the PCT during the year 2004 were included in the study. The referrals made to specialists were recorded through the RR, defined as the quotient between the number of referrals and the visits made, multiplied by 100, during the study year. For each referral, we obtained the annual period (first/second semester), receiving centre (different/self organization), priority (normal/preferential/urgent), issuing specialty (family practice/paediatrics) and receiving specialty (surgical/medical; and by the different specialties), as well as a series of planning indicators/parameters, described in table 1.
|
Furthermore, we obtained information on the following variables: (i) universal: age and gender, (ii) dependent: visits and episodes, and (iii) case-mix and comorbidity. The age variable was used in order to group the patients by healthcare services; this way, age ranges from 0 to 14 were included as Paediatrics patients, and older than 14 years old as Family Practice patients. Some referrals made by other professionals (dentistry, social worker, nursing, etc.) were assigned to these two services.
Visits performed were defined as the contact between the team of physicians and the patient on account of a health demand or problem at the health centre or at patient home. Administrative visits were excluded from the study. In case of an activity being performed by two professionals together, it was assigned to the professional appearing in the appointment schedule. The episode was defined as the process of care for an illness or an explicit request made by the patient (contact with healthcare services), and it was considered equivalent to diagnosis or presenting problems. They were classified according to the International Classification of Primary Care (ICPC).20 A mapping to translate the ICPC codes to the International Classification of Diseases (ICD-9-CM) was made. In order to assess it, a work team consisting of five professionals was arranged (two family physicians and three consulting technicians). Different criteria were applied depending on the relation between the codes being null (from one to none), univocal (from one to one) or multiple (from one to several).
The working algorithm of Grouper ACG® version 6.0 (http://www.acg.jhph.edu), is comprised by a series of consecutive steps until the attainment of the 106 mutually exclusive ACG for each attended patient. In order to create an ACG, we need the age, the gender and the presenting problems or diagnoses codified according to the ICD-9-CM.
The referral average (number of referrals/attended patients) relative to each ACG was obtained by means of dividing the referral average of each category by the referral average of the whole reference population (set of the five centres). Therefore, the efficiency index (EI) was established as the quotient between the observed and the expected referrals according to the ACG arrangement for the set of patients (indirect standardization). An EI value equal to 1 means equal efficiency to the norm or standard (set of centres), while an EI lower than 1 means a higher efficiency (inverse relation).
As a previous step to the analysis, and particularly to the information source belonging to the computerized medical records, the data were carefully checked, its frequency distributions were observed and possible record or codification errors were searched for. The data were obtained in a computerized way, respecting the confidentiality of the records required by law. The measurement of the classification's explanatory power has been established through the determination quotient, dividing the intergroup variation by the total variation; the transformation of variables was carried out by means of the Neperian logarithm and 95% confidence intervals (CI) were detailed. The relationship between quantitative variables was assessed with Spearman's ordinal non-parametric correlation and the ANOVA analysis, and the qualitative variables using the chi-square test. The statistical analysis was established with a significance level of 5%, by means of the application SPSSW version 12.
|
Results |
|---|
|
TopAbstractMethodologyResultsDiscussionAcknowledgementsReferences |
|---|
The inhabitant number assigned to the primary care centres during the year 2004 was 105 778, 84.7% to Family Practice and 15.3% to Paediatrics. In table 1, the general characteristics of the studied series are detailed, as well as some unitary indicators, being a certain variability observed in the characteristics of the PCT. Frequentation was 6.1 visits/inhabitant/year, being higher in the paediatric services (9.8 visits/inhabitant/year). The intensity of healthcare services utilization in the five centres was of 76.9% (CI: 76.5–77.3%), with an average of 5.0 ± 3.6 episodes and 7.9 ± 7.8 visits made per attended patient/year (n = 81 335). The general RR was 7.5% (CI: 7.3–7.7; range: 5.7–9.4; n = 48 513 referrals). The average age of referral patients was 48.3 ± 22.7 years old, being women 55.9% (P = 0.000). The average of referrals/100 attended patients/year was 59.6 (P = 0.000).
The referral distribution per each issuing service is displayed in table 2. The 90.3% of total referrals (n = 48 513) were made by family physicians and 9.7% by paediatricians. Referrals by females accounted for 57% of adult referrals, and were greater if the specialists were integrated into the primary care centre; these constituted 84.1% of referrals and are mainly of surgical type (P = 0.000).
|
The three most requested specialties (accumulated frequency: 43.6%) were orthopaedics and traumatology (15.8%), ophthalmology (15.0%) and dermatology (12.8%), but in different order depending on the services of Family Practice or Paediatrics (table 2; Spearman's ordinal correlation, P = 0.035).
The main reasons for referrals made to the most requested specialties by family physicians or paediatricians are detailed in table 3. Those made for patients with diabetes mellitus (n = 1593; 24.4%) in Family Practice stand out.
|
The total number of episodes (attended morbidity burden) was 404 473. It is worth mentioning that in 27 ACG categories, of the possible 106, no patient was classified, since the study was conducted with attended population; 43.8% of referrals was grouped in five categories, 61.7% in 10 and 74.4% in 15. The most frequently observed categories are 4910: 6–9 other ADG combinations, age >34, 0–1 major ADG (n = 8009; 16.6%), 4100: 2–3 other ADG combinations, age >34 (n = 4595; 9.5%); 4410: 4–5 other ADG combinations, age >44, no major ADG (n = 3234; 6.7%) and 4420: 4–5 other ADG combinations, age >44, 1 major ADG (n = 2891; 6.0%).
The strong correlation between visits, episodes of care and presenting problems stands out (r's = 0.93; P = 0.000 vs r's = 0.86; P = 0.0001), respectively, for all patient/years. The percentage of variance explained by ACG classification with regard to the dependent variables, with logarithmic transformation, was 34.1% for visits and 68.1% for episodes (P = 0.000), and in relation with total costs, was 23.6% (P = 0.0001).
Finally, the EI are described in figure 1 (quotient between the observed and the expected referrals for the whole of attended patients). Once having adjusted the referrals by age, gender and comorbidities (case-mix) during the study year, we observed EI 0.95 (CI: 0.82–1.08); 0.78 (CI: 0.63–0.93); 0.88 (CI: 0.73–1.03); 1.15 (CI: 1.03–1.27) and 1.08 (CI: 0.95–1.21), P = 0.034 for family physicians; and 0.83 (CI: 0.70–0.96); 0.83 (CI: 0.68–0.98); 0.84 (CI: 0.70–0.98); 1.24 (CI: 1.12–1.36) and 1.16 (CI: 1.03–1.29), P = 0.041 for paediatrics in the respective centres; therefore, PCT 1, 2 and 3 proved to be more efficient with regard to the norm.
|
|
Discussion |
|---|
|
TopAbstractMethodologyResultsDiscussionAcknowledgementsReferences |
|---|
The organizational model of most of our country PCT, based on the population assignation on a territorial base and on the growing computerization of its centres, offers a suitable frame to conduct retrospective studies under conditions of usual clinical practice.21 The studies on practical application (efficiency) of ACG carried out in Spain are scarce and they present a general lack of appropriate methodological standardization.22 Due to the characteristics of the centres and to the measure of the used variables, the results must be interpreted cautiously and taking into account the setting of health policies, resource availability and organizational culture involved. Therefore, most comparisons have been done with available Spanish data in order to minimize the variability of organization models and health policies. Our specialized care model is based on organizational requirements, so some specialists are integrated in the primary care centres; the distribution of these integrated specialists is not equal between centres. Available intern data do not reveal significant differences in proportion of referrals between integrated and hospital specialists.
An excessive RR produces undue expenses and exposes the patient to unnecessary risks, while a low RR can delay diagnosis and/or treatment, causing less efficient interventions.15 In this sense, in the US, the Agency for Health Care Policy and Research (AHCPR, 1996), summoned aids for the investigation of referral request, and as a result of this initiative, several studies emerged on a large scale. In general, they were aimed to describe the rates and the pattern of referrals, and most of them showed great interprofessional variability and greater RRs in surgical specialties. The results observed in our study show that the RR was 7.5%, with an average number of 45.9 referrals per each 100 assigned patient/years (mainly in orthopaedic surgery and traumatology, ophthalmology and dermatology); coinciding with most of published studies.10–15,16 These results, without adjusting, seem to be coherent with the usual clinical situations, since most people who frequent the healthcare services are usually chronic patients, with polypathology and polypharmacy. All these results are consistent with published literature.23–26
It is interesting to analyse the existent variability on RRs in different countries, probably attributable to differences in clinical practice, accessibility or coordination models between levels based on a vertical integration. Several studies conducted in the UK and in the US show that in equal population morbidity, British family physicians have a twofold lower RR than Americans.25,27 It seems obvious that a greater morbidity burden leads to a greater resource consumption, a higher number of visits and, finally, to a greater need of referrals to specialized care. This aspect has also been observed in our study, since the number of visits made and the episodes/attended presenting problems are strongly correlated. However, our results have a better agreement with the American ones; despite our health system policy, which is based on a national health service like the British model. Nevertheless, in the Spanish health system, a referral is always needed to see a specialist, unlike the US system where it is not necessary. In this setting, it is likely that the cultural environment, the user's expectations, the training of the professionals and other aspects have an influence, but the coordination mechanisms between the different care levels could turn out to be essential.10,12,15 While in England, primary care purchases the services of specialized care, in the US there are vertical systems that unify primary and specialized care;10,11,15,27 British people give greater power to family physicians, while Americans do to specialized care. In some European countries, where the family physician competes independently with the specialist, a referral profile similar to the American one is obtained.10,24
Standardizing the form of presenting the referral profiles between PCT can lead to more credible and comparable results, for it would be risky to conclude that a centre with a lower RR is more efficient and resolving. In the study, once having adjusted the referrals by age, gender and comorbidities (case-mix), we observed EI for family physicians and paediatrics in the respective centres, in which the PCT 1, 2 and 3 prove to be more efficient. Conceptually, ACG have great clinical validity since the morbidity patterns are relatively constant in time and people included in the group have a good chance of staying in that group throughout the following years.6,7,28,29 Hence, this type of study should be valued in a very positive sense. In a critical sense, with regard to the observed/expected efficiency in the centres, notable differences are observed (ranges 0.78–1.24).
A strictly economic evaluation leads to highlight centres with better efficiency results in RRs, but a more realistic data interpretation needs to consider other factors such as correct clinical performance, clinical governance and health outcomes.29,30 Percentage of RRs variability explained by ACG classification was 23.6% (logarithmic transformation), and so, its interpretation must be done from a global perspective of quality care, including other complementary dimensions (training, healthcare pressure, access to tests, user satisfaction, health outcomes, etc.). Our results, in a similar way to those obtained from other studies, suggest that including information from episodes improves the explained variance in a substantial proportion, probably due to their consistent relationship with resource utilization.
The most outstanding limitations should be linked to the degree of development of our information system, to lower clinical specificity, or to the possible variability and/or reliability in the selection of care episodes by different physicians that can cause contamination effects between groups.31 But the greatest limitation shows up in the external validity of the results, in two aspects; on the one hand, the studied centres are not representative of a general universe, they are centres belonging to a specific healthcare organization, and on the other hand, the number of patients included in the study (n = 81 335) is not very high for this type of generalization.
An aspect of practical interest lies in the fact that the grouper needs a limited number of variables for each patient: age, gender and diagnoses. This simplicity of use fits the characteristics of primary healthcare, with storage of a great volume of information, time limitation of the healthcare visit, coexistence of professionals in the same clinical process and patient reiteration over time.30 This feature contrasts in our environment, with the need to overcome several difficulties for an appropriate development (variability and complexity of primary care, uniformity in diagnoses codification, scarce availability of large and reliable databases, consensus among professionals) in order to have a marked uniformity and quality in obtaining the data and to favour a common language among professionals and agents.2,3,5,9,27 Moreover, the ACG system acts as a relatively neutral instrument to measure healthcare, because it offers few possibilities of perversion. In this sense, the patient's morbidity burden is less likely to be unnecessarily modified than the programming of visits or diagnostic procedures.2,6,7,9
In conclusion, ACG have proved to be an acceptable classification system of patients in usual clinical practice. The RR and pattern were similar to the ones described in previous studies. Adjusted morbidity by ACG explains an important part (68.1%) of the RR variability. The results of the study must be interpreted cautiously even after the adjustment by age, gender and morbidity. Should the results be confirmed it would allow an improvement in the measure of referrals on clinical management in PCT.
|
Acknowledgements |
|---|
|
TopAbstractMethodologyResultsDiscussionAcknowledgementsReferences |
|---|
Thanks to the different professionals of the centres, for their constant input of data on a daily basis, since without their contribution, this study could not have been conducted.
Conflict of interest: None declared.
Key points
|
|
References |
|---|
|
TopAbstractMethodologyResultsDiscussionAcknowledgementsReferences |
|---|
1 Adams EK, Bronstein JM, Raskind-Hood C. Adjusted clinical groups: predictive accuracy for Medicaid enrolees in three states. Health Care Financ Rev (2002) 24:43–61.[Medline]
2 Juncosa S, Bolíbar B, Roset M, Tome R. Performance of an ambulatory case-mix system in primary care in Spain: Ambulatory Care Groups (ACGs). European J Public Health (1999) 9:27–35.
3 Meenan RT, Goodman MJ, Fishman PA, et al. Using risk-adjustment models to identify high-cost risks. Med Care (2003) 41:1301–12.[CrossRef][Web of Science][Medline]
4 Petersen LA, Pietz K, Woodard LD, Byrne M. Comparison of the predictive validity of diagnosis-based risk adjusters for clinical outcomes. Med Care (2005) 43:61–7.[Web of Science][Medline]
5 Sicras-Mainar A, Serrat-Tarrés J. Measurement of relative cost weights as an effect of the retrospective application of adjusted clinical groups in primary care. Gac Sanit (2006) 20:132–41.[CrossRef][Medline]
6 Starfield B, Weiner J, Mumford L, Steinwachs D. Ambulatory Care Groups: a categorization of diagnoses for research and management. Health Ser Res (1991) 26:53–74.
7 Weiner JP, Starfield BH, Steinwachs DM, Mumford LM. Development and application of a population-oriented measure of ambulatory care case-mix. Med Care (1991) 29:452–72.[CrossRef][Web of Science][Medline]
8 Hughes JS, Averill RF, Eisenhanler J, et al. Clinical Risk Groups (CRGs): a classification system for risk adjusted capitation-based payment and health care management. Med Care (2004) 42:81–90.[CrossRef][Web of Science][Medline]
9 Sicras-Mainar A, Serrat-Tarres J, Navarro-Artieda R, Llopart-Lopez J. Prospects of adjusted clinical groups (ACG) in capitated payment risk adjustment. Rev Esp Salud Publica (2006) 80:55–65.[Web of Science][Medline]
10 Alberdi Ordiozola JC, Saenz-Bajo N. Factors affecting primary care referrals to specialised care in the Community of Madrid. Aten Primaria (2006) 37:253–7.[CrossRef][Medline]
11 Alonso JP, Febrel M, Domeco JH. Factores asociados a la derivación inadecuada entre atención primaria y especializada: estudio cualitativo en médicos de atención primaria. Gac Sanit (2000) 14:122–30.[Medline]
12 de Prado Prieto L, Garcia Olmos L, Rodriguez Salvanes F, Otero Puime A. Evaluation of referrals in primary care. Aten Primaria (2005) 28:146–51.
13 Ferradal Garcia JL, Peral Casado A, Villa Martinez B, Burón Llamazares JL. Variabilidad en la derivación de los profesionales de un centro de salud rural. SEMERGEN (2005) 31:351–5.
14 Rodriguez Alcala FJ, Chacón Fuertes J, Esteban Tudela M, et al. Motivos de interconsulta entre atención primaria y el segundo nivel. Aten Primaria (2005) 36:137–43.[CrossRef][Medline]
15 García Olmos F. La investigación sobre demanda derivada. SEMERGEN (2005) 31:345–6.
16 Chan BT, Austin PC. Patient, physician, and community factors affecting referrals to specialists in Ontario, Canada: a population-based, multi-level modelling approach. Med Care (2003) 41:500–11.[CrossRef][Web of Science][Medline]
17 Bolanos-Carmona V, Ocaña-Riola R, Prados-Torres A, Gutierrez-Cuadra P. Variations in health services utilization by primary care patients. Health Serv Manage Res (2002) 15:116–25.[CrossRef][Medline]
18 Hormigo A, García AJ, Martos F, et al. Evaluación retrospectiva del comportamiento de los grupos de cuidados ambulatorios en un centro de salud. Aten Primaria (1998) 21:36–42.[Medline]
19 Orueta JF, Urraca J, Berraondo I, et al. Adjusted Clinical Groups (ACGs) explain the utilization of primary care in Spain based on information registered in the medical records: a cross-sectional study. Health Policy (2006) 76:38–48.[CrossRef][Web of Science][Medline]
20 Lamberts H, Wood M. Clasificación Internacional de la Atención Primaria (CIAP). Clasificación de razones de consulta. (1990) Barcelona: Masson/SG.
21 Gené-Badia J. Possibilities of routine data bases. Aten Primaria (2006) 37:258–9.[CrossRef][Medline]
22 Sackett D, Rosenberg W, Gray J, et al. Evidence based medicine: what it is and what it isn't. BMJ (1996) 312:71–2.
23 Sullivan CO, Omar RZ, Forrest CB, Majeed A. Adjusting for case mix and social class in examining variation in home visits between practices. J Fam Pract (2004) 21:355–63.[CrossRef]
24 Grimshaw JM, Winkens RA, Shirran L, et al. Interventions to improve outpatient referrals from primary care to secondary care. Cochrane Database Syst Rev (2005) 20:CD005471.
25 Sullivan CO, Omar RZ, Ambler G, Majeed A. Case-mix and variation in specialist referrals in general practice. Br J Gen Pract (2005) 55:529–33.[Web of Science][Medline]
26 Weiner J, Starfield B, Stuart M, et al. Ambulatory care practice variation within a medicaid program. Health Serv Res (1996) 30:751–70.[Web of Science][Medline]
27 Forrest CB, Majeed A, Weiner JP, et al. Comparison of specialty referral rates in the United Kingdom and the United States: retrospective cohort analysis. BMJ (2002) 325:370–1.
28 Wahls TL, Barnett MJ, Rosenthal GE. Predicting resource utilization in a veterans health administration primary care population: comparison of methods based on diagnoses and medications. Med Care (2004) 42:123–8.[CrossRef][Web of Science][Medline]
29 Halling A, Fridh G, Ovhed I. Validating the Johns Hopkins ACG Case-Mix System of the elderly in Swedish primary health care. BMC Public Health (2006) 6:171.[CrossRef][Medline]
30 Fishman PA, Goodman MJ, Hornbrook MC, et al. Risk adjustment using automated ambulatory pharmacy data: the RxRisk model. Med Care (2003) 41:84–99.[CrossRef][Web of Science][Medline]
31 Martinez C, Juncosa S, Roset M. ¿Está relacionada la gravedad con la utilización de recursos? Una exploración del Duke Severity of Illness Scale (DUSOI). Aten Primaria (1998) 22:285–92.[Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  L. G. Sandy, T. Bodenheimer, L. G. Pawlson, and B. Starfield The Political Economy Of U.S. Primary Care Health Aff., July 1, 2009; 28(4): 1136 - 1145. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Valderas, B. Starfield, C. B. Forrest, B. Sibbald, and M. Roland Ambulatory Care Provided by Office-Based Specialists in the United States Ann. Fam. Med, March 1, 2009; 7(2): 104 - 111. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||