Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Determinants related to gender differences in general practice utilization: Danish Diet, Cancer and Health Cohort

Determinants related to gender differences in general practice utilization: Danish Diet, Cancer... SCANDINAVIAN JOURNAL OF PRIMARY HEALTH CARE, 2016 VOL. 34, NO. 3, 240–249 http://dx.doi.org/10.1080/02813432.2016.1207141 RESEARCH ARTICLE Determinants related to gender differences in general practice utilization: Danish Diet, Cancer and Health Cohort a b c a Jeanette Therming Jørgensen , John Sahl Andersen , Anne Tjønneland and Zorana Jovanovic Andersen a b Centre for Epidemiology and Screening, Department of Public Health, University of Copenhagen, Copenhagen, Denmark; Section of General Practice, Department of Public Health, University of Copenhagen, Copenhagen, Denmark; Danish Centre for Cancer Research, Danish Cancer Society, Copenhagen, Denmark ABSTRACT ARTICLE HISTORY Objective: This study aims to describe the determinants related to gender differences in the GP Received 2 October 2015 Accepted 4 April 2016 utilization in Danish population aged 50–65 years. Design: Cohort-based cross-sectional study. KEYWORDS Setting: Danish general practice. Cohort; Denmark; gender; Subjects: Totally, 54,849 participants of the Danish Diet, Cancer and Health cohort (50–65 years). general practice; health Main outcome measures: The sum of cohort members’ face-to-face consultations with general service use; lifestyle; practitioner (GP) at the cohort baseline year (1993–1997). We obtained data on GP visits from the unemployment Danish National Health Service Register at the cohort baseline (1993–1997), when information on lifestyle (smoking, body mass index (BMI), alcohol use, physical activity), medical conditions (som- atic and mental), employment, education, gravidity, and hormone therapy (HT) use was collected by questionnaire. Results: Women had on average 4.1 and men 2.8 consultations per year. In a crude model, women had 47% higher rate of GP visits than men (incidence rate ratio: 1.47; 95% Confidence Interval: 1.45–1.50), which remained unchanged after adjustment for lifestyle, socio-demographic and medical factors, but attenuated to 18% (1.18; 1.13–1.24) after adjustment for female factors (gravidity and post-menopausal HT. In a fully adjusted model, subjects with hypertension (1.63; 1.59–1.67), mental illness (1.63; 1.61–1.66), diabetes (1.56; 1.47–1.65), angina pectoris (1.28; 1.21–1.34), and unemployed persons (1.19; 1.18–1.21) had highest rates of GP visits. Conclusions: Gravidity and HT use explain a large proportion, but not all of the gender differ- ence in GP utilization. Medical conditions (somatic and mental) and unemployment are the main determinants of GP utilization in men and women, while lifestyle has minor effect. KEY POINTS Female gender remained a dominant determinant of GP utilization, after adjustment for life- style, socio-demography, medical and gender specific factors, with females consulting their GP 18% more often than males. Female reproductive factors (use of postmenopausal hormone therapy and gravidity) explained a large proportion of the gender variation in use of GP. Strongest determinants for GP use among Danish adults aged 50–65 years were the presence of medical conditions (somatic and mental) and unemployment, while lifestyle factors (e.g., body mass index, alcohol consumption and smoking) had minor effect. Introduction of reasons for encounters. The background for the util- ization of services in the primary health care system is Knowledge about determinants of health care utiliza- of interest to the general practitioners (GPs) and for tion is essential in the daily clinical work and planning public health in general. Existing studies on utilization of a health care system, in order to efficiently meet the of Danish GPs focused primarily on equity in access to needs of the population. General practice is character- health care or frequent attenders.[1,2] Although it is ized by free access and nonselected patients, resulting in a broad spectrum of services provided and a variety well documented that women contact GP more often CONTACT Jeanette Therming Jørgensen jethe@sund.ku.dk Centre for Epidemiology and Screening, Department of Public Health, University of Copenhagen, Copenhagen, Nørregade 10, 1165 København, Denmark Supplemental data for this article can be accessed here. 2016 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. SCANDINAVIAN JOURNAL OF PRIMARY HEALTH CARE 241 than men, little is known about the determinants of this fear of death, alcohol abstinence, low patient satisfac- difference. Vedsted et al. reported that women utilized tion and irritable bowel syndrome as determinants of GPs 47% more often than men,[3] but did not have persistent frequent attendance.[12] Gupta and Greve found that overweight and obesity’s effect on GP use data to examine whether lifestyle and gender-specific in Denmark affected GP use among frequent users factors (reproductive-related contacts, gravidity and use only,[13] suggesting variation in characteristics and of post-menopausal hormone therapy (HT)) explained underlying the mechanism between use of GP among difference. Krasnik et al. found that gender and health frequent attenders and the general population. characteristics, especially functional status and chronic The aim of this study was to examine the determi- diseases, are most important determinants of GP use in nants related to gender differences in use of GP in Denmark, whereas social factors had very little impact, Danish population aged 50–65 years. Furthermore, we but lacked data on gender-specific factors.[1] specifically examined whether gender differences in Musculoskeletal, psychological, and respiratory prob- GP utilization persisted when adjusting for lifestyle, lems have been identified by Moth et al. as the most marital, occupational and educational status, urbaniza- common reasons for encounter in Danish primary care tion, pre-existing diseases (somatic and mental), and during the period of 1993–2009, but also lacked infor- female-specific factors (gravidity and HT use). mation on gender-specific factors.[4] Green et al. investigated gender differences in med- ical care utilization in the United States (all health serv- Materials and Methods ices contacts) and found that the gender differences Study population persisted but were reduced when controlling for gen- der-specific utilization,[5] however the effect on GP con- We have linked data on 57,053 participants of the tacts alone has not yet been quantified. Furthermore, Danish Diet, Cancer, and Health (DCH) cohort to the Green et al. identified attitudinal and behavioral factors Danish National Health Service Register (NHSR) to as important predictors of medical care utilization, obtain data on GP visits. The study was conducted as whereas health knowledge did not affect health care a cohort-based cross-sectional study, where informa- utilization.[5] Evidence on the effect of lifestyle factors tion on GP contacts and confounder information were on GP utilization is conflicting. Body mass index (BMI) collected within the same year, at cohort baseline in has previously been linked to increased GP use, whereas 1993–1997. other lifestyle factors such as diet, physical activity, smoking and alcohol consumption did not influence Danish National Health Service Register attendance rates in the Dutch population.[6] This contra- dicts results from another Dutch study examining risk NHSR is a nationwide register containing information behaviors and use of GP services related to gender,[7] on all contacts within primary health care in Denmark.[14] The register was established for adminis- where no association between BMI and GP use was found. Finally, Vos et al. reported lower GP attendance trative purposes in 1984, and data has been available rate among smoking men compared to nonsmoking for research purposes since 1990.[14,15] In addition to men, and the reverse association in women.[7] citizen-related data, records in NHSR contain informa- A large body of literature examined determinants of tion on the health care provider and the type of ser- frequent attendance to GP, defined typically as the vice provided (e.g., telephone consultation, home-visit, age- and gender-stratified top 10th percentile of GP face-to-face visit, preventive consultation). Reasons for attenders,[2,8–10] and identified social factors encounter or information on specific health problems (unemployment, divorce, low education, and social is only available through NHSR to a limited extent, in support), psychological distress, and physical diseases terms of services codes (e.g., prescription renewal, add- as main determinants.[7] Results from a Dutch study itional services codes), and no diagnoses are available. showed that age, chronic illness, psychosocial prob- GP visits in this study were defined as sum of all face- lems, and analgesics prescriptions, moderately pre- to-face contacts at the year of cohort baseline dicted persistent frequent attendance, whereas gender, (1993–1997) including consultations at GPs office and medically unexplained symptoms, use of psychoactive home-visits during opening hours, while telephone drugs and prescription of antibiotics did not affect fre- consultations and prescription renewals were excluded. quent attendance.[11] Furthermore, information on cohort members visits to Additionally, Koskela et al. identified female gender, psychologist and psychiatrists within primary health obesity (BMI >30 kg/m ), former frequent attendance, sector before cohort baseline and at baseline year 242 J. T. JØRGENSEN ET AL. (since 1990) was obtained from NHSR and used as an before-mentioned diseases. Information on occupa- indicator of pre-existing mental disorders. tional status was constructed based on self-reported levels of physical activity at work. Participants classified themselves according to four different work categories Danish Diet, Cancer, and Health Cohort or as ‘‘have not been working the past year’’, repre- The DCH cohort, described in detail elsewhere,[16]is senting unemployed and individuals outside labor part of the European Prospective Investigation into force. Thus, in this study definition ‘‘unemployed’’ Cancer and Nutrition study (EPIC) and used widely for implies both unemployed and individuals not in the research into lifestyle factors, with focus on diet, labor force. and the risk of cancer and other chronic diseases. Briefly, in 1993–1997 a total of 160,725 individuals, Statistical analysis aged between 50 and 64, born in Denmark, living in Copenhagen or Aarhus, and with no previous records We used negative binomial regression model to exam- in the Danish Cancer Registry,[17] were invited to par- ine association between total number of GP visits at ticipate in the DCH cohort study, and 57,053 individu- the year of cohort baseline and abovementioned cova- als responded and participated in the study. Cohort riates, in five separate models: (1) Model 1, a crude participation involved answering comprehensive ques- model; (2) Model 2, a model adjusted for age, gender tionnaires and interviews concerning dietary intake and lifestyle factors (BMI, alcohol consumption, smok- and lifestyle factors that are known and potential risk ing, physical activity); (3) Model 3, a model adjusted factors in the development of cancer. Additionally, for age, gender, lifestyle and social factors (marital and anthropometric measurements were taken during a occupational status, education, and urbanization); (4) physical examination and various biological materials Model 4, a model adjusted for age, gender, lifestyle, were collected. Due to large cohort sample size and social and medical factors (pre-existing diseases and extensive data collection (e.g., biological material, history of cancer in the family); (5) Model 5, a fully anthropometric measurements, etc.) participants’ base- adjusted model, adjusted for age, gender, lifestyle fac- line information was collected over several years tors, social factors, medical factors and female-specific (1993–1997). The following potential determinants of factors (gravidity, previous or current use of HT). GP contact were obtained from the DCH cohort: gen- Additionally, separate Models 1 and 5 were fit for men der, age, weight, height, alcohol consumption, smoking and women separately. Interaction terms between gen- status, leisure time physical activity, marital status, der and all other covariates were introduced in fully occupational status, educational status, urbanization, adjusted model one at a time, to test potential effect pre-existing somatic diseases, history of cancer in the modification. All analyses were performed as complete- family, previous or current use of HT, number of preg- subjects-analysis. As sensitivity analyses, two additional nancies. Information on pre-existing mental disease models were fitted using alternative outcome including was obtained from the NHSR as described above. Age GP visits in ‘‘near’’ future: participants’ number of visits refers to the participants’ age at the date of the phys- 1 year after baseline (1994–1998) and in 5 years post- ical examination, and a part from age, gender, weight, baseline (1998–2002). Results are available in online height and urbanization, all other variables collected supplement (Table A). Results are presented as inci- from DCH cohort are self-reported. Urbanization was dence rate ratio (IRR) with 95% confidence intervals dichotomized into urban (Copenhagen, Frederiksberg (95%CI). Negative binomial regression procedures or Aarhus municipality) and suburban (remaining sub- (GENMOD) in SAS 3.9 (Copenhagen, Denmark) were urban municipalities around Copenhagen and Aarhus). used to conduct the analyses. The self-reported daily alcohol intake in grams was dichotomized (below the recommended limit and Results above the recommended limit), according to Danish Health and Medicines Authority’s recommendation on Of the 57,053 DCH cohort participants, 571 were weekly alcohol consumption (females: 168 g, male: excluded due to cancer diagnosis prior to cohort base- 252 g) at the time of cohort baseline. Prevalence of line. This was inclusion criteria for the DCH cohort, as pre-existing diseases was estimated based on partici- the original aim of the cohort was to study association pants reporting either being diagnosed with heart between diet and incidence of cancer. Furthermore, attack, high cholesterol, angina pectoris, stroke, hyper- 1633 were excluded due to missing values on one or tension, diabetes, gallstones and intestinal polyps more covariates of interest, leaving 54,849 cohort or self-reported use of medication to treat members for analyses in this paper. Of these 54,849 SCANDINAVIAN JOURNAL OF PRIMARY HEALTH CARE 243 cohort members, 28,643 (52.2%) were women found between GP visits and living in urban area. For (Table 1). A total of 188,709 GP contacts were regis- women, number of pregnancies (1.09; 1.04–1.14) were tered in NHSR at the cohort baseline year, giving 3.44 weakly positively associated with GP visits while previ- mean contacts per DCH participant. A total of 11,192 ous and current users of HT had 27% higher rate of GP (20.4%) cohort participants had no registered GP con- visits (1.27; 1.24–1.30) than nonusers. Further adjust- tacts in NHSR at the baseline year, of whom, majority ment for removal of lump in the breast, hysterectomy, were men (65.2%). and removal of one or both ovaries did not change The average number of visits to GP at cohort base- the estimated gender difference, and these were line in 1993–1997 was 4.06 (standard deviation 4.54) removed from the final model as they did not have for women and 2.76 (4.00) for men (Table 1). Number any effect on the number of GP visits. Significant effect of GP visits increased with age, and was higher in modification by gender was identified for a number of underweight and obese participants, smokers, physic- factors: age, alcohol consumption, smoking, physical ally inactive participants, those who drank alcohol activity, occupational status, heart attack, high choles- below recommended limit, unemployed, participants terol, angina pectoris, hypertension, diabetes intestinal with lower education, and those with pre-existing dis- polyps and mental disorders (Table 3). Lifestyle seemed ease (Table 1). GP utilization did not differ by history to have more pronounced effect on GP use in men of cancer in the family, while married and unmarried than women in this age group, as increasing age and participants had less GP contacts than divorced or wid- smoking lead to higher increase in use of GP in men owed participants (Table 1). The GP contacts in women than women. High cholesterol, hypertension, diabetes, increased with number of pregnancies and HT use intestinal polyps and mental illness led to higher (Table 1). There was no variation in GP use by urbanic- increase in GP visits in men than women. ity level. However, women had consistently, statistically significantly more GP visits than men (Table 1), even Discussion those with pre-existing diseases, e.g., diabetic women contacted GP on average 7.41, while diabetic men con- Statement of principal findings tacted GP 5.71 times per year (p< 0.001). This study yielded three major findings: (1) female gen- In a crude model, women had 47% higher rate of der is a dominant determinant of GP utilization in the GP visits than men (IRR: 1.47; 95% confidence interval age group 50–65 years, even after controlling for life- 1.45–1.50) (Table 2). This gender variation persisted style, socio-demographic, medical (somatic and men- when lifestyle, socio-demographic and medical factors tal), and female reproductive factors, with women were added to the model, but attenuated to 18% consulting GP 18% more than men; (2) female-repro- when female-specific factors were included in a fully ductive factors (gravidity and postmenopausal HT use) adjusted model (1.18; 1.13–1.24). In a fully adjusted explained a large amount of the variation in the GP model, we found no association between age and GP use; (3) pre-existing medical conditions (somatic and visits. Alcohol consumption was weakly, but signifi- mental), unemployment, and HT use in women were cantly inversely associated with GP visits and individu- major determinants of GP utilization. als drinking above the weekly recommended limit had 8% fewer annual visits (0.92; 0.91–0.94) than those adhering recommendations. Current and previous Strengths and weaknesses of the study smoking was weakly positively associated with GP vis- This study benefited from an internationally unique its (1.09; 1.07–1.12 and 1.10; 1.07–1.12, respectively), possibility to link a large Danish cohort with 57,053 while there was weak or no effect of BMI, physical participants recruited from general population to the activity, or marital status. Employment was associated national registry of primary health care utilization, with with lower use of GP (0.81; 0.79–0.82). Similarly, sub- objective assessment of GP utilization. DCH cohort has jects with more than four years of higher education high quality data on lifestyle, education, diseases, and had fewer GP visits than those with no vocational measured height and weight. DCH has been utilized in training (0.70; 0.68–0.72). Pre-existing diseases were a number of epidemiological studies on aetiology of the strongest determinants of GP visits, hypertension (1.63; 1.59–1.67), mental disorders (1.63; 1.61–1.66), dia- cancer and other chronic diseases, related to lifestyle, betes (1.56; 1.47–1.65), angina pectoris (1.28; 1.21–1.34) socio-economic and reproductive factors. However, this and stroke (1.25; 1.16–1.34), as leading determinants of is the first study linking the DCH cohort to the Danish GP use. Having a history of cancer in the family, how- NHSR, and obtaining cohort participants’ information ever had no effect on visits to GP. No association was on GP use. NHSR is considered to be of high validity 244 J. T. JØRGENSEN ET AL. Table 1. Distribution of baseline characteristics among men and women in the DCH cohort (n¼ 54,849). Number of GP visits Female Male N (%) Mean (std) N (%) Mean (std) p Value Gender 28,643 (52.2) 4.06 (4.54) 26,206 (47.8) 2.76 (4.00) <0.0001 Age – – – – – 50–54 11,917 (21.7) 3.84 (4.32) 11,243 (20.5) 2.39 (3.67) <0.0001 55–59 8899 (16.2) 4.18 (4.82) 8129 (14.8) 2.85 (4.28) <0.0001 60–65 7827 (14.3) 4.27 (4.51) 6834 (12.5) 3.27 (4.12) <0.0001 BMI – – – – – Underweight (>18.5 kg/m ) 355 (0.65) 4.43 (7.55) 66 (0.12) 3.59 (4.91) <.2485 Normal weight (18.5–24.9 kg/m ) 14,481 (26.4) 3.63 (4.14) 9110 (16.6) 2.38 (3.83) <0.0001 Overweight (25–29.9 kg/m ) 9805 (17.9) 4.24 (4.45) 13,078 (23.8) 2.73 (3.79) <0.0001 Obese (30 kg/m ) 4002 (7.30) 5.14 (5.45) 3952 (7.21) 3.73 (4.80) <0.0001 Alcohol consumption – – – – – Below the recommended weekly limit 17,456 (31.8) 4.32 (4.96) 15,114 (27.6) 2.81 (4.30) <0.0001 Above the recommended weekly limit 11,187 (20.4) 3.66 (3.74) 11,092 (20.2) 2.70 (3.55) <0.0001 Smoking status – – – – – Never 12,600 (23.0) 3.84 (4.36) 6738 (12.3) 2.39 (3.41) <0.0001 Previous 6731 (12.3) 4.20 (4.40) 9095 (16.6) 2.94 (3.82) <0.0001 Current 9312 (17.0) 4.26 (4.84) 10,373 (18.9) 2.85 (4.47) <0.0001 Physical activity – – – – – No leisure time physical activity 11,799 (21.5) 4.34 (4.94) 13,425 (24.5) 3.01 (4.37) <0.0001 Physical active in leisure time 16,844 (30.7) 3.87 (4.22) 12,781 (23.3) 2.50 (3.56) <0.0001 Marital status – – – – – Unmarried 1775 (3.24) 3.81 (4.92) 1498 (2.73) 2.93 (4.24) <0.0001 Divorced 5538 (10.1) 4.44 (4.97) 3706 (6.76) 3.18 (4.99) <0.0001 Widow/widower 2422 (4.42) 4.37 (5.25) 595 (1.08) 3.19 (4.11) <0.0001 Married 18,908 (34.5) 3.94 (4.25) 20,407 (37.2) 2.66 (3.77) <0.0001 Occupational status – – – – – Unemployed 8018 (14.6) 4.99 (5.69) 4051 (7.4) 4.15 (5.31) <0.0001 Employed 20,625 (37.6) 3.70 (3.94) 22,155 (40.4) 2.51 (3.66) <0.0001 Educational status – – – – – No vocational training 5498 (10.0) 4.95 (5.45) 2606 (4.75) 3.49(4.39) <0.0001 Higher education, <3 years 8943 (16.3) 4.03 (4.19) 3504 (6.39) 2.88(3.61) <0.0001 Higher education, 3–4 years 10,992 (20.0) 3.84 (4.48) 11,131 (20.29) 2.80(4.27) <0.0001 Higher education, >4 years 3210 (5.85) 3.38 (3.65) 8965 (16.34) 2.45(3.63) <0.0001 Municipal – – – – – Suburban 12,712 (23.2) 4.02 (4.44) 11,736 (21.40) 2.74(4.03) <0.0001 Urban 15,931 (29.1) 4.10 (4.61) 14,470 (26.38) 2.78(3.98) <0.0001 Medical conditions – – – – – Heart attack 241 (0.44) 6.87 (8.92) 880 (1.60) 5.00 (5.75) <0.0023 High cholesterol 1797 (3.28) 5.56 (5.14) 2288 (4.17) 4.18 (4.46) <0.0001 Angina pectoris 639 (1.17) 6.52 (6.70) 1010 (1.84) 5.11 (5.52) <0.0001 Stroke 283 (0.52) 6.29 (5.10) 427 (0.78) 5.09 (5.26) <0.0027 Hypertension 4957 (9.04) 5.88 (5.47) 3953 (7.21) 5.06 (5.80) <0.0001 Diabetes 428 (0.78) 7.41 (9.51) 705 (1.29) 5.71 (6.20) <0.0010 Gallstones 2049 (3.74) 5.33 (5.56) 575 (1.05) 3.66 (4.32) <0.0001 Intestinal polyps 876 (1.60) 5.11 (5.35) 1043 (1.90) 3.46 (4.06) <0.0001 Mental illness 2082 (3.80) 6.07 (6.30) 915 (1.67) 5.37 (9.25) <0.0001 History of cancer in the family 14,348 (26.2) 4.14 (4.48) 11,926 (21.7) 2.77 (3.85) <0.0001 Hormone therapy use – – – – – Never 16,057 (29.3) 3.57 (4.03) – – – Previous or current user 12,586 (23.0) 4.69 (5.04) – – – Number of pregnancies – – – – – 0 2479 (4.52) 3.68 (4.22) – – – 1 or more 26,164 (47.7) 4.13 (4.59) – – – Consultations and home-visits. t statistics for univariate associations. due to its administrative purpose, as GPs financial registered contacts with psychologist in primary care reimbursement depends on the accurate registration or referral to psychiatrist. Information on employment of GP services in the register.[15] status was obtained from question on physical activ- This study has several limitations. Information on ity at work, where cohort participants could report pre-existing somatic diseases was self-reported, and being unemployed or not in labor force during the information on the severity of disease and mental last year. Thus, it was not possible to distinguish diseases was unavailable. Therefore, we defined a between unemployed, early retired or individuals on proxy of mental disorders based on the NHSR disability pension. We also lacked information on SCANDINAVIAN JOURNAL OF PRIMARY HEALTH CARE 245 Table 2. General practice utilization determinants among men and women in the DCH cohort (n¼ 54,849). Number of GP visits b c d e f Model 1 Model 2 Model 3 Model 4 Model 5 IRR (95% CI) IRR (95% CI) IRR (95% CI) IRR (95% CI) IRR (95% CI) Gender – – – – – Male (n ¼ 26,206) 1.00 1.00 1.00 1.00 1.00 Female (n ¼ 28,643) 1.47 (1.45–1.50) 1.56 (1.53–1.59) 1.44 (1.41–1.47) 1.44 (1.41–1.46) 1.18 (1.13–1.24) Age –– ––– 50–54 1.00 1.00 1.00 1.00 1.00 55–59 1.13 (1.11–1.16) 1.12 (1.10–1.14) 1.08 (1.06–1.10) 1.06 (1.03–1.08) 1.04 (1.02–1.06) 60–65 1.22 (1.19–1.25) 1.19 (1.16–1.22) 1.06 (1.04–1.09) 1.03 (1.00–1.05) 1.02 (1.00–1.04) BMI –– ––– Underweight (>18.5 kg/m ) 1.00 1.00 1.00 1.00 1.00 Normal weight (18.5–24.9 kg/m ) 0.79 (0.72–0.88) 0.85 (0.77–0.93) 0.88 (0.79–0.97) 0.88 (0.80–0.97) 0.87 (0.79–0.96) Overweight (25–29.9 kg/m ) 0.91 (0.82–1.01) 0.97 (0.87–1.07) 0.98 (0.89–1.08) 0.95 (0.86–1.05) 0.94 (0.86–1.04) Obese (30 kg/m ) 1.17 (1.06–1.30) 1.23 (1.11–1.36) 1.20 (1.09–1.33) 1.08 (0.98–1.19) 1.09 (0.99–1.20) Alcohol consumption – – – – – Below the recommended weekly limit 1.00 1.00 1.00 1.00 1.00 Above the recommended weekly limit 0.90 (0.89–0.92) 0.91 (0.89–0.93) 0.94 (0.92–0.96) 0.92 (0.91–0.94) 0.92 (0.91–0.94) Smoking status – – – – – Never 1.00 1.00 1.00 1.00 1.00 Previous 1.13 (1.11–1.16) 1.13 (1.11–1.16) 1.12 (1.09–1.15) 1.11 (1.08–1.13) 1.10 (1.07–1.12) Current 1.13 (1.11–1.16) 1.15 (1.13–1.18) 1.09 (1.07–1.12) 1.10 (1.08–1.13) 1.09 (1.07–1.12) Physical activity – – – – – No leisure time physical activity 1.00 1.00 1.00 1.00 1.00 Physical active in leisure time 0.87 (0.85–0.88) 0.90 (0.89–0.92) 0.93 (0.92–0.95) 0.95 (0.93–0.97) 0.95 (0.93–0.97) Marital status – – – – – Unmarried 1.00 – 1.00 1.00 1.00 Divorced 1.12 (1.07–1.17) 1.10 (1.06–1.15) 1.09 (1.05–1.14) 1.06 (1.01–1.10) Widow/widower 1.04 (0.99–1.10) 1.03 (0.97–1.08) 1.03 (0.98–1.08) 1.00 (0.95–1.05) Married 0.96 (0.92–1.00) 0.97 (0.94–1.01) 1.00 (0.96–1.04) 0.97 (0.94–1.01) Occupational status – – – – – Unemployed 1.00 – 1.00 1.00 1.00 Employed 0.70 (0.68–0.72) 0.75 (0.73–0.77) 0.80 (0.78–0.82) 0.81 (0.79–0.82) Educational status – – – – – No vocational training 1.00 – 1.00 1.00 1.00 Higher education, <3 years 0.82 (0.80–0.85) 0.93 (0.90–0.96) 0.91 (0.88–0.93) 0.91 (0.88–0.93) Higher education, 3–4 years 0.80 (0.78–0.82) 0.92 (0.89–0.95) 0.89 (0.87–0.91) 0.89 (0.87–0.91) Higher education, >4 years 0.70 (0.68–0.72) 0.88 (0.85–0.91) 0.81 (0.78–0.83) 0.80 (0.78–0.83) Municipal – – – – – Suburban 1.00 – 1.00 1.00 1.00 Urban 1.02 (1.01–1.04) 0.98 (0.96–1.00) 1.00 (0.98–1.01) 1.00 (0.99–1.02) Medical conditions – – – – – Heart attack 1.77 (1.66–1.88) – – 1.19 (1.12–1.27) 1.19 (1.12–1.27) High cholesterol 1.48 (1.43–1.53) – – 1.20 (1.16–1.24) 1.20 (1.17–1.24) Angina pectoris 1.74 (1.66–1.83) – – 1.28 (1.21–1.34) 1.28 (1.21–1.34) Stroke 1.68 (1.56–1.82) – – 1.24 (1.15–1.33) 1.25 (1.16–1.34) Hypertension 1.81 (1.77–1.85) – – 1.63 (1.60–1.67) 1.63 (1.59–1.67) Diabetes 1.99 (1.87–2.11) – – 1.54 (1.46–1.63) 1.56 (1.47–1.65) Gallstones 1.31 (1.26–1.37) – – 1.18 (1.14–1.23) 1.16 (1.12–1.21) Intestinal polyps 1.25 (1.19–1.31) – – 1.17 (1.12–1.22) 1.16 (1.11–1.21) Mental illness 1.71 (1.64–1.77) – – 1.61 (1.55–1.67) 1.63 (1.61–1.66) History of cancer in the family 1.02 (1.00–1.04) – – 1.03 (1.01–1.04) 1.02 (1.01–1.04) Hormone therapy use – – – – – Never 1.00 – – – 1.00 Previous or current user 1.30 (1.27–1.33) 1.27 (1.24–1.30) Number of pregnancies – – – – – 0 1.00 – – – 1.00 1 or more 1.13 (1.08–1.18) – – – 1.09 (1.05–1.14) Consultations and home-visits. Adjusted for gender and age. Adjusted for gender, age and lifestyle factors (BMI, smoking, alcohol consumption, physical activity). Adjusted for model 2 and socio-demographic factors (marital, occupational, educational status, and urbanization). Adjusted for model 3 and medical factors (heart attack, high cholesterol, angina pectoris, stroke, hypertension, diabetes, intestinal polyps, gallstones, and history of cancer in the family). Adjusted for model 4 and female reproductive factors (use of HT, and number of pregnancies). vulnerability or life-events (death in family, divorce, nonparticipants,[16] as well as the study is based on loss of job, etc.), that may have affected GP use.[1] data from 1993–1997 and on the specific age group Another weakness is that participants in DCH 50–65, limiting generalizability of the results to gen- had higher education and income than eral population and other age groups. 246 J. T. JØRGENSEN ET AL. Table 3. General practice utilization determinants in the DCH cohort (n¼ 54,849) by gender. Number of GP visits Female (n¼ 28643) Male (n¼ 26206) b c b c Crude model Adjusted model Crude model Adjusted model IRR (95% CI) IRR (95% CI) IRR (95% CI) IRR (95% CI) p Value Age –– –– – 50–54 1.00 1.00 1.00 1.00 55–59 1.09 (1.06–1.12) 0.98(0.96–1.01) 1.19 (1.15–1.23) 1.12(1.08–1.15) <0.0001 60–65 1.11 (1.08–1.14) 0.92(0.90–0.95) 1.36 (1.32–1.42) 1.14(1.10–1.19) <0.0001 BMI –– –– – Underweight (>18.5 kg/m ) 1.00 1.00 1.00 1.00 Normal weight (18.5–24.9 kg/m ) 0.82 (0.74–0.91) 0.88(0.80–0.97) 0.65 (0.49–0.87) 0.81(0.62–1.06) 0.3728 Overweight (25–29.9 kg/m ) 0.95 (0.86–1.06) 0.96(0.87–1.06) 0.74 (0.56–0.99) 0.87(0.66–1.13) 0.3654 Obese (30 kg/m ) 1.16 (1.04–1.28) 1.07(0.97–1.18) 1.02 (0.76–1.35) 1.03(0.78–1.35) 0.7656 Alcohol consumption – – – – – Below the recommended weekly limit 1.00 1.00 1.00 1.00 Above the recommended weekly limit 0.85 (0.83–0.87) 0.89(0.87–0.91) 0.97 (0.94–1.00) 0.96(0.93–0.98) 0.0002 Smoking status – – – – – Never 1.00 1.00 1.00 1.00 Previous 1.09 (1.06–1.12) 1.07(1.04–1.10) 1.20 (1.15–1.24) 1.14(1.10–1.18) <0.0001 Current 1.11 (1.08–1.14) 1.07(1.04–1.10) 1.18 (1.14–1.23) 1.13(1.09–1.17) 0.0006 Physical activity – – – – – No leisure time physical activity 1.00 1.00 1.00 1.00 Physical active in leisure time 0.89 (0.87–0.91) 0.96(0.94–0.99) 0.84 (0.82–0.87) 0.95(0.93–0.98) 0.0534 Marital status – – – – – Unmarried 1.00 1.00 1.00 1.00 Divorced 1.16 (1.10–1.22) 1.05 (1.00–1.11) 1.07 (1.00–1.16) 1.06(0.99–1.14) 0.6682 Widow/widower 1.11 (1.05–1.18) 1.02 (0.96–1.09) 1.00 (0.89–1.12) 1.02(0.92–1.14) 0.3401 Married 1.03 (0.98–1.08) 0.99(0.94–1.04) 0.88 (0.83–0.94) 0.96(0.90–1.02) 0.3962 Occupational status – – – – – Unemployed 1.00 1.00 1.00 1.00 – Employed 0.74 (0.72–0.76) 0.83 (0.81–0.86) 0.63 (0.61–0.66) 0.75(0.72–0.78) <0.0001 Educational status – – – – – No vocational training 1.00 1.00 1.00 1.00 – Higher education, <3 years 0.82 (0.79–0.85) 0.89(0.86–0.92) 0.82 (0.77–0.87) 0.92(0.87–0.98) 0.5876 Higher education, 3–4 years 0.78 (0.76–0.81) 0.87(0.84–0.90) 0.81 (0.77–0.85) 0.91(0.86–0.95) 0.6859 Higher education, >4 years 0.69 (0.66–0.72) 0.80(0.77–0.84) 0.71 (0.67–0.75) 0.81(0.77–0.85) 0.3323 Municipal – – – – – Suburban 1.00 1.00 1.00 1.00 – Urban 1.02 (1.01–1.05) 1.01(0.99–1.03) 1.03 (1.00–1.06) 0.99(0.96–1.02) 0.1236 Medical conditions – – – – – Heart attack 1.67 (1.48–1.87) 1.14(1.01–1.27) 1.77 (1.64–1.92) 1.15(1.07–1.25) 0.0147 High cholesterol 1.39 (1.32–1.45) 1.19(1.14–1.24) 1.57 (1.49–1.65) 1.22(1.16–1.28) 0.0006 Angina pectoris 1.60 (1.49–1.72) 1.21(1.13–1.30) 1.83 (1.70–1.96) 1.30(1.21–1.40) 0.0003 Stroke 1.54 (1.38–1.71) 1.24(1.12–1.37) 1.76 (1.58–1.97) 1.19(1.07–1.32) 0.1519 Hypertension 1.59 (1.54–1.63) 1.46(1.42–1.50) 2.12 (2.04–2.20) 1.86(1.80–1.94) <0.0001 Diabetes 1.83 (1.68–2.00) 1.48(1.36–1.60) 2.09 (1.92–2.27) 1.57(1.45–1.70) 0.0096 Gallstones 1.33 (1.28–1.39) 1.18(1.13–1.23) 1.30 (1.18–1.43) 1.13(1.03–1.24) 0.7863 Intestinal polyps 1.26 (1.18–1.34) 1.12(1.05–1.19) 1.24 (1.15–1.34) 1.20(1.12–1.28) 0.0312 Mental illness 1.56 (1.50–1.63) 1.45(1.39–1.50) 2.03 (1.89–2.19) 1.87(1.75–2.01) <0.0001 History of cancer in the family 1.04 (1.01–1.06) 1.03(1.01–1.06) 1.00 (0.97–1.03) 1.01(0.98–1.04) 0.2401 Hormone therapy use – – – – – Never 1.00 1.00 – – – Previous or current user 1.31 (1.28–1.34) 1.29(1.26–1.32) – – – Number of pregnancies – – – – – 0 1.00 1.00 – – – 1 1.10 (1.05–1.16) 1.07(1.01–1.12) – – – 2 1.08 (1.03–1.12) 1.07(1.02–1.12) – – – 3 1.11 (1.06–1.16) 1.08(1.03–1.13) – – – 4 1.15 (1.10–1.21) 1.08(1.03–1.14) – – – 5 1.24 (1.17–1.32) 1.11(1.05–1.18) – – – 6 1.29 (1.19–1.39) 1.13(1.05–1.22) – – – 7 1.36 (1.20–1.53) 1.19(1.06–1.33) – – – 8 1.38 (1.19–1.61) 1.21(1.05–1.40) – – – Consultations and home-visits. Adjusted for age. Fully adjusted model. p Value for interaction between gender and given covariate. SCANDINAVIAN JOURNAL OF PRIMARY HEALTH CARE 247 Findings in relation to other studies lifestyle factors has previously been evaluated in rela- tion to frequent attendance in general practice in We found that 20.4% of the cohort members had no Denmark, we present novel estimates of an effect of face-to-face contacts with their GP at baseline year, alcohol use and smoking on GP use in the general consistent with previous results by Gupta and population, which needs to be reproduced. We found Greve.[13] Gender differences persisted but attenuated weak association between GP visits and physical activ- when female-specific factors were included in the ity and none with BMI, in agreement with Gupta and model, but 18% higher GP utilization by women Greve.[13] remained unexplained, in line with previous find- ings.[1,3,5] A U.S. study by Green and Pope reported Socio-demographic factors that female gender remained an independent long- term predictor of higher use of medical services when Employment status was a strong determinant of GP controlling for factors likely to cause gender differen- visits, with employed participants having 19% fewer ces.[5] Likewise, Krasnik et al., who restricted the study GP visits than unemployed. However, as mentioned to participants 50 years and older in order to exclude earlier, some of those classified as unemployed in our possible contacts related to children in parous women, study, may have been retired or on disability pension. still found higher rates of GP use in women.[1] It has Our finding is inconsistent with Krasnik et al., where been suggested that gender differences may be occupational status had no effect on GP use, after explained by differences in health perception and that adjusting for gender, health status (functional limita- women are more sensitive and thus more likely to tions, mental health, chronic diseases), vulnerability report symptoms than men.[3] Furthermore, it has and life-events.[1] Inconsistencies may be due to differ- been suggested that men and women have different ences in the age of participants in the two studies, disease patterns. While women have higher prevalence lack of adjustment for vulnerability and life-events in of non-threatening chronic diseases that are manage- our cohort, or lack of separation of information in our able in GP, men have higher prevalence of more cohort on retired and unemployed. We found a stron- severe, life-threatening chronic diseases requiring hos- ger effect of employment status in men than women, pital admission or treatment within the secondary in line with earlier observation that men are more vul- health care sector.[3,18] The Danish National Institute nerable after unemployment than women.[20] GP visits of Public Health showed in a population-based survey, were inversely associated with education, with least that the prevalence of mental illness is higher among number of visits among those with the highest com- women, in all age groups (from 16 years of age),[19] pleted education, in agreement with earlier findings of however, we find that gender differences in GP utiliza- a systematic decrease in GP contacts with higher edu- tion persist, even when analysis are adjusted for men- cation for both men and women.[21] tal disorders. Medical factors Lifestyle factors Pre-existing diseases were the strongest determinants Surprisingly, GP visits were weakly inversely associated of GP visits, including hypertension, mental disorders, with alcohol consumption, which contradicts results by diabetes, angina pectoris and stroke. This is consistent van Steenkiste et al.[6] Inconsistencies may be with Krasnik et al. who identified functional limitations explained by differences in the Danish and Dutch and chronic diseases as strong determinants of GP util- study populations, general drinking habits, and defin- ization, while general subjective health measures (own ition of alcohol consumption. The inverse association perception of general health and health compared to between drinking alcohol and visits to GP was margin- others) were insignificant.[1] The association between ally stronger in women than men in the present study. GP visits and all: heart attack, high cholesterol, angina Smoking was positively associated with GP visits, which pectoris, hypertension, diabetes, intestinal polyps, and is similar to previous findings by Vos et al. mental disorders were significantly modified by gen- Furthermore, Vos et al. reported lower attendance rates der, with higher number of visits in men than women. among smoking men and the reverse for women, we found both female and male smoker to have higher Female reproductive factors rates of GP visits than nonsmokers. Furthermore, we found significantly stronger association of smoking As estimated effect of 47% higher GP attendance rate with GP visits in men than women. As the effect of in women than men, attenuated the most after 248 J. T. JØRGENSEN ET AL. adjustment for gender-specific variables, including practice that women and men have different health number of pregnancies and previous or current use of seeking behaviors and beliefs. This is important in the HT, related to menopausal symptoms. We found that evaluation of symptoms and the communication with women with one or more previous pregnancies use the patients. It is of special significance in the age their GP 9% more than those with no history of preg- group 50–65 years, where incidence of major chronic nancies. Since GP contacts related to children were diseases and cancer is steeply increasing. It will be removed from the data and the study population is relevant to evaluate whether difference in GP utiliza- aged above 50 years, the influence of parity may seem tion can explain some of this gender difference in strange. Nevertheless, the higher attendance rates overall and cause specific mortality. Results of this among parous women in this age group may be study may furthermore be relevant for policy makers explained by long-term consequences of pregnancy or in primary health care to target specific groups of men delivery, such as problems related to incontinence, in order to raise awareness of importance of contact- bladder infections or genital prolapse.[22,23] This study ing GP with serious symptoms early. For example, presented novel result that a large amount, but not all recent campaign by Danish Cancer Society in Denmark of gender variations in GP utilization is explained by a has targeted men specifically to be more aware of female reproductive factor. early symptoms of colorectal cancer,[27] and to contact their GP early with these symptoms, due to men being at higher risk from and having considerably poorer sur- Meaning of the study: possible mechanisms and vival from colorectal cancer than women. implications In summary, some of the gender differences in health Disclosure statement seeking behavior remained unexplained, even after adjusting for gender-specific utilization and lifestyle The authors report no conflicts of interest. factors. However, psychological factors and preexisting disease may explain some of the remaining gender Ethical approval variation, which have been identified as important Relevant Danish ethical committees and Danish Data determinants of frequent attendance to GP.[24] Protection Agency have approved the study (J.nr. 2013-41- Furthermore, 20–30% of the symptoms presented in 1600), and written informed consent was provided by all par- general practice are classified as medically unexplained ticipants at recruitment. symptoms,[25] and they are more common in women, which may explain some of the gender variations in References GP visits. Although the total number of GP contacts has increased by 17% from 1992 to 2001 in Denmark, [1] Krasnik A, Hansen E, Keiding N, et al. Determinants of number of face-to-face GP contacts exclusively, which general practice utilization in Denmark. Dan Med Bull are used as outcome in this study, increased only DENMARK. 1997;44:542–546. [2] Vedsted P, Christensen MB. Frequent attenders in gen- slightly, by 5%.[26] Similarly, Moth et al. [4] reported eral practice care: a literature review with special refer- only small changes in the reasons for encounter with ence to methodological considerations. Public Health. GP between 1993 and 2009. This implies that current 2005;119:118–137. mechanisms that cause individuals to contact their GP [3] Vedsted P. Kønsforskelle i brug af sundhedsvæsnet most likely do not differ substantially from estimates [Gender differences in the use of the Danish health reported in this study, based on data in 1993–97. care system], English summary. Ugeskr Læger. 2007;169:2403–2408. Furthermore, this study is based on population of [4] Moth G, Olesen F, Vedsted P. Reasons for encounter 50–65-year-old individuals, but results seem rather con- and disease patterns in Danish primary care: changes sistent with related study in younger age groups over 16 years. Scand J Prim Health Care. (20–65) in Denmark.[13] Still, more studies with more 2012;30:70–75. recent data on GP use, and in different ages, both [5] Green CA, Pope CR. Gender, psychosocial factors and younger, and older age groups, in the light of ageing the use of medical services: a longitudinal analysis. Soc Sci Med. 1999;48:1363–1372. populations, are needed. Future studies need better [6] van Steenkiste B, Knevel MF, van den Akker M, et al. data on psychological diseases, vulnerability, life- Increased attendance rate: BMI matters, lifestyles don’t. events, substance abuse, etc. as these may be import- Results from the Dutch SMILE study. Fam Pract. ant GP determinants. Finally, this study will help 2010;27:632–637. deepen understanding of gender differences in GP [7] Vos HMM, Schellevis FG, van den Berkmortel H, et al. attendance. GPs should be aware in daily clinical Does prevention of risk behaviour in primary care SCANDINAVIAN JOURNAL OF PRIMARY HEALTH CARE 249 require a gender-specific approach? A cross-sectional [18] Larsen FB, Nordvig L. Helbred og sygdom. In: Hvordan study. Fam Pract. 2013;30:179–184. har du det? Sundhedsprofil for region og kommuner, [8] Vedsted P, Fink P, Sørensen HT, et al. Physical, mental voksne [How do you feel? Health profile for regions and municipalities, adult population]. Denmark: Region and social factors associated with frequent attendance Midtjylland Center for Folkesundhed; 2008. pp. 20–66. in Danish general practice. A population-based cross- [19] Christensen AI, Ekholm O, Davidsen M, et al. Sundhed sectional study. Soc Sci Med. 2004;59:813–823 og Sygelighed i Danmark 2010 [Health and Morbidity [9] Vedsted P, Olesen F. Social environment and frequent in Denmark 2010]. Statens Inst. Folk. Syddansk Univ. attendance in Danish general practice. Br J Gen Pract. 2005;55:510–515. [20] Sundhedsstyrelsen. Determinanter som er påvirket af [10] Smits FTM, Mohrs JJ, Beem EE, et al. Defining frequent social position [Determinants affected by social pos- attendance in general practice. BMC Fam Pract. ition]. Ulighed i Sundh. – e´rsager og indsatser. 2011. 2008;9:21 p. 74–75. [11] Smits FT, Brouwer HJ, Zwinderman AH, et al. [21] Ministeriet for Sundhed og Forebyggelse. Ulighed i Predictability of persistent frequent attendance in pri- Sundhed [Inequity in Health] [Internet]; [cited 2016 mary care: a temporal and geographical validation Jul 4]. 2013. p. 31. Available from: http://www.sum.dk/ study. PLoS One. 2013;8:e73125. /media/Filer-Publikationer_i_pdf/2013/Ulighed-i- [12] Koskela T-H, Ryynanen O-P, Soini EJ. Risk factors for sundhed-2013/Ulighed-i-sundhed-marts-2013.ashx persistent frequent use of the primary health care [22] Rortveit G, Daltveit AK, Hannestad YS, et al. Urinary services among frequent attenders: a Bayesian incontinence after vaginal delivery or cesarean section. approach. Scand J Prim Health Care. 2010;28:55–61. N Engl J Med. 2003;348:900–907. [13] Gupta ND, Greve J. Overweight and obesity and the [23] Chiaffarino F, Chatenoud L, Dindelli M, et al. utilization of primary care physicians. Health Econ. Reproductive factors, family history, occupation and 2011;20:53–67. risk of urogenital prolapse. Eur J Obstet Gynecol [14] Andersen JS, Olivarius NDF, Krasnik A. The Danish Reprod Biol. 1999;82:63–67. National Health Service Register. Scand J Public [24] Vedsted P, Fink P, Olesen F, et al. Psychological dis- Health. 2011;39:34–37. tress as a predictor of frequent attendance in family [15] De Fine Olivarius N. The Danish National Health practice: a cohort study. Psychosomatics. 2001;42: Service Register a tool for primary health care 416–422. research. Dan Med Bull. 1997;44:449–453. [25] Rosendal M, Olesen F, Fink P. Management of medic- [16] Tjønneland A, Olsen A, Boll K, et al. Study design, ally unexplained symptoms. BMJ. 2005;330:4–5. exposure variables, and socioeconomic determinants [26] Vedsted P, Olesen F. Brug af dansk almen praksis i of participation in Diet, Cancer and Health: a popula- dagtid [Use of Danish general practice during day- tion-based prospective cohort study of 57,053 men time] English summary. Ugeskr Læger. and women in Denmark. Scand J Public Health. 2005;167:3280–3282. 2007;35:432–441 [27] Campaign ‘‘Hold øje mand!’’ by The Danish Cancer [17] Gjerstorff ML. The Danish Cancer Registry. Scand J Society [Internet]; [cited 2016 Jul 4]. Available from: Public Health. 2011;39:42–45. https://www.cancer.dk/holdoejemand http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Scandinavian Journal of Primary Health Care Taylor & Francis

Determinants related to gender differences in general practice utilization: Danish Diet, Cancer and Health Cohort

Loading next page...
 
/lp/taylor-francis/determinants-related-to-gender-differences-in-general-practice-m0CAkvMBpw

References (64)

Publisher
Taylor & Francis
Copyright
© 2016 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
ISSN
0281-3432
DOI
10.1080/02813432.2016.1207141
Publisher site
See Article on Publisher Site

Abstract

SCANDINAVIAN JOURNAL OF PRIMARY HEALTH CARE, 2016 VOL. 34, NO. 3, 240–249 http://dx.doi.org/10.1080/02813432.2016.1207141 RESEARCH ARTICLE Determinants related to gender differences in general practice utilization: Danish Diet, Cancer and Health Cohort a b c a Jeanette Therming Jørgensen , John Sahl Andersen , Anne Tjønneland and Zorana Jovanovic Andersen a b Centre for Epidemiology and Screening, Department of Public Health, University of Copenhagen, Copenhagen, Denmark; Section of General Practice, Department of Public Health, University of Copenhagen, Copenhagen, Denmark; Danish Centre for Cancer Research, Danish Cancer Society, Copenhagen, Denmark ABSTRACT ARTICLE HISTORY Objective: This study aims to describe the determinants related to gender differences in the GP Received 2 October 2015 Accepted 4 April 2016 utilization in Danish population aged 50–65 years. Design: Cohort-based cross-sectional study. KEYWORDS Setting: Danish general practice. Cohort; Denmark; gender; Subjects: Totally, 54,849 participants of the Danish Diet, Cancer and Health cohort (50–65 years). general practice; health Main outcome measures: The sum of cohort members’ face-to-face consultations with general service use; lifestyle; practitioner (GP) at the cohort baseline year (1993–1997). We obtained data on GP visits from the unemployment Danish National Health Service Register at the cohort baseline (1993–1997), when information on lifestyle (smoking, body mass index (BMI), alcohol use, physical activity), medical conditions (som- atic and mental), employment, education, gravidity, and hormone therapy (HT) use was collected by questionnaire. Results: Women had on average 4.1 and men 2.8 consultations per year. In a crude model, women had 47% higher rate of GP visits than men (incidence rate ratio: 1.47; 95% Confidence Interval: 1.45–1.50), which remained unchanged after adjustment for lifestyle, socio-demographic and medical factors, but attenuated to 18% (1.18; 1.13–1.24) after adjustment for female factors (gravidity and post-menopausal HT. In a fully adjusted model, subjects with hypertension (1.63; 1.59–1.67), mental illness (1.63; 1.61–1.66), diabetes (1.56; 1.47–1.65), angina pectoris (1.28; 1.21–1.34), and unemployed persons (1.19; 1.18–1.21) had highest rates of GP visits. Conclusions: Gravidity and HT use explain a large proportion, but not all of the gender differ- ence in GP utilization. Medical conditions (somatic and mental) and unemployment are the main determinants of GP utilization in men and women, while lifestyle has minor effect. KEY POINTS Female gender remained a dominant determinant of GP utilization, after adjustment for life- style, socio-demography, medical and gender specific factors, with females consulting their GP 18% more often than males. Female reproductive factors (use of postmenopausal hormone therapy and gravidity) explained a large proportion of the gender variation in use of GP. Strongest determinants for GP use among Danish adults aged 50–65 years were the presence of medical conditions (somatic and mental) and unemployment, while lifestyle factors (e.g., body mass index, alcohol consumption and smoking) had minor effect. Introduction of reasons for encounters. The background for the util- ization of services in the primary health care system is Knowledge about determinants of health care utiliza- of interest to the general practitioners (GPs) and for tion is essential in the daily clinical work and planning public health in general. Existing studies on utilization of a health care system, in order to efficiently meet the of Danish GPs focused primarily on equity in access to needs of the population. General practice is character- health care or frequent attenders.[1,2] Although it is ized by free access and nonselected patients, resulting in a broad spectrum of services provided and a variety well documented that women contact GP more often CONTACT Jeanette Therming Jørgensen jethe@sund.ku.dk Centre for Epidemiology and Screening, Department of Public Health, University of Copenhagen, Copenhagen, Nørregade 10, 1165 København, Denmark Supplemental data for this article can be accessed here. 2016 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. SCANDINAVIAN JOURNAL OF PRIMARY HEALTH CARE 241 than men, little is known about the determinants of this fear of death, alcohol abstinence, low patient satisfac- difference. Vedsted et al. reported that women utilized tion and irritable bowel syndrome as determinants of GPs 47% more often than men,[3] but did not have persistent frequent attendance.[12] Gupta and Greve found that overweight and obesity’s effect on GP use data to examine whether lifestyle and gender-specific in Denmark affected GP use among frequent users factors (reproductive-related contacts, gravidity and use only,[13] suggesting variation in characteristics and of post-menopausal hormone therapy (HT)) explained underlying the mechanism between use of GP among difference. Krasnik et al. found that gender and health frequent attenders and the general population. characteristics, especially functional status and chronic The aim of this study was to examine the determi- diseases, are most important determinants of GP use in nants related to gender differences in use of GP in Denmark, whereas social factors had very little impact, Danish population aged 50–65 years. Furthermore, we but lacked data on gender-specific factors.[1] specifically examined whether gender differences in Musculoskeletal, psychological, and respiratory prob- GP utilization persisted when adjusting for lifestyle, lems have been identified by Moth et al. as the most marital, occupational and educational status, urbaniza- common reasons for encounter in Danish primary care tion, pre-existing diseases (somatic and mental), and during the period of 1993–2009, but also lacked infor- female-specific factors (gravidity and HT use). mation on gender-specific factors.[4] Green et al. investigated gender differences in med- ical care utilization in the United States (all health serv- Materials and Methods ices contacts) and found that the gender differences Study population persisted but were reduced when controlling for gen- der-specific utilization,[5] however the effect on GP con- We have linked data on 57,053 participants of the tacts alone has not yet been quantified. Furthermore, Danish Diet, Cancer, and Health (DCH) cohort to the Green et al. identified attitudinal and behavioral factors Danish National Health Service Register (NHSR) to as important predictors of medical care utilization, obtain data on GP visits. The study was conducted as whereas health knowledge did not affect health care a cohort-based cross-sectional study, where informa- utilization.[5] Evidence on the effect of lifestyle factors tion on GP contacts and confounder information were on GP utilization is conflicting. Body mass index (BMI) collected within the same year, at cohort baseline in has previously been linked to increased GP use, whereas 1993–1997. other lifestyle factors such as diet, physical activity, smoking and alcohol consumption did not influence Danish National Health Service Register attendance rates in the Dutch population.[6] This contra- dicts results from another Dutch study examining risk NHSR is a nationwide register containing information behaviors and use of GP services related to gender,[7] on all contacts within primary health care in Denmark.[14] The register was established for adminis- where no association between BMI and GP use was found. Finally, Vos et al. reported lower GP attendance trative purposes in 1984, and data has been available rate among smoking men compared to nonsmoking for research purposes since 1990.[14,15] In addition to men, and the reverse association in women.[7] citizen-related data, records in NHSR contain informa- A large body of literature examined determinants of tion on the health care provider and the type of ser- frequent attendance to GP, defined typically as the vice provided (e.g., telephone consultation, home-visit, age- and gender-stratified top 10th percentile of GP face-to-face visit, preventive consultation). Reasons for attenders,[2,8–10] and identified social factors encounter or information on specific health problems (unemployment, divorce, low education, and social is only available through NHSR to a limited extent, in support), psychological distress, and physical diseases terms of services codes (e.g., prescription renewal, add- as main determinants.[7] Results from a Dutch study itional services codes), and no diagnoses are available. showed that age, chronic illness, psychosocial prob- GP visits in this study were defined as sum of all face- lems, and analgesics prescriptions, moderately pre- to-face contacts at the year of cohort baseline dicted persistent frequent attendance, whereas gender, (1993–1997) including consultations at GPs office and medically unexplained symptoms, use of psychoactive home-visits during opening hours, while telephone drugs and prescription of antibiotics did not affect fre- consultations and prescription renewals were excluded. quent attendance.[11] Furthermore, information on cohort members visits to Additionally, Koskela et al. identified female gender, psychologist and psychiatrists within primary health obesity (BMI >30 kg/m ), former frequent attendance, sector before cohort baseline and at baseline year 242 J. T. JØRGENSEN ET AL. (since 1990) was obtained from NHSR and used as an before-mentioned diseases. Information on occupa- indicator of pre-existing mental disorders. tional status was constructed based on self-reported levels of physical activity at work. Participants classified themselves according to four different work categories Danish Diet, Cancer, and Health Cohort or as ‘‘have not been working the past year’’, repre- The DCH cohort, described in detail elsewhere,[16]is senting unemployed and individuals outside labor part of the European Prospective Investigation into force. Thus, in this study definition ‘‘unemployed’’ Cancer and Nutrition study (EPIC) and used widely for implies both unemployed and individuals not in the research into lifestyle factors, with focus on diet, labor force. and the risk of cancer and other chronic diseases. Briefly, in 1993–1997 a total of 160,725 individuals, Statistical analysis aged between 50 and 64, born in Denmark, living in Copenhagen or Aarhus, and with no previous records We used negative binomial regression model to exam- in the Danish Cancer Registry,[17] were invited to par- ine association between total number of GP visits at ticipate in the DCH cohort study, and 57,053 individu- the year of cohort baseline and abovementioned cova- als responded and participated in the study. Cohort riates, in five separate models: (1) Model 1, a crude participation involved answering comprehensive ques- model; (2) Model 2, a model adjusted for age, gender tionnaires and interviews concerning dietary intake and lifestyle factors (BMI, alcohol consumption, smok- and lifestyle factors that are known and potential risk ing, physical activity); (3) Model 3, a model adjusted factors in the development of cancer. Additionally, for age, gender, lifestyle and social factors (marital and anthropometric measurements were taken during a occupational status, education, and urbanization); (4) physical examination and various biological materials Model 4, a model adjusted for age, gender, lifestyle, were collected. Due to large cohort sample size and social and medical factors (pre-existing diseases and extensive data collection (e.g., biological material, history of cancer in the family); (5) Model 5, a fully anthropometric measurements, etc.) participants’ base- adjusted model, adjusted for age, gender, lifestyle fac- line information was collected over several years tors, social factors, medical factors and female-specific (1993–1997). The following potential determinants of factors (gravidity, previous or current use of HT). GP contact were obtained from the DCH cohort: gen- Additionally, separate Models 1 and 5 were fit for men der, age, weight, height, alcohol consumption, smoking and women separately. Interaction terms between gen- status, leisure time physical activity, marital status, der and all other covariates were introduced in fully occupational status, educational status, urbanization, adjusted model one at a time, to test potential effect pre-existing somatic diseases, history of cancer in the modification. All analyses were performed as complete- family, previous or current use of HT, number of preg- subjects-analysis. As sensitivity analyses, two additional nancies. Information on pre-existing mental disease models were fitted using alternative outcome including was obtained from the NHSR as described above. Age GP visits in ‘‘near’’ future: participants’ number of visits refers to the participants’ age at the date of the phys- 1 year after baseline (1994–1998) and in 5 years post- ical examination, and a part from age, gender, weight, baseline (1998–2002). Results are available in online height and urbanization, all other variables collected supplement (Table A). Results are presented as inci- from DCH cohort are self-reported. Urbanization was dence rate ratio (IRR) with 95% confidence intervals dichotomized into urban (Copenhagen, Frederiksberg (95%CI). Negative binomial regression procedures or Aarhus municipality) and suburban (remaining sub- (GENMOD) in SAS 3.9 (Copenhagen, Denmark) were urban municipalities around Copenhagen and Aarhus). used to conduct the analyses. The self-reported daily alcohol intake in grams was dichotomized (below the recommended limit and Results above the recommended limit), according to Danish Health and Medicines Authority’s recommendation on Of the 57,053 DCH cohort participants, 571 were weekly alcohol consumption (females: 168 g, male: excluded due to cancer diagnosis prior to cohort base- 252 g) at the time of cohort baseline. Prevalence of line. This was inclusion criteria for the DCH cohort, as pre-existing diseases was estimated based on partici- the original aim of the cohort was to study association pants reporting either being diagnosed with heart between diet and incidence of cancer. Furthermore, attack, high cholesterol, angina pectoris, stroke, hyper- 1633 were excluded due to missing values on one or tension, diabetes, gallstones and intestinal polyps more covariates of interest, leaving 54,849 cohort or self-reported use of medication to treat members for analyses in this paper. Of these 54,849 SCANDINAVIAN JOURNAL OF PRIMARY HEALTH CARE 243 cohort members, 28,643 (52.2%) were women found between GP visits and living in urban area. For (Table 1). A total of 188,709 GP contacts were regis- women, number of pregnancies (1.09; 1.04–1.14) were tered in NHSR at the cohort baseline year, giving 3.44 weakly positively associated with GP visits while previ- mean contacts per DCH participant. A total of 11,192 ous and current users of HT had 27% higher rate of GP (20.4%) cohort participants had no registered GP con- visits (1.27; 1.24–1.30) than nonusers. Further adjust- tacts in NHSR at the baseline year, of whom, majority ment for removal of lump in the breast, hysterectomy, were men (65.2%). and removal of one or both ovaries did not change The average number of visits to GP at cohort base- the estimated gender difference, and these were line in 1993–1997 was 4.06 (standard deviation 4.54) removed from the final model as they did not have for women and 2.76 (4.00) for men (Table 1). Number any effect on the number of GP visits. Significant effect of GP visits increased with age, and was higher in modification by gender was identified for a number of underweight and obese participants, smokers, physic- factors: age, alcohol consumption, smoking, physical ally inactive participants, those who drank alcohol activity, occupational status, heart attack, high choles- below recommended limit, unemployed, participants terol, angina pectoris, hypertension, diabetes intestinal with lower education, and those with pre-existing dis- polyps and mental disorders (Table 3). Lifestyle seemed ease (Table 1). GP utilization did not differ by history to have more pronounced effect on GP use in men of cancer in the family, while married and unmarried than women in this age group, as increasing age and participants had less GP contacts than divorced or wid- smoking lead to higher increase in use of GP in men owed participants (Table 1). The GP contacts in women than women. High cholesterol, hypertension, diabetes, increased with number of pregnancies and HT use intestinal polyps and mental illness led to higher (Table 1). There was no variation in GP use by urbanic- increase in GP visits in men than women. ity level. However, women had consistently, statistically significantly more GP visits than men (Table 1), even Discussion those with pre-existing diseases, e.g., diabetic women contacted GP on average 7.41, while diabetic men con- Statement of principal findings tacted GP 5.71 times per year (p< 0.001). This study yielded three major findings: (1) female gen- In a crude model, women had 47% higher rate of der is a dominant determinant of GP utilization in the GP visits than men (IRR: 1.47; 95% confidence interval age group 50–65 years, even after controlling for life- 1.45–1.50) (Table 2). This gender variation persisted style, socio-demographic, medical (somatic and men- when lifestyle, socio-demographic and medical factors tal), and female reproductive factors, with women were added to the model, but attenuated to 18% consulting GP 18% more than men; (2) female-repro- when female-specific factors were included in a fully ductive factors (gravidity and postmenopausal HT use) adjusted model (1.18; 1.13–1.24). In a fully adjusted explained a large amount of the variation in the GP model, we found no association between age and GP use; (3) pre-existing medical conditions (somatic and visits. Alcohol consumption was weakly, but signifi- mental), unemployment, and HT use in women were cantly inversely associated with GP visits and individu- major determinants of GP utilization. als drinking above the weekly recommended limit had 8% fewer annual visits (0.92; 0.91–0.94) than those adhering recommendations. Current and previous Strengths and weaknesses of the study smoking was weakly positively associated with GP vis- This study benefited from an internationally unique its (1.09; 1.07–1.12 and 1.10; 1.07–1.12, respectively), possibility to link a large Danish cohort with 57,053 while there was weak or no effect of BMI, physical participants recruited from general population to the activity, or marital status. Employment was associated national registry of primary health care utilization, with with lower use of GP (0.81; 0.79–0.82). Similarly, sub- objective assessment of GP utilization. DCH cohort has jects with more than four years of higher education high quality data on lifestyle, education, diseases, and had fewer GP visits than those with no vocational measured height and weight. DCH has been utilized in training (0.70; 0.68–0.72). Pre-existing diseases were a number of epidemiological studies on aetiology of the strongest determinants of GP visits, hypertension (1.63; 1.59–1.67), mental disorders (1.63; 1.61–1.66), dia- cancer and other chronic diseases, related to lifestyle, betes (1.56; 1.47–1.65), angina pectoris (1.28; 1.21–1.34) socio-economic and reproductive factors. However, this and stroke (1.25; 1.16–1.34), as leading determinants of is the first study linking the DCH cohort to the Danish GP use. Having a history of cancer in the family, how- NHSR, and obtaining cohort participants’ information ever had no effect on visits to GP. No association was on GP use. NHSR is considered to be of high validity 244 J. T. JØRGENSEN ET AL. Table 1. Distribution of baseline characteristics among men and women in the DCH cohort (n¼ 54,849). Number of GP visits Female Male N (%) Mean (std) N (%) Mean (std) p Value Gender 28,643 (52.2) 4.06 (4.54) 26,206 (47.8) 2.76 (4.00) <0.0001 Age – – – – – 50–54 11,917 (21.7) 3.84 (4.32) 11,243 (20.5) 2.39 (3.67) <0.0001 55–59 8899 (16.2) 4.18 (4.82) 8129 (14.8) 2.85 (4.28) <0.0001 60–65 7827 (14.3) 4.27 (4.51) 6834 (12.5) 3.27 (4.12) <0.0001 BMI – – – – – Underweight (>18.5 kg/m ) 355 (0.65) 4.43 (7.55) 66 (0.12) 3.59 (4.91) <.2485 Normal weight (18.5–24.9 kg/m ) 14,481 (26.4) 3.63 (4.14) 9110 (16.6) 2.38 (3.83) <0.0001 Overweight (25–29.9 kg/m ) 9805 (17.9) 4.24 (4.45) 13,078 (23.8) 2.73 (3.79) <0.0001 Obese (30 kg/m ) 4002 (7.30) 5.14 (5.45) 3952 (7.21) 3.73 (4.80) <0.0001 Alcohol consumption – – – – – Below the recommended weekly limit 17,456 (31.8) 4.32 (4.96) 15,114 (27.6) 2.81 (4.30) <0.0001 Above the recommended weekly limit 11,187 (20.4) 3.66 (3.74) 11,092 (20.2) 2.70 (3.55) <0.0001 Smoking status – – – – – Never 12,600 (23.0) 3.84 (4.36) 6738 (12.3) 2.39 (3.41) <0.0001 Previous 6731 (12.3) 4.20 (4.40) 9095 (16.6) 2.94 (3.82) <0.0001 Current 9312 (17.0) 4.26 (4.84) 10,373 (18.9) 2.85 (4.47) <0.0001 Physical activity – – – – – No leisure time physical activity 11,799 (21.5) 4.34 (4.94) 13,425 (24.5) 3.01 (4.37) <0.0001 Physical active in leisure time 16,844 (30.7) 3.87 (4.22) 12,781 (23.3) 2.50 (3.56) <0.0001 Marital status – – – – – Unmarried 1775 (3.24) 3.81 (4.92) 1498 (2.73) 2.93 (4.24) <0.0001 Divorced 5538 (10.1) 4.44 (4.97) 3706 (6.76) 3.18 (4.99) <0.0001 Widow/widower 2422 (4.42) 4.37 (5.25) 595 (1.08) 3.19 (4.11) <0.0001 Married 18,908 (34.5) 3.94 (4.25) 20,407 (37.2) 2.66 (3.77) <0.0001 Occupational status – – – – – Unemployed 8018 (14.6) 4.99 (5.69) 4051 (7.4) 4.15 (5.31) <0.0001 Employed 20,625 (37.6) 3.70 (3.94) 22,155 (40.4) 2.51 (3.66) <0.0001 Educational status – – – – – No vocational training 5498 (10.0) 4.95 (5.45) 2606 (4.75) 3.49(4.39) <0.0001 Higher education, <3 years 8943 (16.3) 4.03 (4.19) 3504 (6.39) 2.88(3.61) <0.0001 Higher education, 3–4 years 10,992 (20.0) 3.84 (4.48) 11,131 (20.29) 2.80(4.27) <0.0001 Higher education, >4 years 3210 (5.85) 3.38 (3.65) 8965 (16.34) 2.45(3.63) <0.0001 Municipal – – – – – Suburban 12,712 (23.2) 4.02 (4.44) 11,736 (21.40) 2.74(4.03) <0.0001 Urban 15,931 (29.1) 4.10 (4.61) 14,470 (26.38) 2.78(3.98) <0.0001 Medical conditions – – – – – Heart attack 241 (0.44) 6.87 (8.92) 880 (1.60) 5.00 (5.75) <0.0023 High cholesterol 1797 (3.28) 5.56 (5.14) 2288 (4.17) 4.18 (4.46) <0.0001 Angina pectoris 639 (1.17) 6.52 (6.70) 1010 (1.84) 5.11 (5.52) <0.0001 Stroke 283 (0.52) 6.29 (5.10) 427 (0.78) 5.09 (5.26) <0.0027 Hypertension 4957 (9.04) 5.88 (5.47) 3953 (7.21) 5.06 (5.80) <0.0001 Diabetes 428 (0.78) 7.41 (9.51) 705 (1.29) 5.71 (6.20) <0.0010 Gallstones 2049 (3.74) 5.33 (5.56) 575 (1.05) 3.66 (4.32) <0.0001 Intestinal polyps 876 (1.60) 5.11 (5.35) 1043 (1.90) 3.46 (4.06) <0.0001 Mental illness 2082 (3.80) 6.07 (6.30) 915 (1.67) 5.37 (9.25) <0.0001 History of cancer in the family 14,348 (26.2) 4.14 (4.48) 11,926 (21.7) 2.77 (3.85) <0.0001 Hormone therapy use – – – – – Never 16,057 (29.3) 3.57 (4.03) – – – Previous or current user 12,586 (23.0) 4.69 (5.04) – – – Number of pregnancies – – – – – 0 2479 (4.52) 3.68 (4.22) – – – 1 or more 26,164 (47.7) 4.13 (4.59) – – – Consultations and home-visits. t statistics for univariate associations. due to its administrative purpose, as GPs financial registered contacts with psychologist in primary care reimbursement depends on the accurate registration or referral to psychiatrist. Information on employment of GP services in the register.[15] status was obtained from question on physical activ- This study has several limitations. Information on ity at work, where cohort participants could report pre-existing somatic diseases was self-reported, and being unemployed or not in labor force during the information on the severity of disease and mental last year. Thus, it was not possible to distinguish diseases was unavailable. Therefore, we defined a between unemployed, early retired or individuals on proxy of mental disorders based on the NHSR disability pension. We also lacked information on SCANDINAVIAN JOURNAL OF PRIMARY HEALTH CARE 245 Table 2. General practice utilization determinants among men and women in the DCH cohort (n¼ 54,849). Number of GP visits b c d e f Model 1 Model 2 Model 3 Model 4 Model 5 IRR (95% CI) IRR (95% CI) IRR (95% CI) IRR (95% CI) IRR (95% CI) Gender – – – – – Male (n ¼ 26,206) 1.00 1.00 1.00 1.00 1.00 Female (n ¼ 28,643) 1.47 (1.45–1.50) 1.56 (1.53–1.59) 1.44 (1.41–1.47) 1.44 (1.41–1.46) 1.18 (1.13–1.24) Age –– ––– 50–54 1.00 1.00 1.00 1.00 1.00 55–59 1.13 (1.11–1.16) 1.12 (1.10–1.14) 1.08 (1.06–1.10) 1.06 (1.03–1.08) 1.04 (1.02–1.06) 60–65 1.22 (1.19–1.25) 1.19 (1.16–1.22) 1.06 (1.04–1.09) 1.03 (1.00–1.05) 1.02 (1.00–1.04) BMI –– ––– Underweight (>18.5 kg/m ) 1.00 1.00 1.00 1.00 1.00 Normal weight (18.5–24.9 kg/m ) 0.79 (0.72–0.88) 0.85 (0.77–0.93) 0.88 (0.79–0.97) 0.88 (0.80–0.97) 0.87 (0.79–0.96) Overweight (25–29.9 kg/m ) 0.91 (0.82–1.01) 0.97 (0.87–1.07) 0.98 (0.89–1.08) 0.95 (0.86–1.05) 0.94 (0.86–1.04) Obese (30 kg/m ) 1.17 (1.06–1.30) 1.23 (1.11–1.36) 1.20 (1.09–1.33) 1.08 (0.98–1.19) 1.09 (0.99–1.20) Alcohol consumption – – – – – Below the recommended weekly limit 1.00 1.00 1.00 1.00 1.00 Above the recommended weekly limit 0.90 (0.89–0.92) 0.91 (0.89–0.93) 0.94 (0.92–0.96) 0.92 (0.91–0.94) 0.92 (0.91–0.94) Smoking status – – – – – Never 1.00 1.00 1.00 1.00 1.00 Previous 1.13 (1.11–1.16) 1.13 (1.11–1.16) 1.12 (1.09–1.15) 1.11 (1.08–1.13) 1.10 (1.07–1.12) Current 1.13 (1.11–1.16) 1.15 (1.13–1.18) 1.09 (1.07–1.12) 1.10 (1.08–1.13) 1.09 (1.07–1.12) Physical activity – – – – – No leisure time physical activity 1.00 1.00 1.00 1.00 1.00 Physical active in leisure time 0.87 (0.85–0.88) 0.90 (0.89–0.92) 0.93 (0.92–0.95) 0.95 (0.93–0.97) 0.95 (0.93–0.97) Marital status – – – – – Unmarried 1.00 – 1.00 1.00 1.00 Divorced 1.12 (1.07–1.17) 1.10 (1.06–1.15) 1.09 (1.05–1.14) 1.06 (1.01–1.10) Widow/widower 1.04 (0.99–1.10) 1.03 (0.97–1.08) 1.03 (0.98–1.08) 1.00 (0.95–1.05) Married 0.96 (0.92–1.00) 0.97 (0.94–1.01) 1.00 (0.96–1.04) 0.97 (0.94–1.01) Occupational status – – – – – Unemployed 1.00 – 1.00 1.00 1.00 Employed 0.70 (0.68–0.72) 0.75 (0.73–0.77) 0.80 (0.78–0.82) 0.81 (0.79–0.82) Educational status – – – – – No vocational training 1.00 – 1.00 1.00 1.00 Higher education, <3 years 0.82 (0.80–0.85) 0.93 (0.90–0.96) 0.91 (0.88–0.93) 0.91 (0.88–0.93) Higher education, 3–4 years 0.80 (0.78–0.82) 0.92 (0.89–0.95) 0.89 (0.87–0.91) 0.89 (0.87–0.91) Higher education, >4 years 0.70 (0.68–0.72) 0.88 (0.85–0.91) 0.81 (0.78–0.83) 0.80 (0.78–0.83) Municipal – – – – – Suburban 1.00 – 1.00 1.00 1.00 Urban 1.02 (1.01–1.04) 0.98 (0.96–1.00) 1.00 (0.98–1.01) 1.00 (0.99–1.02) Medical conditions – – – – – Heart attack 1.77 (1.66–1.88) – – 1.19 (1.12–1.27) 1.19 (1.12–1.27) High cholesterol 1.48 (1.43–1.53) – – 1.20 (1.16–1.24) 1.20 (1.17–1.24) Angina pectoris 1.74 (1.66–1.83) – – 1.28 (1.21–1.34) 1.28 (1.21–1.34) Stroke 1.68 (1.56–1.82) – – 1.24 (1.15–1.33) 1.25 (1.16–1.34) Hypertension 1.81 (1.77–1.85) – – 1.63 (1.60–1.67) 1.63 (1.59–1.67) Diabetes 1.99 (1.87–2.11) – – 1.54 (1.46–1.63) 1.56 (1.47–1.65) Gallstones 1.31 (1.26–1.37) – – 1.18 (1.14–1.23) 1.16 (1.12–1.21) Intestinal polyps 1.25 (1.19–1.31) – – 1.17 (1.12–1.22) 1.16 (1.11–1.21) Mental illness 1.71 (1.64–1.77) – – 1.61 (1.55–1.67) 1.63 (1.61–1.66) History of cancer in the family 1.02 (1.00–1.04) – – 1.03 (1.01–1.04) 1.02 (1.01–1.04) Hormone therapy use – – – – – Never 1.00 – – – 1.00 Previous or current user 1.30 (1.27–1.33) 1.27 (1.24–1.30) Number of pregnancies – – – – – 0 1.00 – – – 1.00 1 or more 1.13 (1.08–1.18) – – – 1.09 (1.05–1.14) Consultations and home-visits. Adjusted for gender and age. Adjusted for gender, age and lifestyle factors (BMI, smoking, alcohol consumption, physical activity). Adjusted for model 2 and socio-demographic factors (marital, occupational, educational status, and urbanization). Adjusted for model 3 and medical factors (heart attack, high cholesterol, angina pectoris, stroke, hypertension, diabetes, intestinal polyps, gallstones, and history of cancer in the family). Adjusted for model 4 and female reproductive factors (use of HT, and number of pregnancies). vulnerability or life-events (death in family, divorce, nonparticipants,[16] as well as the study is based on loss of job, etc.), that may have affected GP use.[1] data from 1993–1997 and on the specific age group Another weakness is that participants in DCH 50–65, limiting generalizability of the results to gen- had higher education and income than eral population and other age groups. 246 J. T. JØRGENSEN ET AL. Table 3. General practice utilization determinants in the DCH cohort (n¼ 54,849) by gender. Number of GP visits Female (n¼ 28643) Male (n¼ 26206) b c b c Crude model Adjusted model Crude model Adjusted model IRR (95% CI) IRR (95% CI) IRR (95% CI) IRR (95% CI) p Value Age –– –– – 50–54 1.00 1.00 1.00 1.00 55–59 1.09 (1.06–1.12) 0.98(0.96–1.01) 1.19 (1.15–1.23) 1.12(1.08–1.15) <0.0001 60–65 1.11 (1.08–1.14) 0.92(0.90–0.95) 1.36 (1.32–1.42) 1.14(1.10–1.19) <0.0001 BMI –– –– – Underweight (>18.5 kg/m ) 1.00 1.00 1.00 1.00 Normal weight (18.5–24.9 kg/m ) 0.82 (0.74–0.91) 0.88(0.80–0.97) 0.65 (0.49–0.87) 0.81(0.62–1.06) 0.3728 Overweight (25–29.9 kg/m ) 0.95 (0.86–1.06) 0.96(0.87–1.06) 0.74 (0.56–0.99) 0.87(0.66–1.13) 0.3654 Obese (30 kg/m ) 1.16 (1.04–1.28) 1.07(0.97–1.18) 1.02 (0.76–1.35) 1.03(0.78–1.35) 0.7656 Alcohol consumption – – – – – Below the recommended weekly limit 1.00 1.00 1.00 1.00 Above the recommended weekly limit 0.85 (0.83–0.87) 0.89(0.87–0.91) 0.97 (0.94–1.00) 0.96(0.93–0.98) 0.0002 Smoking status – – – – – Never 1.00 1.00 1.00 1.00 Previous 1.09 (1.06–1.12) 1.07(1.04–1.10) 1.20 (1.15–1.24) 1.14(1.10–1.18) <0.0001 Current 1.11 (1.08–1.14) 1.07(1.04–1.10) 1.18 (1.14–1.23) 1.13(1.09–1.17) 0.0006 Physical activity – – – – – No leisure time physical activity 1.00 1.00 1.00 1.00 Physical active in leisure time 0.89 (0.87–0.91) 0.96(0.94–0.99) 0.84 (0.82–0.87) 0.95(0.93–0.98) 0.0534 Marital status – – – – – Unmarried 1.00 1.00 1.00 1.00 Divorced 1.16 (1.10–1.22) 1.05 (1.00–1.11) 1.07 (1.00–1.16) 1.06(0.99–1.14) 0.6682 Widow/widower 1.11 (1.05–1.18) 1.02 (0.96–1.09) 1.00 (0.89–1.12) 1.02(0.92–1.14) 0.3401 Married 1.03 (0.98–1.08) 0.99(0.94–1.04) 0.88 (0.83–0.94) 0.96(0.90–1.02) 0.3962 Occupational status – – – – – Unemployed 1.00 1.00 1.00 1.00 – Employed 0.74 (0.72–0.76) 0.83 (0.81–0.86) 0.63 (0.61–0.66) 0.75(0.72–0.78) <0.0001 Educational status – – – – – No vocational training 1.00 1.00 1.00 1.00 – Higher education, <3 years 0.82 (0.79–0.85) 0.89(0.86–0.92) 0.82 (0.77–0.87) 0.92(0.87–0.98) 0.5876 Higher education, 3–4 years 0.78 (0.76–0.81) 0.87(0.84–0.90) 0.81 (0.77–0.85) 0.91(0.86–0.95) 0.6859 Higher education, >4 years 0.69 (0.66–0.72) 0.80(0.77–0.84) 0.71 (0.67–0.75) 0.81(0.77–0.85) 0.3323 Municipal – – – – – Suburban 1.00 1.00 1.00 1.00 – Urban 1.02 (1.01–1.05) 1.01(0.99–1.03) 1.03 (1.00–1.06) 0.99(0.96–1.02) 0.1236 Medical conditions – – – – – Heart attack 1.67 (1.48–1.87) 1.14(1.01–1.27) 1.77 (1.64–1.92) 1.15(1.07–1.25) 0.0147 High cholesterol 1.39 (1.32–1.45) 1.19(1.14–1.24) 1.57 (1.49–1.65) 1.22(1.16–1.28) 0.0006 Angina pectoris 1.60 (1.49–1.72) 1.21(1.13–1.30) 1.83 (1.70–1.96) 1.30(1.21–1.40) 0.0003 Stroke 1.54 (1.38–1.71) 1.24(1.12–1.37) 1.76 (1.58–1.97) 1.19(1.07–1.32) 0.1519 Hypertension 1.59 (1.54–1.63) 1.46(1.42–1.50) 2.12 (2.04–2.20) 1.86(1.80–1.94) <0.0001 Diabetes 1.83 (1.68–2.00) 1.48(1.36–1.60) 2.09 (1.92–2.27) 1.57(1.45–1.70) 0.0096 Gallstones 1.33 (1.28–1.39) 1.18(1.13–1.23) 1.30 (1.18–1.43) 1.13(1.03–1.24) 0.7863 Intestinal polyps 1.26 (1.18–1.34) 1.12(1.05–1.19) 1.24 (1.15–1.34) 1.20(1.12–1.28) 0.0312 Mental illness 1.56 (1.50–1.63) 1.45(1.39–1.50) 2.03 (1.89–2.19) 1.87(1.75–2.01) <0.0001 History of cancer in the family 1.04 (1.01–1.06) 1.03(1.01–1.06) 1.00 (0.97–1.03) 1.01(0.98–1.04) 0.2401 Hormone therapy use – – – – – Never 1.00 1.00 – – – Previous or current user 1.31 (1.28–1.34) 1.29(1.26–1.32) – – – Number of pregnancies – – – – – 0 1.00 1.00 – – – 1 1.10 (1.05–1.16) 1.07(1.01–1.12) – – – 2 1.08 (1.03–1.12) 1.07(1.02–1.12) – – – 3 1.11 (1.06–1.16) 1.08(1.03–1.13) – – – 4 1.15 (1.10–1.21) 1.08(1.03–1.14) – – – 5 1.24 (1.17–1.32) 1.11(1.05–1.18) – – – 6 1.29 (1.19–1.39) 1.13(1.05–1.22) – – – 7 1.36 (1.20–1.53) 1.19(1.06–1.33) – – – 8 1.38 (1.19–1.61) 1.21(1.05–1.40) – – – Consultations and home-visits. Adjusted for age. Fully adjusted model. p Value for interaction between gender and given covariate. SCANDINAVIAN JOURNAL OF PRIMARY HEALTH CARE 247 Findings in relation to other studies lifestyle factors has previously been evaluated in rela- tion to frequent attendance in general practice in We found that 20.4% of the cohort members had no Denmark, we present novel estimates of an effect of face-to-face contacts with their GP at baseline year, alcohol use and smoking on GP use in the general consistent with previous results by Gupta and population, which needs to be reproduced. We found Greve.[13] Gender differences persisted but attenuated weak association between GP visits and physical activ- when female-specific factors were included in the ity and none with BMI, in agreement with Gupta and model, but 18% higher GP utilization by women Greve.[13] remained unexplained, in line with previous find- ings.[1,3,5] A U.S. study by Green and Pope reported Socio-demographic factors that female gender remained an independent long- term predictor of higher use of medical services when Employment status was a strong determinant of GP controlling for factors likely to cause gender differen- visits, with employed participants having 19% fewer ces.[5] Likewise, Krasnik et al., who restricted the study GP visits than unemployed. However, as mentioned to participants 50 years and older in order to exclude earlier, some of those classified as unemployed in our possible contacts related to children in parous women, study, may have been retired or on disability pension. still found higher rates of GP use in women.[1] It has Our finding is inconsistent with Krasnik et al., where been suggested that gender differences may be occupational status had no effect on GP use, after explained by differences in health perception and that adjusting for gender, health status (functional limita- women are more sensitive and thus more likely to tions, mental health, chronic diseases), vulnerability report symptoms than men.[3] Furthermore, it has and life-events.[1] Inconsistencies may be due to differ- been suggested that men and women have different ences in the age of participants in the two studies, disease patterns. While women have higher prevalence lack of adjustment for vulnerability and life-events in of non-threatening chronic diseases that are manage- our cohort, or lack of separation of information in our able in GP, men have higher prevalence of more cohort on retired and unemployed. We found a stron- severe, life-threatening chronic diseases requiring hos- ger effect of employment status in men than women, pital admission or treatment within the secondary in line with earlier observation that men are more vul- health care sector.[3,18] The Danish National Institute nerable after unemployment than women.[20] GP visits of Public Health showed in a population-based survey, were inversely associated with education, with least that the prevalence of mental illness is higher among number of visits among those with the highest com- women, in all age groups (from 16 years of age),[19] pleted education, in agreement with earlier findings of however, we find that gender differences in GP utiliza- a systematic decrease in GP contacts with higher edu- tion persist, even when analysis are adjusted for men- cation for both men and women.[21] tal disorders. Medical factors Lifestyle factors Pre-existing diseases were the strongest determinants Surprisingly, GP visits were weakly inversely associated of GP visits, including hypertension, mental disorders, with alcohol consumption, which contradicts results by diabetes, angina pectoris and stroke. This is consistent van Steenkiste et al.[6] Inconsistencies may be with Krasnik et al. who identified functional limitations explained by differences in the Danish and Dutch and chronic diseases as strong determinants of GP util- study populations, general drinking habits, and defin- ization, while general subjective health measures (own ition of alcohol consumption. The inverse association perception of general health and health compared to between drinking alcohol and visits to GP was margin- others) were insignificant.[1] The association between ally stronger in women than men in the present study. GP visits and all: heart attack, high cholesterol, angina Smoking was positively associated with GP visits, which pectoris, hypertension, diabetes, intestinal polyps, and is similar to previous findings by Vos et al. mental disorders were significantly modified by gen- Furthermore, Vos et al. reported lower attendance rates der, with higher number of visits in men than women. among smoking men and the reverse for women, we found both female and male smoker to have higher Female reproductive factors rates of GP visits than nonsmokers. Furthermore, we found significantly stronger association of smoking As estimated effect of 47% higher GP attendance rate with GP visits in men than women. As the effect of in women than men, attenuated the most after 248 J. T. JØRGENSEN ET AL. adjustment for gender-specific variables, including practice that women and men have different health number of pregnancies and previous or current use of seeking behaviors and beliefs. This is important in the HT, related to menopausal symptoms. We found that evaluation of symptoms and the communication with women with one or more previous pregnancies use the patients. It is of special significance in the age their GP 9% more than those with no history of preg- group 50–65 years, where incidence of major chronic nancies. Since GP contacts related to children were diseases and cancer is steeply increasing. It will be removed from the data and the study population is relevant to evaluate whether difference in GP utiliza- aged above 50 years, the influence of parity may seem tion can explain some of this gender difference in strange. Nevertheless, the higher attendance rates overall and cause specific mortality. Results of this among parous women in this age group may be study may furthermore be relevant for policy makers explained by long-term consequences of pregnancy or in primary health care to target specific groups of men delivery, such as problems related to incontinence, in order to raise awareness of importance of contact- bladder infections or genital prolapse.[22,23] This study ing GP with serious symptoms early. For example, presented novel result that a large amount, but not all recent campaign by Danish Cancer Society in Denmark of gender variations in GP utilization is explained by a has targeted men specifically to be more aware of female reproductive factor. early symptoms of colorectal cancer,[27] and to contact their GP early with these symptoms, due to men being at higher risk from and having considerably poorer sur- Meaning of the study: possible mechanisms and vival from colorectal cancer than women. implications In summary, some of the gender differences in health Disclosure statement seeking behavior remained unexplained, even after adjusting for gender-specific utilization and lifestyle The authors report no conflicts of interest. factors. However, psychological factors and preexisting disease may explain some of the remaining gender Ethical approval variation, which have been identified as important Relevant Danish ethical committees and Danish Data determinants of frequent attendance to GP.[24] Protection Agency have approved the study (J.nr. 2013-41- Furthermore, 20–30% of the symptoms presented in 1600), and written informed consent was provided by all par- general practice are classified as medically unexplained ticipants at recruitment. symptoms,[25] and they are more common in women, which may explain some of the gender variations in References GP visits. Although the total number of GP contacts has increased by 17% from 1992 to 2001 in Denmark, [1] Krasnik A, Hansen E, Keiding N, et al. Determinants of number of face-to-face GP contacts exclusively, which general practice utilization in Denmark. Dan Med Bull are used as outcome in this study, increased only DENMARK. 1997;44:542–546. [2] Vedsted P, Christensen MB. Frequent attenders in gen- slightly, by 5%.[26] Similarly, Moth et al. [4] reported eral practice care: a literature review with special refer- only small changes in the reasons for encounter with ence to methodological considerations. Public Health. GP between 1993 and 2009. This implies that current 2005;119:118–137. mechanisms that cause individuals to contact their GP [3] Vedsted P. Kønsforskelle i brug af sundhedsvæsnet most likely do not differ substantially from estimates [Gender differences in the use of the Danish health reported in this study, based on data in 1993–97. care system], English summary. Ugeskr Læger. 2007;169:2403–2408. Furthermore, this study is based on population of [4] Moth G, Olesen F, Vedsted P. Reasons for encounter 50–65-year-old individuals, but results seem rather con- and disease patterns in Danish primary care: changes sistent with related study in younger age groups over 16 years. Scand J Prim Health Care. (20–65) in Denmark.[13] Still, more studies with more 2012;30:70–75. recent data on GP use, and in different ages, both [5] Green CA, Pope CR. Gender, psychosocial factors and younger, and older age groups, in the light of ageing the use of medical services: a longitudinal analysis. Soc Sci Med. 1999;48:1363–1372. populations, are needed. Future studies need better [6] van Steenkiste B, Knevel MF, van den Akker M, et al. data on psychological diseases, vulnerability, life- Increased attendance rate: BMI matters, lifestyles don’t. events, substance abuse, etc. as these may be import- Results from the Dutch SMILE study. Fam Pract. ant GP determinants. Finally, this study will help 2010;27:632–637. deepen understanding of gender differences in GP [7] Vos HMM, Schellevis FG, van den Berkmortel H, et al. attendance. GPs should be aware in daily clinical Does prevention of risk behaviour in primary care SCANDINAVIAN JOURNAL OF PRIMARY HEALTH CARE 249 require a gender-specific approach? A cross-sectional [18] Larsen FB, Nordvig L. Helbred og sygdom. In: Hvordan study. Fam Pract. 2013;30:179–184. har du det? Sundhedsprofil for region og kommuner, [8] Vedsted P, Fink P, Sørensen HT, et al. Physical, mental voksne [How do you feel? Health profile for regions and municipalities, adult population]. Denmark: Region and social factors associated with frequent attendance Midtjylland Center for Folkesundhed; 2008. pp. 20–66. in Danish general practice. A population-based cross- [19] Christensen AI, Ekholm O, Davidsen M, et al. Sundhed sectional study. Soc Sci Med. 2004;59:813–823 og Sygelighed i Danmark 2010 [Health and Morbidity [9] Vedsted P, Olesen F. Social environment and frequent in Denmark 2010]. Statens Inst. Folk. Syddansk Univ. attendance in Danish general practice. Br J Gen Pract. 2005;55:510–515. [20] Sundhedsstyrelsen. Determinanter som er påvirket af [10] Smits FTM, Mohrs JJ, Beem EE, et al. Defining frequent social position [Determinants affected by social pos- attendance in general practice. BMC Fam Pract. ition]. Ulighed i Sundh. – e´rsager og indsatser. 2011. 2008;9:21 p. 74–75. [11] Smits FT, Brouwer HJ, Zwinderman AH, et al. [21] Ministeriet for Sundhed og Forebyggelse. Ulighed i Predictability of persistent frequent attendance in pri- Sundhed [Inequity in Health] [Internet]; [cited 2016 mary care: a temporal and geographical validation Jul 4]. 2013. p. 31. Available from: http://www.sum.dk/ study. PLoS One. 2013;8:e73125. /media/Filer-Publikationer_i_pdf/2013/Ulighed-i- [12] Koskela T-H, Ryynanen O-P, Soini EJ. Risk factors for sundhed-2013/Ulighed-i-sundhed-marts-2013.ashx persistent frequent use of the primary health care [22] Rortveit G, Daltveit AK, Hannestad YS, et al. Urinary services among frequent attenders: a Bayesian incontinence after vaginal delivery or cesarean section. approach. Scand J Prim Health Care. 2010;28:55–61. N Engl J Med. 2003;348:900–907. [13] Gupta ND, Greve J. Overweight and obesity and the [23] Chiaffarino F, Chatenoud L, Dindelli M, et al. utilization of primary care physicians. Health Econ. Reproductive factors, family history, occupation and 2011;20:53–67. risk of urogenital prolapse. Eur J Obstet Gynecol [14] Andersen JS, Olivarius NDF, Krasnik A. The Danish Reprod Biol. 1999;82:63–67. National Health Service Register. Scand J Public [24] Vedsted P, Fink P, Olesen F, et al. Psychological dis- Health. 2011;39:34–37. tress as a predictor of frequent attendance in family [15] De Fine Olivarius N. The Danish National Health practice: a cohort study. Psychosomatics. 2001;42: Service Register a tool for primary health care 416–422. research. Dan Med Bull. 1997;44:449–453. [25] Rosendal M, Olesen F, Fink P. Management of medic- [16] Tjønneland A, Olsen A, Boll K, et al. Study design, ally unexplained symptoms. BMJ. 2005;330:4–5. exposure variables, and socioeconomic determinants [26] Vedsted P, Olesen F. Brug af dansk almen praksis i of participation in Diet, Cancer and Health: a popula- dagtid [Use of Danish general practice during day- tion-based prospective cohort study of 57,053 men time] English summary. Ugeskr Læger. and women in Denmark. Scand J Public Health. 2005;167:3280–3282. 2007;35:432–441 [27] Campaign ‘‘Hold øje mand!’’ by The Danish Cancer [17] Gjerstorff ML. The Danish Cancer Registry. Scand J Society [Internet]; [cited 2016 Jul 4]. Available from: Public Health. 2011;39:42–45. https://www.cancer.dk/holdoejemand

Journal

Scandinavian Journal of Primary Health CareTaylor & Francis

Published: Jul 2, 2016

Keywords: Cohort; Denmark; gender; general practice; health service use; lifestyle; unemployment

There are no references for this article.