Gender Differences in Healthcare Utilization of Patients With Diabetes

July 12, 2012
Heike U. Kr&auml

,
mer, MSc

,
Gernot R&uuml

,
ter, MD

,
Ben Sch&ouml

,
ttker, MPH

,
Dietrich Rothenbacher, MD, MPH

,
Thomas Rosemann, MD, PhD

,
Joachim Szecsenyi, MD

,
Hermann Brenner, MD, MPH

,
Elke Raum, MD, MPH

Volume 18, Issue 7

Gender differences were found in healthcare utilization in patients with type 2 diabetes mellitus in Germany, despite a high rate of enrollment in a disease management program.

Objectives: To examine gender differences in healthcare utilization including outpatient and inpatient medical care for patients with type 2 diabetes mellitus (T2DM), despite participation in T2DM-specifi c disease management programs (DMP-DM).

Study Design: Baseline data from a cohort study in sout hwest Germany including 1146 patients with T2DM recruited between October 2008 and March 2010 were used.

Methods: After bivariate analyses, multivariate Poisson and logistic regression models were used to estimate the effect of sex on the number of general practitioner (GP) and medical specialist appointments, prescribed medications, hospitalizations, and inpatient rehabilitations, with additional consideration of glycemic control levels. Poor glycemic control (PGC) was defi ned as glycated hemoglobin >7.5%.

Results: In total, 905 participants had acceptable glycemic control and 237 participants had poor glycemic control. PGC was more prevalent in men than in women (23% vs 18%). Bivariate analyses among participants with PGC showed significantly fewer GP and medical specialist appointments, a lower number of medications, and longer rehabilitation stays in men than in women. Multivariate regression analyses among participants with PGC confi rmed statistically signifi cant gender differences for GP appointments and number of prescribed medications (P <.05) for men compared with women. Gender differences regarding inpatient care were less evident.

Conclusions: Our data disclosed major gender differences in healthcare utilization of diabetes patients in Germany despite a high DMP-DM rate. Future research should focus attention on gender-specific approaches to healthcare delivery to improve quality and access to care.

(Am J Manag Care. 2012;18(7):362-369)Cross-sectional analyses of baseline data of a large cohort study of patients with type 2 diabetes mellitus (T2DM) of whom 80% were enrolled in a T2DM-specifi c disease management program (DMP-DM) were performed to investigate gender differences in healthcare utilization.

  •  The number of general practitioner appointments and the number of prescribed medications were significantly lower in men than in women with glycated hemoglobin >7.5%.

  •  Gender differences regarding inpatient care were far less evident.

  •  Health authorities should concentrate on gender-specific healthcare differences being still inherent in DMP-DM in order to improve quality of healthcare.

Diabetes accounts for a large share of excess morbidity and mortality globally.1 Patients with type 2 diabetes mellitus (T2DM) are more susceptible to macro- or microangiopathic complications, such as myocardial infarction (MI), stroke, and peripheral arterial occlusive disease, than people without T2DM.2 During the last decade, differences between men and women with T2DM have been intensively investigated, revealing a lower quality of life and mental well-being in women than in men, as well as a shorter survival in diabetic women than in diabetic men after acute MI.3,4 Since glycemic control and diabetes-related complications are associated with healthcare utilization, the individual glycemic status has to be continuously checked in order to prevent an increase in comorbidity.5

Healthcare utilization seems to be higher among women than among men, especially at younger ages.6,7 Although there are analyses of the association between healthcare utilization and T2DM, studies analyzing disparities between men and women are still rare.8,9 Generally, it is assumed that T2DM-specifi c disease management programs (DMPDM) by sickness insurance funds improve outcome and process quality of medical care and limit gender-specifi c utilization differences due to their managing character.10 Although it was shown that DMP-DM improve process quality in Germany,10 gender differences in quality of life between DMP-DM patients remain.11

The primary aim of this study was to investigate gender differences in healthcare utilization of patients with T2DM in Germany participating in large part in DMP-DM of sickness insurance funds with additional consideration of quality of glycemic control.

METHODSStudy Design and Study Population

This analysis was based on data from the baseline examination of the DIANA study (Type 2 Diabetes Mellitus: New Approaches to Optimize Medical Care in General Practice), an epidemiological prospective cohort study of patients with T2DM conducted in the Ludwigsburg- Heilbronn area located in southwest Germany. The study was initiated in 2008 to address (short- and longterm) diabetes-related outcomes and to evaluate potential for health services improvements in patients with T2DM. Participants 18 years and older with physician-diagnosed T2DM were recruited according to a standardized protocol by 38 general practitioners (GPs) during regular practice visits between October 2008 and March 2010. The study protocol was approved by the Ethics Committees of the medical faculty of the University of Heidelberg (reference S186/2008) and of the Chamber of Physicians of Baden-Württemberg (reference B-2008-168).

Inclusion criteria for patients were prevalent T2DM, a visit to one of the participating study practices between October 2008 and March 2010, and sufficient knowledge of the German language. We excluded nursing home residents as well as patients seen by the general practitioners for palliative or emergency care only. A total of 1146 unselected patients with physician-diagnosed T2DM gave written informed consent and completed a self-administered standardized questionnaire at baseline. Medical information was obtained from the GPs by a standardized questionnaire, and a blood sample was collected by the recruiting physicians for glycated hemoglobin (A1C) measurement. A1C was assessed by a central laboratory,using ion-exchange high-pressure liquid chromatography (HPLC) (G8, Tosoh Biosciences).

Definition of Key Variables

This analysis was based on data collected from patients’ and physicians’ questionnaires at baseline and the A1C levels reported by the cooperating central laboratory.

The covariates, such as age at time of recruitment, gender, level of school education, marital status, place of residence, living condition, occupational status, smoking history, and alcohol consumption, were obtained from the participant questionnaire. To estimate general health status, the fi rst question of the short-form-12 (SF-12) questionnaire (“In general, would you say your health is … ?”; response categories “poor,” “fair,” “good,” “very good,” and “excellent”) was used.12

Information on diabetes duration, participation in a DMP-DM, body mass index (BMI) (kg/m²), and total number of prescribed medications (diabetes medications and other medications combined) was taken from the physician questionnaire.

Information on prevalent comorbidities including hypertension, hypercholesterolemia, coronary heart disease, heart failure, MI, stroke, intermittent claudication, diabetic retinopathy, diabetic neuropathy, diabetic nephropathy, depression, and cancer was taken from the physician questionnaire. In the few cases (3.8%) where no medical information was available from the GPs, information was taken from the participant questionnaire. Prevalent coronary heart disease including history of myocardial infarction or stroke, heart failure, and intermittent claudication was summarized as cardiovascular disease.

According to the recommendations of the International Diabetes Federation (IDF), glycemic control was classifi ed by an A1C level >7.5% indicating poor glycemic control (PGC).13 This classifi cation is commonly used worldwide in order to anticipate the level of diabetes progression; ie, T2DM patients with an A1C level >7.5% have a poorer disease prognosis.

Our primary outcomes were the number of outpatient appointments with general practitioners or with medical specialists within the last 3 months as well as total numbers of prescribed medications, hospitalizations, and inpatient rehabilitations including length of stay (in days) within the last 12 months. The information was available from the patient questionnaire.

Statistical Analysis

Descriptive statistics included means and proportions, respectively, on participants’ characteristics and healthcare utilization stratifi ed by gender and glycemic control. Differences between groups were tested for statistical signifi cance by 2-sided t tests and χ² tests. The Wilcoxon rank sum test was used to evaluate hospitalization and rehabilitation length of stay (2-tailed), including only patients who reported a stay. We used Poisson regression models to fit outpatient appointments (count data) or total number of prescribed medication and A1C measurements. Poisson regression returns (unadjusted and adjusted) rate ratios (RRs). Furthermore, logistic regression models were used to estimate (unadjusted and adjusted) odds ratios (ORs) for hospitalization or rehabilitation stay within the last 12 months (dichotomous) and glycemic control status (A1C level). The multivariate Poisson and logistic models were adjusted for variables selected by backward selection (P <.1) in order to identify the main independent determinants. Furthermore, effect modifi cation by glycemic status was assessed using the Breslow-Day test. For statistical testing, an alpha level of 5% was applied, and 95% confi dence intervals (CIs) calculated. SAS 9.2 (SAS Institute, Cary, North Carolina) was used throughout.

RESULTS

Overall, 624 men (54.5%) and 522 women (45.5%) participated in this study (Table 1). About 23% of the men and 18% of the women had PGC. Men were significantly younger than women (mean age ± standard deviation [SD]: men: 67.2 ± 10.1 years; women: 69.7 ± 10.4 years). In the PGC group, men were on average 5 years younger than women (men: 63.6 ± 11.1 years; women: 68.8 ± 10.8 years). Diabetes duration (measured in years since diabetes diagnosis) was similar for men (8.9 ± 7.2 years) and women (8.8 ± 7.2 years). Participation in a disease management program was slightly less common in men than in women (79.4% vs 81.6%).

There were statistically significant differences between men and women concerning physician-diagnosed comorbidities: men showed a higher prevalence of coronary heart disease, myocardial infarction, intermittent claudication, stroke, and nephropathy than women, whereas women more often had a diagnosed depression than men. The self-reported general health status did not differ between men and women —even if PGC status was considered. The mean BMI was almost identical between men and women. Alcohol consumption was far more common in men (77.7%) than in women (45.6%). Similarly, current smoking was more often reported by men than by women (14.8% vs 8.8%).

Overall, the vast majority of patients had completed up to 9 years of school education (72.2%). More men than women had completed 13 years or more of education. More men than women were still married (81.9% vs 59.7%). About 9% of all men but 18% of all women reported living alone. The majority of the patients were already retired (63%). Employment was more common in men, especially in the group with PGC (40.7% vs 16.7%).

The mean total number (± standard deviation) of GP appointments during the last 3 months was 2.5 ± 2.4 for men and 2.6 ± 2.5 for women, respectively (Table 2). Among patients with PGC, men had on average signifi cantly fewer appointments with their GPs or medical specialists than women (GP appointments: men: 2.4 ± 2.9, women: 3.3 ± 3.0; medical specialist appointments: men: 0.8 ± 1.3, women: 1.2 ± 1.5) (GP: P = .03; medical specialist: P = .05). The mean total number of prescribed medications was 5.7 ± 2.6 in men and 6.1 ± 2.5 in women, respectively. A statistically signifi cant gender difference in prescribed medication was evident in the group of patients with PGC: women received more medication than men (men: 5.8 ± 2.6, women: 6.8 ± 2.3, P = .0002).

About 17% of the participants reported a hospitalization during the last 12 months. There were small, non-signifi cant gender differences in regard to hospitalization in the respective PGC groups: more women (24.7%) than men with PGC (16.7%) were hospitalized, but men with PGC stayed on average about 1 day longer in the hospital than women with PGC. In the group of patients with an A1C <7.5%, hospitalized women stayed on average 3 days longer in the hospital than men.

Overall, about 8% of the men and 9% of the women reported a rehabilitation stay during the last 12 months. In the group of patients with PGC, women stayed on average 11 days longer than men, and in the group with an A1C <7.5%, women who had a rehabilitation stayed on average 6 days longer than men (P = .01).

Table 3 shows the results of the Poisson regression analyses of the outpatient care setting: overall, GP appointments of men and women seemed to be very similar. However, after stratification for glycemic status and full adjustment for the aforementioned covariates, men had signifi cantly fewer GP appointments than women (fully adjusted RR = 0.70, 95% CI: 0.53-0.91) in the PGC group but not in the non-PGC group.

Men, in general, had fewer specialist appointments than women (fully adjusted RR = 0.83; 95% CI: 0.69-1.00). This trend was similar for both glycemic control groups: for patients with an A1C <7.5% (fully adjusted RR = 0.84, 95% CI: 0.69-1.02) and for patients with PGC (fully adjusted RR = 0.85, 95% CI: 0.54-1.35).

We found a lower medication rate in men than women. This difference persisted after control for covariates including comorbidities in the fully adjusted model (fully adjusted RR = 0.94; 95% CI: 0.89-0.99). Also, among patients with PGC, we observed a statistically signifi cant gender difference (fully adjusted RR = 0.89; 95% CI: 0.80-0.99).

Table 4 shows the results of logistic regression analyses for the inpatient care setting. No significant variation by gender was found, neither for hospitalization nor for rehabilitation (hospitalization: fully adjusted OR for men compared with women = 0.92; 95% CI: 0.65-1.29; rehabilitation: fully adjusted OR = 0.84; 95% CI: 0.54-1.32). Also, no significant association was found after stratification for glycemic status.

There was no effect modification by glycemic status for outpatient or inpatient healthcare utilization (Breslow-Day test or Cochran-Mantel-Haenszel statistics: P >.05).

DISCUSSION

In this cohort of patients with T2DM recruited in the primary care setting in Germany, we found a higher cardiovascular comorbidity in men, whereas women suffered more often from depression, irrespective of quality of glycemic control and self-reported general health status. Analyzing the healthcare utilization in outpatient and inpatient care settings, we found no or marginal differences between men and women regarding the number of GP and medical specialist appointments, the number of prescribed medications, hospitalization, and rehabilitation. However, after stratifying for glycemic control, in the group of patients with PGC, men had significantly lower numbers of GP appointments and numbers of prescribed medications than women.

Compared with women without PGC, women with PGC had higher rates of prevalent comorbidities, like coronaryheart disease, intermittent claudication, and diabetic retinopathy or neuropathy, and therefore differences between men and women in prevalences were less pronounced in the group of patients with PGC than in the group without PGC. A higher number of comorbidities, either diabetes-related or non—diabetes-related, has been associated with an increase of healthcare use,14 but previous studies also highlighted the fact that a higher burden of disease and a lower general health status leads to a stronger increase of healthcare utilization in women than in men.6 Female perception of illnesses is different from the male and it is more culturally and socially accepted for women to be ill and seek professional help than it is for men.15 However, Christensen et al recently showed that in the elderly, gender-specifi c differences are generally less evident since the need for healthcare services might be similarly high in old men and women due to an increase of chronic diseases and frailty with age.16

A lower healthcare utilization in men compared with women with PGC might also be explained by socioeconomic factors: previous investigations found an association between a higher healthcare utilization and a lower socioeconomic status17,18 and a single marital status rate,19 which is mostly the case in women. Differences in education between men and women were more pronounced in the PGC group than in the group with good glycemic control, and the highest rate of single status was found among women with PGC. Overall, socioeconomic factors might contribute to healthcare utilization, but may infl uence healthcare utilization less than individual health status and quality of life.20

In general, our finding of a lower healthcare utilization of men than of women is in accordance with previous international examinations either with unselected populations or with diabetes patients7,8,21: our results in the outpatient care setting confi rm current statistics in Germany, for example on GP appointments showing a strong discrepancy between men and women and indicating a potential overutilization by women, especially in those with PGC.22 However, we could not confirm the finding of an analysis from Finland and Norway showing a higher inpatient care use among women than among men.23 This may be due to the very low numbers of hospitalizations and rehabilitations of our study population. A higher medication prescription rate in women compared with men, but higher inpatient costs in men than in women, were also previously described by Stock et al using data of a large German statutory health insurance focusing on 6 chronic diseases including diabetes.24 It is also in concordance with McFarlane et al and Wexler et al, indicating that women with diabetes are treated less aggressively compared with diabetic men when cardiovascular comorbidities are considered, although the prescription rate is often higher in women than in men.25,26

This study was the first analyzing total and gender-specific associations between different categories of healthcare utilization and glycemic status defi ned by A1C. A major strength is the comprehensive adjustment for numerous important covariates. Information was generated by participants and physician questionnaires and A1C tests. The data quality was overall high with a very low rate of missing values, and crucial information about diabetes, eg, prevalent comorbidities, diabetes duration, and prescribed medications, is assumed as valid since this information was obtained from the physicians. Furthermore, the almost equal distribution of men and women in our study population allowed gender-stratified analyses.

To our knowledge, no previous study has examined gender differences in healthcare utilization (ie, outpatient care visits, medication use, hospitalization, and rehabilitation) and glycemic control (measured by A1C) among diabetes patients in Germany while simultaneously adjusting for a wide range of important covariates.

Limitations of the study need to be considered in interpreting the results. First, the information on outpatient appointments and on hospital and rehabilitation length of stay was taken from the participant questionnaire, as the GPs were not asked for this information. Therefore, underreporting might be possible, as it is known that women tend to report healthcare use more carefully because being ill is culturally more accepted for women than for men.15 We also had no information on the reasons for hospitalization or rehabilitation which might explain differences in the length of stay between men and women. Furthermore, lack of statistical significance in some of the analyses, especially in inpatient care, might be explained by sample size limitations rather than a lack of association.

Since this is a cross-sectional analysis, causality of PGC and healthcare utilization cannot be determined. Furthermore, we could not differentiate by health insurance status, which has been shown to influence healthcare utilization.27 However, since more than 80% of the study participants attend a DMPDM and more than 90% of the German population is generally insured by a statutory health insurer, our population mainly represents publicly insured patients with T2DM in Germany.

Our study documents the need to further understand the gender-specific complex and possibly synergistic effect of diabetes, its risk factors, and its comorbidities on healthcare utilization. Health authorities should concentrate on gender-specific differences in order to improve both process and outcome quality of healthcare, especially the oversupply of outpatient care in women compared with men with T2DM. Examples might include introducing gender-focused diabetes treatment and training programs offered by statutory health insurances. Future research should focus on gender differences in the access to healthcare and its potential effect on diabetes progression.Acknowledgments

The authors wish to thank all the participants of the DIANA study and our cooperating study partners. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Author Affiliations: From Division of Clinical Epidemiology and Aging Research (HUK, BS, DR, HB, ER), German Cancer Research Center, Heidelberg, Germany; Practice of General Medicine (GR), Benningen/Neckar, Germany; Institute of Epidemiology and Medical Biometry (DR), Ulm University, Ulm, Germany; Department of General Practice and Health Services Research (TR), University of Zürich, University Hospital of Zürich, Zürich, Switzerland; Department of General Practice and Health Services Research (JS), University Hospital Heidelberg, Heidelberg, Germany.

Funding Source: This study was supported by a grant of the Federal Ministry of Education and Research (study identifi cation number: 01GX0746) and is registered at ClinicalTrials.gov (identifi cation number: NCT00742547). The sponsor of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

Author Disclosures: The authors (HUK, GR, BS, DR, TR, JS, HB, ER) report no relationship or fi nancial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (HUK, GR, TR, JS, HB, ER); acquisition of data (HUK, GR, DR, HB, ER); analysis and interpretation of data (HUK, HB, ER); drafting of the manuscript (HUK, DR, JS, ER); critical revision of the manuscript for important intellectual content (HUK, GR, BS, JS, HB, ER); statistical analysis (HUK, ER); provision of study materials or patients (HUK, GR, DR); obtaining funding (TR, JS, HB, ER); administrative, technical, or logistic support (DR, TR); and supervision (GR, HB, ER).

Address correspondence to: Hermann Brenner, MD, MPH, German Cancer Research Center, Division of Clinical Epidemiology and Aging Research, Im Neuenheimer Feld 581, 69120 Heidelberg, Germany. E-mail: h.brenner@dkfz.de.1. Danaei G, Lawes CM, Vander Hoorn S, Murray CJ, Ezzati M. Global and regional mortality from ischaemic heart disease and stroke attributable to higher-than-optimum blood glucose concentration: comparative risk assessment. Lancet. 2006;368(9548):1651-1659.

2. Wändell PE, Gafvels C. Patients with type 2 diabetes aged 35-64 years at four primary health care centres in Stockholm County, Sweden: prevalence and complications in relation to gender and socioeconomic status. Diabetes Res Clin Pract. 2004;63(3):195-203.

3. Unden AL, Elofsson S, Andreasson A, Hillered E, Eriksson I, Brismar K. Gender differences in self-rated health, quality of life, quality of care, and metabolic control in patients with diabetes. Gend Med. 2008; 5(2):162-180.

4. Ouhoummane N, Abdous B, Emond V, Poirier P. Impact of diabetes and gender on survival after acute myocardial infarction in the Province of Quebec, Canada--a population-based study. Diabet Med. 2009; 26(6):609-616.

5. Fu AZ, Qiu Y, Radican L, Yin DD, Mavros P. Impact of concurrent macrovascular co-morbidities on healthcare utilization in patients with type 2 diabetes in Europe: a matched study. Diabetes Obes Metab. 2010; 12(7):631-637.

6. Case A, Paxson C. Sex differences in morbidity and mortality. Demography. 2005;42(2):189-214.

7. Vegda K, Nie JX, Wang L, Tracy CS, Moineddin R, Upshur RE. Trends in health services utilization, medication use, and health conditions among older adults: a 2-year retrospective chart review in a primary care practice. BMC Health Serv Res. 2009;9:217.

8. Shalev V, Chodick G, Heymann AD, Kokia E. Gender differences in healthcare utilization and medical indicators among patients with diabetes. Public Health. 2005;119(1):45-49.

9. Nilsson PM, Theobald H, Journath G, Fritz T. Gender differences in risk factor control and treatment profi le in diabetes: a study in 229 Swedish primary health care centres. Scand J Prim Health Care. 2004; 22(1):27-31.

10. Schäfer I, Küver C, Gedrose B, et al. The diabetes management program for type 2 diabetes in Germany enhances process quality of diabetes care - a follow-up survey of patient’s experiences. BMC Health Serv Res. 2010;10:55.

11. Miksch A, Hermann K, Trieschmann J, et al. Gender-specifi c differences in the quality of life of patients with type 2 diabetes with or without participating in DMP [in German]. Gesundheitswesen. 2008;70(4): 250-255.

12. Ware J Jr, Kosinski M, Keller SD. A 12-item short-form health survey: construction of scales and preliminary tests of reliability and validity. Med Care. 1996;34(3):220-233.

13. International Diabetes Federation (IDF) Clinical Guidelines Task Force. Global guideline for type-2 diabetes. Brussels: International Diabetes Federation; 2005.

14. Struijs JN, Baan CA, Schellevis FG, Westert GP, van den Bos GA. Comorbidity in patients with diabetes mellitus: impact on medical health care utilization. BMC Health Serv Res. 2006;6:84.

15. De Visser RO, Smith JA, McDonnell EJ. ‘That’s not masculine’: masculine capital and health-related behaviour. J Health Psychol. 2009; 14(7):1047-1058.

16. Christensen K, Doblhammer G, Rau R, Vaupel JW. Ageing populations: the challenges ahead. Lancet. 2009;374(9696):1196-1208.

17. Adamson J, Ben-Shlomo Y, Chaturvedi N, Donovan J. Ethnicity, socio-economic position and gender--do they affect reported healthcare seeking behaviour? Soc Sci Med. 2003;57(5):895-904.

18. Coeli CM, Faerstein E, Chor D, Lopes CS, Werneck GL. Gender differences in the socioeconomic gradient in self-reported diabetes: does health service access play a role? Diabetes Res Clin Pract. 2009;86(2): 134-139.

19. August KJ, Sorkin DH. Marital status and gender differences in managing a chronic illness: the function of health-related social control. Soc Sci Med. 2010;71(10):1831-1838.

20. Geitona M, Zavras D, Kyriopoulos J. Determinants of healthcare utilization in Greece: implications for decision-making. Eur J Gen Pract. 2007;13(3):144-150.

21. Santosh J, Crampton P. Gender differences in general practice utilisation in New Zealand. J Prim Health Care. 2009;1(4):261-269.

22. Robert Koch Institut (RKI). Gesundheit in Deutschland aktuell - Telefonischer Gesundheitssurvey (GEDA) 2008/2009. http://www.gbe-bund .de. Updated September 21, 2010. Accessed September 27, 2011.

23. Suominen-Taipale AL, Martelin T, Koskinen S, Holmen J, Johnsen R. Gender differences in health care use among the elderly population in areas of Norway and Finland: a cross-sectional analysis based on the HUNT study and the FINRISK Senior Survey. BMC Health Serv Res. 2006;6:110.

24. Stock SA, Stollenwerk B, Redaelli M, Civello D, Lauterbach KW. Sex differences in treatment patterns of six chronic diseases: an analysis from the German statutory health insurance. J Womens Health. 2008; 17(3):343-354.

25. McFarlane SI, Castro J, Kaur J, et al. Control of blood pressure and other cardiovascular risk factors at different practice settings: outcomes of care provided to diabetic women compared to men. J Clin Hypertens. 2005;7(2):73-80.

26. Wexler DJ, Grant RW, Meigs JB, Nathan DM, Cagliero E. Sex disparities in treatment of cardiac risk factors in patients with type 2 diabetes. Diabetes Care. 2005;28(3):514-520.

27. Gruber S, Kiesel M. Inequality in health care utilization in Germany? theoretical and empirical evidence for specialist consultation. J Public Health. 2010;18:351-365.