, Volume 59, Issue 2, pp 275–285 | Cite as

The impact of gender on the long-term morbidity and mortality of patients with type 2 diabetes receiving structured personal care: a 13 year follow-up study

  • Marlene Ø. Krag
  • Lotte Hasselbalch
  • Volkert Siersma
  • Anni B. S. Nielsen
  • Susanne Reventlow
  • Kirsti Malterud
  • Niels de Fine Olivarius



The aim of this study was to assess gender differences in mortality and morbidity during 13 follow-up years after 6 years of structured personal care in patients with type 2 diabetes mellitus.


In the Diabetes Care in General Practice (DCGP) multicentre, cluster-randomised, controlled trial ( registration no. NCT01074762), 1,381 patients newly diagnosed with type 2 diabetes were randomised to receive 6 years of either structured personal care or routine care. The intervention included regular follow-up, individualised goal setting and continuing medical education of general practitioners participating in the intervention. Patients were re-examined at the end of intervention. This observational analysis followed 970 patients for 13 years thereafter using national registries. Outcomes were all-cause mortality, incidence of diabetes-related death, any diabetes-related endpoint, myocardial infarction, stroke, peripheral vascular disease and microvascular disease.


In women, but not men, a lower HR for structured personal care vs routine care emerged for any diabetes-related endpoint (0.65, p = 0.004, adjusted; 73.4 vs 107.7 events per 1,000 patient-years), diabetes-related death (0.70, p = 0.031; 34.6 vs 45.7), all-cause mortality (0.74, p = 0.028; 55.5 vs 68.5) and stroke (0.59, p = 0.038; 15.6 vs 28.9). This effect was different between men and women for diabetes-related death (interaction p = 0.015) and all-cause mortality (interaction p = 0.005).


Compared with routine care, structured personal diabetes care reduced all-cause mortality and diabetes-related death in women but not in men. This gender difference was also observed for any diabetes-related outcome and stroke but was not statistically significant after extensive multivariate adjustment. These observational results from a post hoc analysis of a randomised controlled trial cannot be explained by intermediate outcomes like HbA1c alone, but involves complex social and cultural issues of gender. There is a need to rethink treatment schemes for both men and women to gain benefit from intensified treatment efforts.


Gender General practice Intervention Mortality Myocardial infarction Primary care Sex Stroke Type 2 diabetes mellitus 



Cardiovascular disease


Diabetes Control and Complications Trial


Diabetes Care in General Practice


General practitioner


Myocardial infarction


Men generally have a lower life expectancy than women [1] and people with type 2 diabetes are at increased risk of premature death, especially from cardiovascular disease (CVD), compared with people without diabetes [2, 3]. The relative protection against CVD and death observed in women vs men is reversed in patients with diabetes [4]. Women with diabetes have a proportionally greater risk of cardiovascular death and stroke than men with diabetes [2, 4]. Control of blood glucose [5], blood pressure [6] and cholesterol [7] may decrease the risk of diabetic complications. Gender-based behaviour change and attitudes towards diabetes are common observations. Women tend to diet more [8, 9, 10], use the healthcare system more often [11] and more often report that diabetes has a negative impact on their lives [12], but they exercise less than men [9, 10, 13]. Despite these known gender differences in diabetes-related behaviour and diabetes outcomes there is only limited evidence on the impact of gender on the effectiveness of diabetes interventions. However, it has been suggested that the health service should be concerned with inequalities between men and women in the management of patients at risk of coronary heart disease [14].

In the pragmatic, open, multicentre, randomised controlled trial Diabetes Care in General Practice (DCGP) ( registration no. NCT01074762), the intervention of structured personal care reduced the incidence of any diabetes-related outcome and myocardial infarction (MI) in patients newly diagnosed with type 2 diabetes [15]. However, the observed effect of structured personal care on reducing HbA1c measured 6 years after diagnosis was present only in women [13]. With evidence of gender differences in disease outcomes, behaviours and attitudes in people with type 2 diabetes, it is relevant to investigate the impact of gender on intervention outcomes in diabetes trials. Hence, at the end of 6 years of structured diabetes intervention, we assessed gender differences in mortality and cardiovascular and microvascular complications over a follow-up period of 13 years, using registry data.


Study population

The DCGP study was a pragmatic, open, controlled trial in which general practices were randomised to provide structured personal care or routine care [16]. In the DCGP study, 474 Danish general practitioners (GPs) treated 1,381 patients with diabetes newly diagnosed based on specific diagnostic criteria (Fig. 1) [16]. Among these patients, 1,369 (99.1%) were of western European descent. Based on onset of insulin treatment, approximately 97.5% of the patients were considered to have type 2 diabetes. The present study population comprises the 970 patients who survived and were re-examined at the end of 6 years of intervention. Of these, 478 were women, and 492 were men. The study was approved by the Research Ethics Committee of Copenhagen and Frederiksberg (V.100.869/87) and oral consent was given by all patients.
Fig. 1

Patient flow through trial. y, year


In Denmark, routine care for type 2 diabetes is usually provided in the primary care setting and costs are covered by a tax-based health insurance system. In the intervention group, follow-up every 3 months and annual screening for diabetes complications were supported by a questionnaire sent to GPs 1 month before the next expected consultation. GPs were asked to work with patients to define the best possible goals for controlling important risk factors, with an emphasis placed on glycaemic control [16]. At each quarterly consultation, GPs were asked to evaluate patients’ achievements in light of the goals set.

GPs were introduced to possible solutions to therapeutic problems through six annual half-day seminars, annual descriptive feedback reports on individual patients and folders and leaflets for doctors and for patients. Generally, the importance of diet was stressed and doctors were recommended to postpone, if possible, the start of glucose-lowering drugs until at least 3 months after diabetes diagnosis to observe the effect of any weight loss. GPs were encouraged to recommend increased physical exercise and simple dietary rules [16]. In cases of persistent hyperglycaemia, hypertension and/or dyslipidaemia, pharmacological treatment was recommended. No individual patient-specific advice on treatment was given to GPs, who were allowed to deviate from the recommendations in an effort to individualise treatment. None of the intervention procedures were explicitly based on the sex of the patient. Patients were never approached by the study centre.

GPs in the routine care group were free to choose any treatment and to change it over time [16]. The practices providing routine care were not contacted by the study coordinating centre during the trial after patient inclusion had stopped. In September 1995, the intervention was terminated and the 6 year examination was initiated. No further attempt was made to maintain patients in the randomised groups or to influence their therapy.

Clinical follow-up

The clinical 6 year follow-up examination was completed for 970 (93.4%) of 1,039 surviving patients after a median (interquartile range) of 5.57 (4.96–6.16) years in the structured personal care group and 5.85 (5.30–6.45) years in the routine care group. At this follow-up examination, GPs recorded body weight, blood pressure, glucose-lowering medication, number of consultations the preceding year and whether a patient had ever been treated at a diabetes clinic. In questionnaires, patients gave information about whether they lived alone, smoking habits and leisure-time physical activity. Analysis of fasting blood samples and freshly voided morning urine samples was centralised. The fraction of HbA1c was determined using the same ion-exchange HPLC method throughout the study. The reference interval was 5.4–7.4% (36–57 mmol/mol). This method was later compared with a newer Diabetes Control and Complications Trial (DCCT)-aligned HPLC method. The reference range may cautiously be converted to 4.8–6.7% (29–50 mmol/mol) if the DCGP analytical method had been DCCT-aligned [16]. A description of other variables and definitions has previously been published [16]. Hypertension was defined as systolic/diastolic blood pressure ≥160/90 mmHg and/or the use of antihypertensive and/or diuretic drugs. Microalbuminuria was defined as urinary albumin concentration ≥15 to <200 mg/l and proteinuria as ≥200 mg/l.

At 6 year follow-up, GPs answered questions about patient motivation for best possible control and treatment, patient attitudes to study participation, influence of patient’s own efforts on the course of diabetes treatment, a realistic goal for fasting whole-blood glucose and the GP’s opinion on whether knowledge that the patient was participating in a study influenced consultations. Patients answered five questions concerning altered habits, food habits, home glucose monitoring, attitudes towards diabetes and social support. All questions were based on a literature review and interviews with people who have type 2 diabetes. Experienced GPs and sociologists reviewed the questions, which were revised after pilot testing.

Registry-based follow-up

After the end of intervention, patients were followed up for 13 years using Danish registries. The vital and emigration status of all patients were certified through the Danish Civil Registration System [17], in which everyone living in Denmark is registered with a permanent and unique personal identification number allowing linkage between study populations and all national registries. Surviving patients were censored on 31 December 2008. The Danish Register of Causes of Death contains information about underlying and possible contributory causes of death [18]. In four cases the cause of death was not recorded in this registry. The Danish National Patient Register includes information on almost all contacts with hospitals in Denmark [19] (e.g. discharge diagnosis[es] and surgical procedures performed). The seven outcomes used in the registry-based follow-up were made with reference to those in the UK Prospective Diabetes Study (UKPDS) [20]: any diabetes-related endpoint, diabetes-related deaths, all-cause mortality, MI, stroke, peripheral vascular disease and microvascular disease.

Statistical analysis

The incidences of death and other outcomes were analysed univariably with logrank tests and multivariably in Cox regression models. In the latter, 95% CIs and p values were determined using a sandwich estimator for the variance to account for clustering of patients within practices [19]. Patients with missing values for one or more variables were omitted from the analyses where these variables were included. Absolute risks for each outcome were calculated as the number of patients experiencing the corresponding outcome divided by the sum of the risk times (i.e. from the start of the 6 year follow-up examination to the first occurrence of the outcome, death or end of follow-up). Patients with any occurrence of an outcome preceding the 6 year follow-up were excluded from the analyses pertaining to that outcome. Two multivariate models are presented: one model adjusting for age, diabetes duration and clustering and one model with additional adjustment for BMI, hypertension, HbA1c, total cholesterol, sedentary physical activity, current smoking and receipt of glucose-lowering medication. Whether the effect of randomisation differed between gender groups was tested by the interaction of patient sex and randomisation group in a joint model for men and women. Analyses were done in SAS v9.2 (SAS Institute, Cary, NC, USA). The level of statistical significance was p < 0.05.


After 6 years of intervention no gender-specific differences in the effect of the intervention on intermediate outcomes was seen, except for HbA1c (Table 1). In the intervention and control group, respectively, HbA1c concentration was 8.6% (70 mmol/mol) and 9.4% (79 mmol/mol) in women and 8.8% (73 mmol/mol) and 9.0% (75 mmol/mol) in men (interaction p = 0.003). There was no difference between men and women in the effect of intervention on referral to a diabetes clinic, although since diabetes diagnosis, fewer women had been referred in the structured personal care group than in the routine care group (17.3% vs 31.3%, p = 0.003).
Table 1

Characteristics of patients at end of intervention


Men (n = 492)

Women (n = 478)

Interaction p valueb

n (routine/structured)

Routine care

Structured personal care

p valuea

n (routine/structured)

Routine care

Structured personal care

p valuea


  Age, years


67.1 ± 10.8

66.7 ± 10.5



70.0 ± 10.7

70.1 ± 11.1



  Diabetes duration, years


5.9 ± 0.7

5.5 ± 0.9



6.0 ± 0.9

5.6 ± 0.9



  Live alone


44 (20.7)

64 (26.0)



96 (50.8)

116 (44.8)




  BMI, kg/m2


28.3 ± 4.1

28.8 ± 4.5



29.4 ± 5.8

29.0 ± 5.5





155 (70.5)

182 (66.9)



157 (78.1)

215 (77.6)




  HbA1c, %c


9.0 ± 1.6

8.8 ± 1.7



9.4 ± 1.9

8.6 ± 1.3



  HbA1c, mmol/mol








  Total cholesterol, mmol/l


6.0 ± 1.2

5.8 ± 1.5



6.5 ± 1.2

6.3 ± 1.2



  Fasting triacylglycerol, mmol/l


2.3 ± 1.8

2.3 ± 3.0



2.3 ± 1.2

2.0 ± 1.3



  Serum creatinine, μmol/l


102 ± 28

105 ± 67



90 ± 29

89 ± 25



  Urinary albumin








    Normal: <15 mg/l


110 (53.4)

139 (53.9)


119 (63.3)

179 (70.2)


    Microalbuminuria: ≥15 to <200 mg/l


83 (40.3)

101 (39.2)


60 (31.9)

72 (28.2)


    Proteinuria: ≥200 mg/l


13 (6.3)

18 (7.0)


9 (4.8)

4 (1.6)



  Sedentary (leisure-time) physical activityd


54 (25.5)

57 (23.5)



72 (38.5)

85 (33.2)



  Current smokerd


73 (34.9)

104 (42.3)



43 (22.9)

58 (22.6)



Process of care



6.7 ± 4.6

8.0 ± 7.7



8.0 ± 6.4

8.4 ± 5.0



  Diabetes-related consultations/yeard


4.3 ± 3.4

5.0 ± 3.7



4.5 ± 3.6

5.2 ± 3.1



  Ever treated at a diabetes clinicd


48 (21.8)

44 (16.2)



63 (31.3)

48 (17.3)



  Glucose-lowering therapyd








    Diet only


73 (33.2)

81 (29.9)


58 (28.9)

78 (28.2)


    Oral glucose-lowering medicine


122 (55.5)

161 (59.4)


109 (54.2)

165 (59.6)




25 (11.4)

29 (10.7)


34 (16.9)

34 (12.3)


Values are means (SD) or n (% of randomisation group)

The p values are from multivariate generalised linear models (ordinary linear regression for continuous variables, logistic regression for binary variables and Poisson regression for count variables with log[diabetes duration] as offset) where the effect of structured care vs routine care is adjusted for age and diabetes duration. Clustering with GPs is accounted for by the use of generalised estimating equations

aTests effect of randomisation within gender groups

bTests whether the effect of randomisation differs between gender groups

cReference range: 5.4–7.4%

dData from questionnaires to patients (behavioural) or their GPs (process of care)

The intervention did not have a statistically significant effect on patients’ attitudes. While men in the intervention group tended to feel that they had less social support than the men in the control group, the possible effect of the intervention went in the opposite direction for women (Table 2). This tendency is in line with the observation that, when considering the influence of a patient’s own efforts on treatment course, the intervention GPs considered that the men’s efforts were at a lower level than that of the control GPs’ male patients. Again, the intervention had an effect in the opposite direction for women (interaction p = 0.011, Table 2). In both randomisation arms, women were considered to comply better with dietary advice than men.
Table 2

Attitudes and opinions of patients and GPs at end of intervention


Men (n = 492)

Women (n = 478)

Interaction p valueb

n (routine/structured)

Routine care

Structured personal care

p valuea

n (routine/structured)

Routine care

Structured personal care

p valuea

Information from patient questionnaires

  Altered habits after diagnosis










121 (57.9)

155 (64.3)


123 (65.8)

156 (60.5)


    Yes, but very little


42 (20.1)

43 (17.8)


34 (18.2)

55 (21.3)




46 (22.0)

43 (17.8)


30 (16.0)

47 (18.2)


  Food habits








    Diabetes diet


48 (22.9)

72 (29.9)


76 (40.4)

99 (38.4)


    Full diet without sugar


118 (56.2)

127 (52.7)


91 (48.4)

128 (49.6)


    Diet as for non-diabetic individuals


44 (20.9)

42 (17.4)


21 (11.2)

31 (12.0)


  Performs home blood or urinary glucose monitoring


58 (28.3)

79 (32.5)



56 (29.8)

61 (23.8)



  Attitudes towards diabetes








    The illness is unproblematic


101 (48.8)

115 (47.9)


95 (51.4)

152 (59.4)


    Work/worked with the illness


88 (42.5)

103 (42.9)


66 (35.7)

79 (30.9)


    It is a strain


18 (8.7)

22 (9.2)


24 (13.0)

25 (9.8)


  Social support








    Full support


161 (78.9)

173 (72.4)


94 (51.4)

142 (56.6)


    Handle it by oneself


30 (14.7)

45 (18.8)


57 (31.2)

71 (28.3)


    Feels alone/misunderstood


13 (6.4)

21 (8.8)


32 (17.5)

38 (15.1)


For the patient in question, the GP’s opinion

  Patient’s motivation for best possible control and treatment over past year








    Very good


65 (29.6)

58 (21.4)


53 (26.6)

56 (20.3)




69 (31.4)

92 (34.0)


60 (30.2)

106 (38.4)




52 (23.6)

74 (27.3)


49 (24.6)

73 (26.5)




34 (15.4)

47 (17.4)


37 (18.6)

41 (14.9)


  Patient’s attitude to study participation








    Happy with the attention


48 (22.2)

138 (51.9)


43 (22.2)

136 (50.0)


    No special importance


159 (73.6)

102 (38.4)


138 (71.1)

115 (42.3)


    Irritated or bothered


9 (4.2)

26 (9.8)


13 (6.7)

21 (7.7)


  The influence of patient’s own efforts on treatment course










144 (65.8)

148 (54.6)


117 (59.1)

171 (62.0)


    None in particular


36 (16.4)

51 (18.8)


29 (14.7)

49 (17.8)




39 (17.8)

72 (26.6)


52 (26.3)

56 (20.3)


  Realistic goal for fasting whole-blood glucose








    ≤7 mmol/l


58 (27.2)

92 (34.1)


51 (26.2)

98 (35.5)


     > 7 to 8 mmol/l


51 (23.9)

65 (24.1)


37 (19.0)

74 (26.8)


    >8 to 9 mmol/l


29 (13.6)

48 (17.8)


32 (16.4)

35 (12.7)


    >9 mmol/l


75 (35.2)

65 (24.1)


75 (38.4)

69 (25.0)


  Use of fact that patient was participating in study during consultations








    Used vigorously


2 (0.9)

41 (15.3)


4 (2.0)

36 (13.0)


    Used moderately


17 (7.8)

123 (45.9)


17 (8.7)

120 (43.3)


    Only mentioned when necessary


199 (91.3)

104 (38.8)


175 (89.3)

121 (43.7)


Values are given as n (% of randomisation group)

The p values are multivariate (multinomial) logistic regression models where the effect of structured care vs routine care is adjusted for age and diabetes duration. Clustering with GP is accounted for by the use of generalised estimating equations

aTests effect of randomisation within gender group

bTests whether the effect of randomisation differs between genders

During 13 years of follow-up, no statistically significant reductions in outcomes were observed for men when comparing the structured personal intervention group with the routine care group (Table 3). In women, however, a lower HR (95% CI) and absolute risk for personal structured vs routine care emerged for any diabetes-related endpoint (0.65 [0.48, 0.87], p = 0.004, adjusted for age, diabetes duration, clustering, physical activity, smoking and clinical variables; 73.4 vs 107.7 events per 1,000 patient-years), diabetes-related death (0.70 [0.50, 0.96], p = 0.031; 34.6 vs 45.7), all-cause mortality (0.74 [0.57, 0.97], p = 0.028; 55.5 vs 68.5) and stroke (0.59 [0.36, 0.97], p = 0.038; 15.6 vs 28.9). This effect differed between men and women for diabetes-related death (interaction p = 0.015, Table 3) and all-cause mortality (interaction p = 0.005). Hence, survival for women who received structured care improved whereas there was a tendency towards a poorer survival for men following structured care (Fig. 2).
Table 3

Outcomes from registry-based monitoring for 13 years after intervention was terminated


No. of patients without outcome at end of intervention (routine/structured)

No. (%) of patients with outcome

Absolute risk (events per 1,000 patient-years)

HR (95% CI)a for structured care vs routine care

Routine care

Structured personal care

p valueb

Routine care

Structured personal care

p valuec

Adjusted for age, diabetes duration and clustering

p valued

Interaction p valuee

Additionally adjusted for physical activity, smoking and clinical variablesf

p valued

Interaction p valuee


Any diabetes-related endpoint

  6–19 years men


87 (63.0)

104 (56.8)





0.89 (0.65, 1.21)



0.90 (0.67, 1.22)




  6–19 years women


104 (72.2)

117 (60.0)





0.65 (0.49, 0.87)



0.65 (0.48, 0.87)




Diabetes-related deaths

  6–19 years men


84 (38.4)

110 (40.6)





1.20 (0.90, 1.60)



1.09 (0.81, 1.47)




  6–19 years women


82 (41.0)

91 (33.0)





0.72 (0.54, 0.96)



0.70 (0.50, 0.96)




All-cause mortality

  6–19 years men


136 (61.8)

179 (65.8)





1.21 (0.96, 1.52)



1.11 (0.88, 1.40)




  6–19 years women


123 (61.2)

146 (52.7)





0.78 (0.62, 0.99)



0.74 (0.57, 0.97)





  6–19 years men


64 (33.2)

70 (29.1)





0.95 (0.68, 1.32)



1.00 (0.69, 1.44)




  6–19 years women


58 (31.7)

61 (23.7)





0.72 (0.49, 1.05)



0.71 (0.45, 1.12)





  6–19 years men


43 (21.7)

49 (20.0)





1.04 (0.67, 1.61)



0.78 (0.48, 1.28)




  6–19 years women


46 (24.2)

37 (14.6)





0.51 (0.33, 0.78)



0.59 (0.36, 0.97)




Peripheral vascular disease

  6–19 years men


12 (5.6)

17 (6.3)





1.12 (0.49, 2.54)



1.10 (0.49, 2.46)




  6–19 years women


9 (4.6)

6 (2.2)





0.37 (0.13, 1.06)



0.67 (0.20, 2.27)




Microvascular disease

  6–19 years men


28 (13.3)

31 (11.9)





0.95 (0.57, 1.60)



0.98 (0.57, 1.67)




  6–19 years women


29 (14.8)

31 (11.5)





0.74 (0.42, 1.30)



0.90 (0.47, 1.70)




aHR is calculated in a Cox proportional hazard regression model. The corresponding 95% CIs and p values are determined using a sandwich estimator for the variance to account for clustering of patients within practices

b p value from a Rao–Scott χ 2 test: a Pearson χ 2 test adjusted for clustering of patients with GPs

c p value from a logrank test

dTests the effect of randomisation within gender groups

eTests whether the effect of randomisation differs between gender groups

fBesides age, diabetes duration and clustering, this model also adjusted for BMI, hypertension, HbA1c, total cholesterol, sedentary physical activity, current smoking and receipt of glucose-lowering medication

Fig. 2

Survival after end of intervention according to randomisation arm and gender. Solid blue line, structured care, women; solid red line, structured care, men; dotted blue line, routine care, women; dotted red line, routine care, men


During 13 years of follow-up after the completion of 6 years of structured personal diabetes care, women experienced lower all-cause mortality and lower incidences of diabetes-related death, any diabetes-related endpoint, and stroke compared with women in the control group. Such effects were not seen in men. The gender difference was statistically significant for all-cause mortality and diabetes-related death.

Gender perspective

The structured personal care intervention provided focused treatment strategies for lowering blood glucose, blood pressure and cholesterol but the intervention did not take a patient’s gender into consideration in any way. The intervention, however, lowered HbA1c in women but not men (Table 1). The HbA1c level has been shown to have a graded positive association with risk of stroke in women [21] and mortality increases with HbA1c in type 2 diabetes [22]. The lowering of HbA1c could therefore contribute to explaining the positive outcome for women. However, the difference in mortality outcome persisted after adjustment for HbA1c. Turnbull et al, in a meta-analysis, found that treatment allocation to more intensive glucose control reduced the risk of major cardiovascular events but not all-cause or cardiovascular mortality [23]. Intensive multifactorial therapy in high-risk patients with type 2 diabetes and well-established microalbuminuria, however, has previously been shown to reduce death from any cause and cardiovascular death [24].

Gender differences in diabetes outcomes are well documented. However, we are not aware of other studies assessing the impact of gender on endpoints in structured diabetes interventions. A large meta-analysis found women with diabetes to be at more than 40% higher risk of incident coronary heart disease than men with diabetes [25]. Moreover, a relatively higher increase in mortality [3, 26], fatal CVD [4, 27] and stroke [3, 28] has been found among women diagnosed with diabetes compared with men. However, one meta-analysis found that the excessive relative risk of CVD in women with type 2 diabetes was absent when adjusting for classical CVD risk factors [29]. With our gender-based results showing improved morbidity and mortality outcomes for women receiving structured personal care, but without any obvious explanation from improved intermediate outcomes (except for HbA1c, for which we adjust), we need to discuss how gender really matters in diabetes and diabetes care.

In medical research, gender issues are usually presented as cross-sectional measures of difference, indicating a dichotomous and essentialist understanding of men and women. However, to explain the impact of gender on our study outcome, a more complex conceptual and theoretical framework is needed [30]. In this discussion, we shall regard gender in the context of social and cultural interaction (‘doing gender’) beyond an inborn property [30]. Taksdal and Widerberg have presented a framework assessing gender as ‘biology’, ‘identity’, ‘symbol’ and ‘structure’ [31]. We shall apply this framework to discuss potential hypotheses for the impact of gender, reflected by the difference in the intervention effect on all-cause mortality between men and women in our study.

In medicine, gender in terms of ‘biology’ (usually reported as ‘sex’) has traditionally been regarded as most relevant. Generally, women lose their female cardiovascular protection when suffering from type 2 diabetes [26]. This has been explained by epigenetic changes prompting more unfavourable presentation of oestrogen receptors associated with higher oxidative stress, pro-inflammatory profile and increased atherosclerotic plaque formation [32].

Gender in terms of ‘identity’ is linked with how people think, feel and behave when they incorporate masculinity and femininity and perform masculine or feminine roles. Women disclose their diabetes status and integrate management more readily into their lives, whereas men are more reluctant to talk about their diabetes and are less observant of self-management practices [8]. Women find it more stressful to accommodate their own needs and health concerns into daily life since they often see themselves as caretakers and givers rather than receivers [33]. We previously reported a more adaptive attitude towards treatment among women, and this could lead to better treatment adherence and disease outcomes [13]. Women report poorer social support compared with men [34], possibly linked to a poorer self-perceived quality of life [35]. This, together with health status, is related to increased mortality [36]. Hence, the structured diabetes intervention might have provided disease-related support and attention, which improved disease behaviour and self-perceived quality of life, leading to positive long-term outcomes for women.

Gender in terms of ‘symbol’ pertains to cultural images of masculinity and femininity—which is necessary to be considered a real man or a real woman [37]. Negotiating work and healthcare have been identified as barriers to disease self-management and acceptance of disease [38]. Men have been found to be less worried about long-term outcomes and to be more troubled by limitations to their personal freedom following diabetes diagnosis [34]. Men expect less benefit from self-management [34] and rely more on self-directed learning [8]. This may contribute to the poorer outcome among men in the structured diabetes intervention.

Gender in terms of ‘structure’ deals with work, economy, power and privileges. In most societies men are better educated, have higher positions in society, are more financially independent and take greater control of decision making than women. Several studies have shown that men with diabetes and CVD are more likely than women with comparable conditions to receive more intensive medical treatments such as statins, antihypertensive drugs and acetylsalicylates [26, 39], which would be expected to lead to better treatment outcomes. As we provided a focused, structured and personalised intervention for both men and women, quality of care could be assumed to be similar in both groups. Therefore, a possible treatment bias might have been levelled out, adding to improved outcomes among women.

Structured personal diabetes care could provide women with significant attention and support and thus provide an incentive to treatment adherence. Women accept disease and implement disease management more easily [13], which might affect long-term outcomes. Masculinity may be challenged by diabetes, demanding daily consideration and lifestyle changes [34]. The structured approach could conflict with men’s tendency to trust self-directed learning instead of self-management.

Strengths and weaknesses of the study

This is a post hoc analysis of a randomised controlled trial and the results should be interpreted as observational. The detailed information on possible confounders, however, allowed for extensive adjustment of HRs.

The outcomes of this study were drawn from the Danish national registries. The Danish Register of Causes of Death covers the entire population of Denmark [18] and the Danish National Patient Registry has covered discharges from Danish hospitals since 1977 [19]. From 1995 onwards this registry has also covered outpatients, but contacts with the few and small private specialised hospitals were not included in the registry until 2007. The private hospitals may be considered relatively unimportant in the present analyses, as hardly any of the outcomes of interest are treated there.

Vital status was confirmed for all our study participants. The cause-specific mortality, in our study diabetes-related deaths, relies on the validity of the diagnoses in the national registries. These methodological considerations, however, are not relevant for the outcome of all-cause mortality. The validity has not been established for all the non-fatal outcomes. In one study, the predictive value of MI as primary diagnosis or underlying cause of death was 93.6% and the sensitivity was 77.6% in comparison with definite or possible MI [40]. For a stroke diagnosis the predictive value was 81–86% in the Danish National Patient Register when evaluated in an audit of patient records [41].

In the nationwide DCGP study, time-dependent changes in definitions of diseases and in registration and coding practices are unlikely to cause differential misclassification according to treatment allocation. This assumption of non-differential misclassification is supported by the fact that the diagnoses in the registries are almost entirely provided by GPs unaware of patients’ randomisation status.

There are several arguments to support the generalisability of the present results to the wider population of patients with type 2 diabetes: the study sample was population-based; patients were included with no upper age limit; the setting was general practice where most patients with type 2 diabetes are treated; the elements of the intervention resemble standard procedures in general practice and a relatively high number of general practices participated. Due to our application of individualised treatment goals it is, however, uncertain whether patients subjected to treatment-to-target will show the same gender difference.

Clinical implications

We present a post hoc observational analysis of a randomised trial comparing structured personal diabetes care with routine diabetes care. Of seven predefined outcomes, the intervention reduced all-cause mortality, diabetes-related death, any diabetes-related outcome and stroke in women, but not in men, and this gender difference was statistically significant for all-cause mortality and diabetes-related death. After 6 years of intervention, HbA1c was only lowered in women, but the improvements in outcomes for women were preserved after adjustment for HbA1c. Hence, we propose that the improved outcomes in women may be explained by complex social and cultural issues of gender. There is a need to further explore the gender-specific effects of major intervention trials in order to rethink the way we provide medical care to both men and women, so that both men and women benefit from intensified treatment efforts.



We thank the patients, GPs and ophthalmologists who volunteered to take part in this study.


Funding was provided by the Danish Medical Research Council, The Danish Research Foundation for General Practice, the Health Insurance Foundation, the Danish Ministry of Health, Novo Nordisk Farmaka Denmark Ltd, the Pharmacy Foundation, the Foundation for General Practice in Copenhagen, Frederiksberg, Tårnby and Dragør, the Doctor Sofus Carl Emil Friis and his Wife Olga Doris Friis Trust, the Danish Medical Association Research Fund, the Velux Foundation, the Rockwool Foundation, Novo Nordisk Ltd, the Danish Diabetes Association, Oda and Hans Svenningsen’s Fund, the A. P. Møller Foundation for the Advancement of Medical Science, the Novo Nordisk Foundation, Captain Axel Viggo Mørch and his Wife’s Trust, the Danish Eye Health Society, Mogens and Jenny Vissing’s Trust and Bernhard and Marie Klein’s Trust. None of the funding sources were involved in any of the following: study design; collection, analysis and interpretation of data; writing of the report; the decision to submit the article for publication. The researchers are independent from funders.

Duality of interest

The authors declare that there is no duality of interest associated with this manuscript.

Contribution statement

NdFO developed the research question and wrote the protocol for this follow-up study together with the other authors. NdFO was responsible for the original study design, randomisation, intervention delivery and data collection and obtained funding. VS performed the statistical analyses. All authors made substantial contributions to the analysis and interpretation of data. KM contributed with gender theory. The paper was written by MØK, LH, ABSN and NdFO and the other authors revised it critically for important intellectual content. All authors, external and internal, had full access to all of the data (including statistical reports and tables) in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. All authors have approved the final version of the manuscript. NdFO is guarantor.


  1. 1.
    World Health Organization (2014) World Health Statistics 2014 [report online]. Available from Accessed 14 Dec 2014
  2. 2.
    Hu G, Jousilahti P, Qiao Q, Katoh S, Tuomilehto J (2005) Sex differences in cardiovascular and total mortality among diabetic and non-diabetic individuals with or without history of myocardial infarction. Diabetologia 48:856–861PubMedCrossRefGoogle Scholar
  3. 3.
    Mulnier HE, Seaman HE, Raleigh VS, Soedamah-Muthu SS, Colhoun HM, Lawrenson RA (2006) Mortality in people with type 2 diabetes in the UK. Diabet Med 23:516–521PubMedCrossRefGoogle Scholar
  4. 4.
    Huxley R, Barzi F, Woodward M (2006) Excess risk of fatal coronary heart disease associated with diabetes in men and women: meta-analysis of 37 prospective cohort studies. BMJ 332:73–78PubMedPubMedCentralCrossRefGoogle Scholar
  5. 5.
    Ray KK, Seshasai SR, Wijesuriya S et al (2009) Effect of intensive control of glucose on cardiovascular outcomes and death in patients with diabetes mellitus: a meta-analysis of randomised controlled trials. Lancet 373:1765–1772PubMedCrossRefGoogle Scholar
  6. 6.
    UK Prospective Diabetes Study Group (1998) Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPDS 38. BMJ 317:703–713PubMedCentralCrossRefGoogle Scholar
  7. 7.
    Cholesterol Treatment Trialists (CTT) Collaborators, Kearney PM, Blackwell L et al (2008) Efficacy of cholesterol-lowering therapy in 18,686 people with diabetes in 14 randomised trials of statins: a meta-analysis. Lancet 371:117–125CrossRefGoogle Scholar
  8. 8.
    Mathew R, Gucciardi E, De Melo M, Barata P (2012) Self-management experiences among men and women with type 2 diabetes mellitus: a qualitative analysis. BMC Fam Pract 13:122PubMedPubMedCentralCrossRefGoogle Scholar
  9. 9.
    Nothwehr F, Stump T (2000) Health-promoting behaviors among adults with type 2 diabetes: findings from the Health and Retirement Study. Prev Med 30:407–414PubMedCrossRefGoogle Scholar
  10. 10.
    Chiu CJ, Wray LA (2011) Gender differences in functional limitations in adults living with type 2 diabetes: biobehavioral and psychosocial mediators. Ann Behav Med 41:71–82PubMedCrossRefGoogle Scholar
  11. 11.
    Jonsson PM, Sterky G, Gafvels C, Ostman J (2000) Gender equity in health care: the case of Swedish diabetes care. Health Care Women Int 21:413–431PubMedCrossRefGoogle Scholar
  12. 12.
    Glasgow RE, Ruggiero L, Eakin EG, Dryfoos J, Chobanian L (1997) Quality of life and associated characteristics in a large national sample of adults with diabetes. Diabetes Care 20:562–567PubMedCrossRefGoogle Scholar
  13. 13.
    Nielsen AB, de Fine Olivarius N, Gannik D, Hindsberger C, Hollnagel H (2006) Structured personal diabetes care in primary health care affects only women’s HbA1c. Diabetes Care 29:963–969PubMedCrossRefGoogle Scholar
  14. 14.
    Frich JC, Malterud K, Fugelli P (2006) Women at risk of coronary heart disease experience barriers to diagnosis and treatment: a qualitative interview study. Scand J Prim Health Care 24:38–43PubMedCrossRefGoogle Scholar
  15. 15.
    Hansen LJ, Siersma V, Beck-Nielsen H, de Fine Olivarius N (2013) Structured personal care of type 2 diabetes: a 19 year follow-up of the study Diabetes Care in General Practice (DCGP). Diabetologia 56:1243–1253PubMedCrossRefGoogle Scholar
  16. 16.
    Olivarius NF, Beck-Nielsen H, Andreasen AH, Horder M, Pedersen PA (2001) Randomised controlled trial of structured personal care of type 2 diabetes mellitus. BMJ 323:970–975PubMedPubMedCentralCrossRefGoogle Scholar
  17. 17.
    Pedersen CB (2011) The Danish civil registration system. Scand J Public Health 39:22–25PubMedCrossRefGoogle Scholar
  18. 18.
    Helweg-Larsen K (2011) The Danish register of causes of death. Scand J Public Health 39:26–29PubMedCrossRefGoogle Scholar
  19. 19.
    Lynge E, Sandegaard JL, Rebolj M (2011) The Danish National Patient Register. Scand J Public Health 39:30–33PubMedCrossRefGoogle Scholar
  20. 20.
    Holman RR, Paul SK, Bethel MA, Matthews DR, Neil HA (2008) 10-year follow-up of intensive glucose control in type 2 diabetes. N Engl J Med 359:1577–1589PubMedCrossRefGoogle Scholar
  21. 21.
    Zhao W, Katzmarzyk PT, Horswell R, Wang Y, Johnson J, Hu G (2014) Sex differences in the risk of stroke and HbA1c among diabetic patients. Diabetologia 57:918–926PubMedPubMedCentralCrossRefGoogle Scholar
  22. 22.
    Ma WY, Li HY, Pei D et al (2012) Variability in hemoglobin A1c predicts all-cause mortality in patients with type 2 diabetes. J Diabetes Complicat 26:296–300PubMedCrossRefGoogle Scholar
  23. 23.
    Control Group, Turnbull FM, Abraira C et al (2009) Intensive glucose control and macrovascular outcomes in type 2 diabetes. Diabetologia 52:2288–2298CrossRefGoogle Scholar
  24. 24.
    Gaede P, Lund-Andersen H, Parving HH, Pedersen O (2008) Effect of a multifactorial intervention on mortality in type 2 diabetes. N Engl J Med 358:580–591PubMedCrossRefGoogle Scholar
  25. 25.
    Peters SA, Huxley RR, Woodward M (2014) Diabetes as risk factor for incident coronary heart disease in women compared with men: a systematic review and meta-analysis of 64 cohorts including 858,507 individuals and 28,203 coronary events. Diabetologia 57:1542–1551PubMedCrossRefGoogle Scholar
  26. 26.
    Baviera M, Santalucia P, Cortesi L et al (2014) Sex differences in cardiovascular outcomes, pharmacological treatments and indicators of care in patients with newly diagnosed diabetes: analyses on administrative database. Eur J Intern Med 25:270–275PubMedCrossRefGoogle Scholar
  27. 27.
    Lee C, Joseph L, Colosimo A, Dasgupta K (2012) Mortality in diabetes compared with previous cardiovascular disease: a gender-specific meta-analysis. Diabetes Metab 38:420–427PubMedCrossRefGoogle Scholar
  28. 28.
    Peters SA, Huxley RR, Woodward M (2014) Diabetes as a risk factor for stroke in women compared with men: a systematic review and meta-analysis of 64 cohorts, including 775,385 individuals and 12,539 strokes. Lancet 383:1973–1980PubMedCrossRefGoogle Scholar
  29. 29.
    Kanaya AM, Grady D, Barrett-Connor E (2002) Explaining the sex difference in coronary heart disease mortality among patients with type 2 diabetes mellitus: a meta-analysis. Arch Intern Med 162:1737–1745PubMedCrossRefGoogle Scholar
  30. 30.
    Holge-Hazelton B, Malterud K (2009) Gender in medicine—does it matter? Scand J Public Health 37:139–145PubMedCrossRefGoogle Scholar
  31. 31.
    Taksdal A, Widerberg K (1992) Forståelser av kjønn. Gyldendal, Oslo [in Norwegian]Google Scholar
  32. 32.
    Dantas AP, Fortes ZB, de Carvalho MH (2012) Vascular disease in diabetic women: why do they miss the female protection? Exp Diabetes Res 2012:570598PubMedPubMedCentralCrossRefGoogle Scholar
  33. 33.
    Hepworth J (1999) Gender and the capacity of women with NIDDM to implement medical advice. Scand J Public Health 27:260–266PubMedCrossRefGoogle Scholar
  34. 34.
    Gafvels C, Lithner F, Borjeson B (1993) Living with diabetes: relationship to gender, duration and complications. A survey in northern Sweden. Diabet Med 10:768–773PubMedCrossRefGoogle Scholar
  35. 35.
    Schunk M, Reitmeir P, Schipf S et al (2012) Health-related quality of life in subjects with and without type 2 diabetes: pooled analysis of five population-based surveys in Germany. Diabet Med 29:646–653PubMedCrossRefGoogle Scholar
  36. 36.
    Haring R, Feng YS, Moock J et al (2011) Self-perceived quality of life predicts mortality risk better than a multi-biomarker panel, but the combination of both does best. BMC Med Res Methodol 11:103PubMedPubMedCentralCrossRefGoogle Scholar
  37. 37.
    Courtenay WH (2000) Constructions of masculinity and their influence on men's well-being: a theory of gender and health. Soc Sci Med 50:1385–1401PubMedCrossRefGoogle Scholar
  38. 38.
    Fort MP, Alvarado-Molina N, Pena L, Mendoza Montano C, Murrillo S, Martinez H (2013) Barriers and facilitating factors for disease self-management: a qualitative analysis of perceptions of patients receiving care for type 2 diabetes and/or hypertension in San Jose, Costa Rica and Tuxtla Gutierrez, Mexico. BMC Fam Pract 14:131PubMedPubMedCentralCrossRefGoogle Scholar
  39. 39.
    Tonstad S, Rosvold EO, Furu K, Skurtveit S (2004) Undertreatment and overtreatment with statins: the Oslo Health Study 2000-2001. J Intern Med 255:494–502PubMedCrossRefGoogle Scholar
  40. 40.
    Madsen M, Davidsen M, Rasmussen S, Abildstrom SZ, Osler M (2003) The validity of the diagnosis of acute myocardial infarction in routine statistics: a comparison of mortality and hospital discharge data with the Danish MONICA registry. J Clin Epidemiol 56:124–130PubMedCrossRefGoogle Scholar
  41. 41.
    Krarup LH, Boysen G, Janjua H, Prescott E, Truelsen T (2007) Validity of stroke diagnoses in a National Register of Patients. Neuroepidemiology 28:150–154PubMedCrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2015

Authors and Affiliations

  • Marlene Ø. Krag
    • 1
  • Lotte Hasselbalch
    • 1
  • Volkert Siersma
    • 1
  • Anni B. S. Nielsen
    • 1
  • Susanne Reventlow
    • 1
  • Kirsti Malterud
    • 1
    • 2
    • 3
  • Niels de Fine Olivarius
    • 1
  1. 1.The Research Unit for General Practice and Section of General Practice, Department of Public HealthUniversity of CopenhagenCopenhagen KDenmark
  2. 2.Research Unit for General Practice, Uni Health ResearchBergenNorway
  3. 3.Department of Global Public Health and Primary CareUniversity of BergenBergenNorway

Personalised recommendations