Time trends in healthy lifestyle among adults in Germany: Results from three national health interview and examination surveys between 1990 and 2011.

BACKGROUND: The combined impact of multiple healthy behaviors on health exceeds that of single behaviors. This study aimed to estimate trends in the prevalence of a healthy lifestyle among adults in Germany.
METHODS: A data set of 18,058 adults aged 25-69 years from three population-based national health examination surveys 1990-92, 1997-99 and 2008-11 with complete information for five healthy behavior factors was used. A 'daily intake of both fruits and vegetables, 'sufficient physical exercise', 'no current smoking' and 'no current risk drinking' were assessed with self-reports and 'normal body weight' was calculated based on measured body weight and height. A dichotomous 'healthy lifestyle' indicator was defined as meeting at least four out of five healthy behaviors. Age-standardized prevalence was calculated stratified by sex, age groups (25-34, 35-44, 45-54 and 55-69 years) and education level (low, medium and high). Trends were expressed in relative change (RC) between 1990-92 and 2008-11.
RESULTS: In Germany, the overall prevalence of healthy lifestyle increased from 9.3% in 1990-92 to 13.5% in 1997-99 and to 14.7% in 2008-11 (RC: +58.1%). The prevalence increased among men and women and in all age groups, with the exception of men aged 45-54 years. The RC of increasing healthy lifestyle prevalence between 1990-92 and 2008-11 was stronger albeit on a higher level among women compared to men. Therefore, the gender difference in healthy lifestyle has increased, but age-related differences have overall decreased in this period. Among high educated men the prevalence of a healthy lifestyle increased between 1990-92 and 2008-11 from 10.6% to 16.3% (p = 0.01) and among high educated women from 16.4% to 30.3% and also among medium educated women (10.9 to 16.6, p<0.01), but no significant increase in healthy lifestyle prevalence was observed among men with low and medium education and among women with low education level.
CONCLUSIONS: The prevalence of a lifestyle with at least four out of five healthy behaviors markedly increased from 1990-92 to 2008-11. Nevertheless, additional health promotion interventions are needed to improve the number of combined healthy behavior factors and the awareness in the population that each additional healthy behavior factor leads to a further improvement in health, especially in men in the age-range 45 to 54 years, and among persons with low education level.

Introduction

A healthy lifestyle is a combination of several healthy behaviors that reduce the likelihood of ill-health and premature death [1]. These usually include non-smoking, sufficient physical activity, healthy eating and low-risk alcohol drinking. Furthermore, normal body weight often is considered to be a healthy behavior related factor. Different approaches have been applied to combine several healthy behavior factors to composite indices of healthy lifestyle [2-5]. Studies using such measures demonstrated a clear dose-response relationship: the higher the number of combined healthy behavior factors, the larger are the health benefits [6-13]. A recent cohort study using NHANES data with a 35 years follow-up period observed a risk reduction of 74% for all-cause mortality among persons with five healthy behaviors compared to those without healthy behavior factors [14]. In addition, an unhealthy lifestyle with a lower number of healthy behavior factors has been linked to a higher risk of developing type 2 diabetes and cardiovascular diseases and to premature death [15-17]. Another recent study based on the European Social Survey 2014 observed that only 5.8% of adults in Europe were adhering to five healthy behaviors with strong differences between the 20 countries under study [18]. Hungary with 1.3% and the Czech Republic with 1.9% were the countries with the lowest prevalence of five healthy behaviors and the United Kingdom with 8.6% and Finland with 9.2% were those with the highest prevalence [18]. In Germany in 2009–10, 7.1% of women and 3.2% of men were adhering to five healthy behaviors and 29.1% of women and 17.8% of men showed at least four out of five healthy behaviors [19]. Studies estimating the potential of reducing social inequalities in mortality in Europe by simulating scenarios in which educational groups have the same healthy lifestyle pattern as the high-educated group, showed that the potential is very high for countries with a marked social gradient of healthy behaviors [20-22]. Thus, information on multiple healthy behavior factors combined is crucial for public health monitoring and policy planning. However, only a few studies have analyzed trends over time in healthy lifestyle in the general adult population and these observed only little net changes in the prevalence of healthy lifestyles [23].

For Germany, time trends of prevalence of a healthy lifestyle in the adult population have not been investigated so far. Therefore, the aim of this study is to estimate trends in the prevalence of a healthy lifestyle among adults in Germany according to gender, age and level of education.

Materials and methods

Study design, setting and participants

We used data of the German Federal Health Monitoring System from three cross-sectional German national health interview and examination surveys (GNHIESs) with information about health status, risk factors and health behavior for the general adult population [24]. The surveys were conducted from 1990 to 1992 (GNHIES 1990–92), 1997 to 1999 (GNHIES 1997–99) and 2008 to 2011 (GNHIES 2008–11) [25-27]. Detailed information on the study designs and methods for the three surveys were published previously [25-28]. In brief, all three surveys were based on two-stage cluster random sampling designs. One hundred twenty to 180 sample points (clusters) were randomly selected for each survey proportional to the population structure of the Federal Republic of Germany. In a second step, address information were randomly drawn from local population registries in the sampled communities. The eligible study populations were adults living in Germany aged 18 to 79 years for GNHIES 1997–99 and GNHIES 2008–11 and 25- to 69-year-old adults with German nationality for GNHIES 1990–92. Hence, we restricted our trend analysis to the age range 25 to 69 years [28]. The response rates were 70% for GNHIES 1990–92, 61% for GNHIES 1997–99 and 42% for GNHIES 2008–11 [28]. A non-responder analysis for GNHIES 2008–11 indicated a high representativeness between the study sample and the German population structure [29]. All participants were informed about the study objectives, examination and interview processes and applicable data protection guidelines. Verbal consent was witnessed and formally recorded in the surveys 1990–92 and 1997–1999. The GNHIES 1997–99 and GNHIES 2008–11 participants signed an informed written consent prior to participation. All surveys were conducted according to the Federal and State Commissioners for Data Protection guidelines and the GNHIES 2008–11 study protocol was approved by the Charité –University Medicine Berlin ethics committee; No. EA2/047/08 [27]. The 1997–99 and 2008–11 13 surveys conform to the principles of the Helsinki Declaration.

Measurements and variables

In line with previous studies [13, 14, 19], we selected five healthy behavior factors to construct a healthy lifestyle index (HLI)–sufficient physical activity, low-risk alcohol drinking, a healthy diet (fruit and vegetable intake), non-smoking, and normal body mass index (BMI). Appropriate data with comparable assessment methods across the three surveys were identified for each healthy behavior factor. Physical exercise and smoking data were assessed with self-administered questionnaires as described elsewhere [28]. Briefly, physical exercise was assessed with the question ‘How often do you engage in physical exercise?’ and the regular duration in hours per week was assessed with categorical answer options. Smoking habits were assessed with the questions allowing a distinction between ‘current’, ‘former’ and ‘never’ smoking [6, 9]. In all surveys, alcohol as well as fruit and vegetable consumption were assessed with self-administered food-frequency questionnaires; however, there are differences in the data collection across the surveys in food groups, reference time, intake frequencies and additional information on portion sizes. To compare information on alcohol consumption across surveys we used frequencies and quantities of alcoholic beverage intake. A detailed description of the alcohol assessment is given in the supplement (S1 File).

A daily consumption of both fruits and vegetables was used as an indicator for a healthy dietary pattern. To compare the information on fruit and vegetable consumption across the three surveys we used only frequency information since portion sizes were not obtained in GNHIES 1990–92 and GNHIES 1997–99. Information on frequency of consumption of fruits and vegetables was assessed in GNHIES 1990–92 with the question ‘How often do you consume these particular foods?’ without a specific reference period. Participants reported their consumption of ‘cooked vegetables’, ‘tinned vegetables’, ‘salad’, ‘raw vegetables’, and ‘fresh fruits’ using the frequency categories ‘(almost) daily’, ‘several times a week’, ‘about once a week’, ‘2 to 3 times a month’, ‘once a month or less’, and ‘never’. In GNHIES 1997–99, data on frequency of intake during the past 12 months was collected for ‘cooked vegetables’, ‘tinned vegetables’, ‘lettuce, raw salad, raw vegetables’, and ‘fresh fruits’ using the frequency categories ‘several times a day’, ‘daily or almost daily’, ‘several times a week’, ‘about once a week’, ‘2 to 3 times a month’, ‘once a month or less’, and ‘never’. In GNHIES 2008–11, the frequency of intake during the last 4 weeks was assessed for ‘ raw vegetables’, ‘cooked vegetables’, ‘legumes’, ‘fresh fruits’ and ‘cooked fruits’. The answering categories were ‘more than 5 times a day’, ‘4 to 5 times a day’, ‘3 times a day’, ‘twice a day’, ‘once a day’, ‘5 to 6 times a week’, ‘3 to 4 times a week’, ‘1 to 2 times a week’, ‘ 2 to 3 times a month, ‘once a month or less’, and ‘never’. The dietary assessment method in GNHIES 2008–11 has been described in detail previously [30]. To examine trends over time from surveys with different dietary collection methods, we standardized the frequency of fruit and vegetable consumption by recoding the information into a dichotomous variable reflecting a daily intake of both fruits and vegetables or not.

In all surveys, body weight and body height were measured during physical examination in a standardized manner by trained personal. BMI was calculated as the ratio of a person’s body weight to the square of body height (kg/m2).

Outcome variables

Smoking status was categorized into ‘current smoking’ versus ‘no current smoking’. This cut-off was selected because it was used in healthy lifestyle scores of previous studies which demonstrated significant associations between healthy lifestyle and health outcomes and mortality [6, 9]. ‘Sufficient physical exercise’ was defined as reporting regular physical exercise of at least two hours per week. This cut-off point was selected because from the available answer options it comes closest to the minimum level of health-enhancing physical activity of 2.5 hours per week recommended by the World Health Organization (WHO) [31]. Moreover, in a previous study we demonstrated that this level of physical exercise is associated with an improved cognitive function across the life span [32]. ‘No current risk drinking’ was defined as consuming ≤ 20 grams of pure alcohol per day in men and ≤ 10 grams per day in women in line with evidence-based drinking guidelines for German adults [33, 34]. ‘Daily fruits and vegetables’ intake was defined as once or more per day versus less than daily intake of fruits and vegetables. ‘Normal weight’ was defined according to WHO guidelines as having a BMI in the range between 18.5 to less than 25 kg/m2 [35]. A ‘Healthy lifestyle index’ (HLI) was constructed by assigning one point for each healthy behavior factor resulting in a score from zero to five; where zero is interpreted as high-risk and five as low-risk lifestyle. Furthermore, a dichotomous variable with a cut-off of ‘at least four’ healthy behavior factors (HLI ≥ 4) was used and hereafter referred to as ‘healthy lifestyle’. The selection of this cut-off was based on a meta-analysis which indicated that a combination of at least four healthy behavior factors is associated with a reduction of all-cause mortality by 66% [13].

Stratification variables

We stratified the analyses by gender and used the following age strata: 25 to 34, 35 to 44, 45 to 54 and 55 to 69 years. ‘Level of education’ was defined according to the International Standard Classification of Education (ISCED) 2011 and EUROSTAT guidelines [36]. Based on self-reported information on the highest education level, education of participants was classified as low (ISCED levels 0–2), medium (ISCED levels 3–4) or high education (ISCED levels 5–8) [36].

Study size

The total numbers of participants in the age group 25 to 69 years were 7,466 for GNHIES 1990–92 (3,641 men and 3,825 women), 5,825 for GNHIES 1997–99 (2,831 men and 2,994 women) and 5,375 for GNHIES 2008–11 (2,538 men and 2,837 women). After exclusion of individuals with missing data for at least one healthy behavior factor used for the HLI, the final study sample consisted of 7,382 participants for GNHIES 1990–92, 5,603 for GNHIES 1997–99 and 5,073 for GNHIES 2008–11 (Table 1). Those numbers relate to item response rates for the HLI of 98.9% for the GNHIES 1990–92, 96.2% for the GNHIES 1997–99 and 94.4% for the GNHIES 2008–11. The total study sample comprised 18,058 participants, 8,716 men and 9,342 women.

Table 1

Distribution of the study samples according to sex and age.

	GNHIES 1990–92	GNHIES 1997–99	GNHIES 2008–11	Total
	1990–92	1997–99	2008–11
	n (%)	n (%)	n (%)	n
Total sample	7466	5825	5375	18666
Study sample	7382	5603	5073	18058
Men	3594 (48.7)	2742 (48.9)	2380 (46.9)	8716
Age
25–34	922 (25.7)	614 (22.4)	378 (15.9)	1914
35–44	839 (23.3)	698 (25.5)	455 (19.3)	1992
45–54	905 (25.2)	578 (21.1)	634 (26.6)	2117
55–69	928 (25.8)	852 (31.1)	913 (38.4)	2693
Women	3788 (51.3)	2861 (51.1)	2693 (53.1)	9342
Age
25–34	1022 (27.0)	646 (22.6)	410 (15.2)	2078
35–44	856 (22.6)	720 (25.2)	536 (19.9)	2112
45–54	883 (23.3)	587 (20.5)	755 (28.0)	2225
55–69	1027 (27.1)	908 (31.7)	992 (36.8)	2927

Statistical methods

We used the software SAS Version 9.4 (SAS Institute, Cary, NC, USA) for the statistical analyses. Within the analyses, a weighting procedure was applied to all three surveys to adjust for deviations from the German population structure at each survey period according to age, sex, region and education level. In addition, the results were age-standardized to the German population structure as of 31 December 2010 to control for demographic changes and differences in age distribution across the three survey samples. Trend estimations were performed, while adjusting for survey design effects of the cluster sampling designs of each survey. Age-standardized and weighted prevalence and 95% confidence intervals (CI) were calculated for each healthy behavior factor separately and for the HLI based on three cross-sectional survey samples collected at three different time periods (1990–92, 1997–99 and 2008–11). Trends were expressed with relative change (RC), i.e. (value in survey 3 minus value in survey 1) / (value in survey 1) × 100%. The basis for comparison is always the estimate from the first survey (GNHIES 1990–92). Logistic regression was used to test time trends for statistical significance with the survey wave as a categorical variable in the model. The criterion for statistical significance was set at p<0.05 [28].

Results

Healthy behavior factors

Time trends for the five healthy behavior factors are presented for men and women separately in Fig 1. Among men and women, the prevalence of sufficient physical exercise and no current risk drinking have increased between 1990–92 and 2008–11 (RC: sufficient physical exercise +44.9% in men; +118.2% in women); no current risk drinking +64.8% in men and +53.1% in women; all p<0.01), while daily fruits and vegetables intake has decreased (RC: -38.6% in men and -29.2% in women; both p<0.01). More pronounced differences were observed for no current risk drinking with higher increases in prevalences between 1990–92 and 1997–99 compared to 1997–99 and 2008–11. The decrease in daily fruits and vegetables intake was more distinct between 1997–99 and 2008–11 in comparison to 1990–92 and 1997–99. The prevalence of no current smoking has increased among men (+7.2%, p = 0.04) but decreased among women (-4.3%, p = 0.10). Between 1990–92 and 2008–11 no statistical relevant changes can be observed among men and women for the prevalence of normal weight. In all age groups sufficient physical exercise and no current risk drinking increased as well as daily fruits and vegetables intake decreased over time (S1 and S2 Tables). No current smoking increased over time among men in the age groups 35–44 years (RC: +17.9%, p<0.01) and 55–69 years (RC: +13.7%, p<0.01) but decreased in 45–69 year old women (45–54 years RC: -12.7%, 55–69 years RC: -8.2%, both p<0.01). A higher prevalence of normal weight in 2008–11 compared to 1990–92 was only observed among older women (45–54 years RC: +14.5%, p = 0.08; 55–69 years RC: +26.3%, p<0.01). The increasing trend in sufficient physical exercise and no current risk drinking and the decreasing trend in daily fruits and vegetables intake over time can be observed in all educational groups albeit on a different level (S1 Fig). No significant changes in the prevalence of normal weight over time between different educated groups can be observed.

Fig 1

Prevalence (%, 95%-CI) of individual healthy behavior factors among men (A) and women (B) aged 25–69 years.

Healthy lifestyle index (HLI)

In total, the proportion with a high number of combined healthy behavior factors has increased between 1990–92 and 2008–11 (Fig 2). The proportions of men with none or only one healthy behavior factor have declined (relative change: -62.9% and -19.2%, both p < 0.0001), while the proportions of those with three (RC: +25.8%, p = 0.10), four (RC: +42.4%, p = 0.009) and five factors (RC: +66.7%, p < 0.0001) have increased. Among women, the proportion of those with four and five healthy behavior factors has strongly increased (RC: +44.7% and +428.6, both p<0.0001), while the proportions of those with less than 4 factors have declined. However, this result is only statistically significant among women with one (RC: -19.3%, p = 0.0004) or two (RC: -8.5%, p = 0.02) healthy behavior factors. In general, among all age groups, the proportions of men and women with none and one healthy behavior factors have declined between 1990–92 and 2008–11, while the proportions of adults with four and five factors have increased (Fig 3). However, not all trends are statistically significant.

Fig 2

Proportions of specific numbers of healthy behavior factors among men and women aged 25–69 years.

Fig 3

Proportions of specific numbers of healthy behavior factors among men (A) and women (B) by age groups.

Healthy lifestyle (HLI ≥ 4)

Table 2 presents the proportions of each health behavior factor by HLI ≥ 4. The proportion of sufficient physical exercise, daily fruits and vegetables, no current smoking, and normal weight as proportion of HLI ≥ 4 increased over time. No current risk drinking did not show any changes over time.

Table 2

Proportion (%, 95%-CI) of each health behaviour factor by healthy lifestyle index (HLI) ≥ 4.

	HLI ≥ 4
	1990–92	1997–99	2008–11
Sufficient physical exercise	33.9 (30.9–37.1)	45.0 (41.6–48.4)	42.7 (39.1–46.3)
No current risk drinking	16.5 (14.8–18.3)	17.2 (15.7–18.9)	18.3 (16.7–20.1)
Daily fruits and vegetables	19.2 (17.4–21.2)	28.0 (25.6–30.6)	38.6 (35.6–41.7)
No current smoking	13.3 (11.9–14.9)	19.4 (17.7–21.2)	21.0 (19.1–22.9)
Normal weight	21.6 (19.6–23.7)	30.3 (27.7–32.9)	31.5 (28.6–34.5)

The proportions of adults with a healthy lifestyle across the surveys are presented in Table 3. HLI ≥ 4 prevalence increased in total from 9.3% in 1990–92 to 13.5% in 1997–99 and to 14.7% in 2008–11 (RC 1990-92-2008-11: +58.1%, p<0.01). The corresponding percentages were for men 7.5%, 9.3% and 10.9% (RC: +45.3%, p<0.01) and for women 11.1%, 17.7% and 18.6% (RC: +67.6%, p<0.01). The prevalence increased in all age groups, with the exception of men aged 45–54 years. The relative change of increasing healthy lifestyle prevalence between 1990–92 and 2008–11 was stronger albeit on a higher level among women compared to men. Therefore, gender difference in healthy lifestyle has increased, but age-related differences have overall decreased in this period.

Table 3

Proportions (%) of adhering to a healthy lifestyle among men and women according to sex, age group and education.

	1990–92	1997–99	2008–11	Relative change	p#
				1990–92 to 2008–11
	% (95% CI)	% (95% CI)	% (95% CI)	%
Total	9.3 (8.3–10.3)	13.5 (12.3–14.7)	14.7 (13.4–16.2)	58.1	< .0001
Age
25–34	12.6 (11.0–14.4)	14.9 (12.8–17.2)	16.3 (13.3–19.7)	29.4	0.08
35–44	8.1 (6.6–9.8)	13.5 (11.8–15.5)	15.6 (13.1–18.6)	92.6	< .0001
45–54	9.5 (8.0–11.3)	13.8 (11.6–16.4)	12.9 (10.9–15.2)	35.8	0.005
55–69	7.8 (6.4–9.5)	12.2 (10.4–14.3)	14.7 (12.3–17.4)	88.5	< .0001
Education
Low	7.3 (5.9–9.1)	8.3 (6.5–10.4)	9.0 (6.4–12.7)	23.3	0.57
Medium	8.8 (7.8–10.0)	12.8 (11.6–14.2)	12.6 (11.0–14.3)	43.2	< .0001
High	12.5 (10.8–14.5)	19.7 (17.3–22.3)	22.2 (19.8–24.8)	77.6	< .0001
Men	7.5 (6.4–8.8)	9.3 (8.1–10.8)	10.9 (9.4–12.5)	45.3	0.003
Age
25–34	11.6 (9.5–14.1)	10.3 (8.0–13.2)	13.6 (10.0–18.2)	17.2	0.40
35–44	5.7 (4.1–8.0)	8.0 (5.9–10.7)	12.4 (9.1–16.8)	117.5	0.004
45–54	7.9 (6.0–10.4)	9.1 (6.7–12.2)	7.1 (5.0–9.9)	-10.1	0.61
55–69	5.7 (4.1–7.8)	10.1 (7.8–12.8)	11.3 (8.7–14.5)	98.2	0.003
Education
Low	3.4 (1.9–6.0)	4.0 (2.0–7.8)	7.6 (3.9–14.3)	123.5	0.19
Medium	6.9 (5.8–8.2)	8.2 (6.7–9.9)	8.5 (6.9–10.5)	23.2	0.29
High	10.6 (8.4–13.2)	13.9 (11.5–16.8)	16.3 (13.5–19.4)	53.8	0.01
Women	11.1 (9.9–12.4)	17.7 (15.9–19.7)	18.6 (16.6–20.8)	67.6	< .0001
Age
25–34	13.6 (11.3–16.3)	19.6 (16.5–23.3)	19.0 (14.8–24.0)	39.7	0.01
35–44	10.5 (8.4–13.1)	19.4 (16.4–22.8)	18.9 (15.4–23.1)	80.0	< .0001
45–54	11.2 (9.0–13.7)	18.7 (15.2–22.7)	18.9 (15.6–22.6)	68.8	0.0003
55–69	9.8 (7.8–12.2)	14.3 (11.7–17.4)	18.0 (14.6–21.9)	83.7	0.0005
Education
Low	9.0 (7.2–11.2)	10.5 (8.2–13.5)	10.0 (6.7–14.6)	11.1	0.65
Medium	10.9 (9.2–12.8)	17.6 (15.6–19.9)	16.6 (14.1–19.3)	52.3	< .0001
High	16.4 (13.5–19.9)	30.4 (25.9–35.2)	30.3 (26.3–34.6)	84.8	< .0001

# p for trend

1 Healthy lifestyle index ≥ 4

Some sex differences in the trends for age and education were observed. While in women the proportion with a healthy lifestyle increased between 1990–92 and 2008–11 in all age groups, among men increases were only observed in the age groups 35 to 44 years and 55 to 69 years. The proportion of women with a healthy lifestyle increased between 1990–92 and 2008–11 only in the high and medium education groups, while among men an increase was only observed in the high education group.

Discussion

In this trend analysis, based on three population-based German national health interview and examination surveys, it is observed that the prevalence of a healthy lifestyle defined as having at least four out of five healthy behavior factors, overall increased explicitly among adults in the period between 1990–92 and 1997–99 and further slightly between 1997–99 and 2008–11. Age-related differences that were observed in 1990–92 with younger persons having a healthy lifestyle more often than older persons attenuated over time until 2008–11, among men and women. Gender-related differences that were observed in 1990–92 with women having more often a healthy lifestyle than men became slightly larger over time. Educational differences that were observed in 1990–92 showing that those with high education had a healthy lifestyle more often than those with low education became larger among women but not among men.

Health implications

Cohort studies consistently demonstrated that the relative risk for all-cause and cardiovascular mortality proportionally decreased with a higher number of combined healthy behavior factors [6, 12, 13]. Persons with a combination of at least four healthy behavior factors, as chosen in this trend study, can expect a reduction of all-cause mortality by 66% according to a meta-analysis including 15 cohort studies with more than 513,000 participants [13]. Hence, since the prevalence of a combination of at least four healthy behavior factors has increased between 1990–92 and 2008–11 in Germany, it can be expected that morbidity and mortality related with unhealthy behavior will subsequently decrease. In line with this assumption, the cardiovascular disease mortality has declined in Germany in the last decades and mortality rates are expected to further decline in Germany until 2025 [37-39]. This is on the one hand due to a better health care with improved detection and treatment of cardiovascular and metabolic diseases and on the other hand due to an improved health-related behavior [28]. Moreover, a previous trend study based on the same surveys indicated that mean systolic blood pressure, total cholesterol and serum glucose levels significantly declined in the period between 1990–92 and 2008–10 among men and women [28]. Those reductions in cardio-metabolic risk profiles and deaths may be partially explained by the observed increase in combined healthy behavior factors among adults in Germany. However, it is likely that an improved health care also contributed to the described developments with higher detection rates of undiagnosed hypertension, hyperlipidemia and diabetes and higher prescription and use of antihypertensive, lipid-lowering and antidiabetic drugs [28, 40–42].

Trend patterns

To the best of our knowledge, trend studies on combined healthy behavior factors in adults using population-based, country-wide data are rare. A study from the USA observed only modest improvements in healthy lifestyle in adults in the period between 1994 and 2007 using data of the Behavioral Risk Factor Surveillance System [23]. Comparing single healthy behavior changes, non-smoking prevalence increased in the USA, while there was little change in fruit and vegetable consumption and physical activity, and prevalence of healthy weight even decreased [23]. The main drivers for increasing healthy lifestyle prevalence in our study were increases in sufficient physical exercise and no current risk drinking. In contrast daily fruits and vegetables intake decreased over time, which is also observed in the German National Nutrition Monitoring [43]. Yet, the largest part of observed improvements in single healthy behavior factors as well as in the HLI ≥ 4 occurred between 1990–92 and 1997–99 In Germany in the 1990s there was a discussion on a stronger legal alcohol limit when driving, which resulted in a revised law in 2001 [44]. Furthermore, the national legislation has developed the banning of alcohol use at work within the labor protection law and employer agreements at the workplace [45].

Gender effects

The prevalence of a healthy lifestyle was significantly higher in women compared to men in all three observation periods. The relative change of increasing healthy lifestyle prevalence between 1990–92 and 2008–11 was stronger albeit on a higher level among women compared to men, leading to an increasing gender differences in healthy lifestyle over time. The strongest increases among all single healthy behavior factors were observed to be sufficient physical exercise in women and no current risk drinking in men. We are not aware of any other trend study to compare results on developments of gender-related differences in healthy lifestyle over time. However, Walther et al. also observed that female sex was a predictor for a higher healthy lifestyle score in a cohort of Swiss adults [46]. Several other studies using healthy lifestyle indices from the USA, Asia and Europe consistently showed that women are more likely to have a healthier lifestyle compared to men [6, 47–49].

Age effects

In the observation period 1990–92, younger persons had overall higher healthy lifestyle prevalence than older persons, with a more pronounced pattern in men. Over time, healthy lifestyle prevalence increased in most age groups with the exception of the age groups 25–34 and 45–55 years among men and a less pronounced pattern in the age group 25–34 years among women leading to a less pronounced age difference in healthy lifestyle in 2008–11 compared to 1990–92. Among women, age differences in healthy lifestyle disappeared, with a prevalence of 18–19% in all age groups in 2008–11. Among men, the age group 45–55 years has the lowest healthy lifestyle prevalence at 7% and may therefore be a target group for health promotion interventions. The study sample consists of different generations including birth cohorts ranking from 1921 to 1985. People aged 55–69 years in 1990–92 belong to a birth cohort born between 1921/23–1935/37 and are referred to by social scientist as the ‘Silent Generation’, while people aged 55–69 years in 2008–10 belong to a birth cohort born between 1939/41–1953/55 of which the most belong to the ‘Baby Boomers’ generation [50]. It is likely that sociocultural differences between the generations exist which influences their lifestyles. The strong increase in sufficient physical exercise in the oldest age group 55–69 years is an important contributor to the positive trend in healthy lifestyle in this age group, especially among women. In the Baby Boomer generation in the course of the second wave feminist movement of the late 1960s and early 1970s [51] women adopted lifestyles which were formerly more common among men such as performing sports and exercise [52-54]. This may partly explain the higher prevalence of physical exercise in the Baby Boomer generation compared to the Silence Generation. In line with this observation, a trend study from Spain observed the same pattern that older people showed initially lower leisure-time physical activity levels than younger people and that this difference became smaller over time especially among women [55].

Education effects

The prevalence of a healthy lifestyle was significantly higher among men and women with high compared to low education level in all three observation periods, this observation is in line with the previous finding from the ‘German Health Update’ study 2009 (GEDA 2009) [56]. An important driver for the overall increasing healthy lifestyle prevalence between 1990–92 and 2008–11 could be the educational expansion that took place in Germany in the last decades [57]. Studies simulated scenarios in which the low-educated adapt the health behaviors of the high-educated showing that societies can expect significant improvements in health equality [20-22]. This could partly have happened in Germany between 1990–92 and 2008–11 when the societal group of people with low education became smaller and the group of people with high education became larger over time. The educational expansion was driven by an disproportionate increase of females achieving high education certificates [58], leading to an improvement of gender equality in education. This is in line with the observation that healthy lifestyle prevalence has more strongly increased in females than in males. However, the educational expansion cannot explain differences in healthy lifestyle between educational groups. Among women, those with medium and high education showed stronger increases in healthy lifestyle over time than those with low education. This was different among men, where those with low education showed stronger increases in healthy lifestyle over time than those with medium and high education. However, only among men with high education the increase in healthy lifestyle prevalence was statistically significant. The result of increasing educational inequalities in healthy lifestyle among women is in line with previous trend studies showing increasing educational inequalities over time in smoking and physical exercise behaviors among adults in Germany [59, 60].

Limitations

Several limitations should be considered when interpreting our findings. All healthy behavior factors, except normal body weight, were based on self-reported information and thus, we cannot exclude the possibility that reporting bias occurred, e.g. because of limited cognitive ability to report complex behaviors or socially desirable answering [61, 62]. In addition, the surveys were conducted over a period of more than 20 years bringing about inevitably changes in survey and assessment methods making comparisons over time more difficult. We have paid much attention to these aspects when selecting variables and defining indicators, but we cannot fully exclude the possibility that methodological differences between the surveys have influenced the results [28, 63]. The initial physical exercise question has not changed across the three surveys, but the answer categories have been slightly adapted. The initial smoking question was the same in the first two surveys but was adapted in the third survey. The BMI assessment method based on physical examination data on body weight and height has remained consistent across the three surveys [28]. A consumption of both fruits and vegetables each day is a very rough indicator for a general healthy diet which may include much more aspects e.g. intake of highly processed foods, saturated fat, fiber, sugar intake. However, comprehensive information on the diet was not available in all three surveys. Consumption was assessed with self-administered food-frequency questionnaires with inconsistent reference periods and slightly different food items and answering categories for the three surveys. Although we tried to standardize this by condensing this information, this may have resulted in systematic differences in estimates for the surveys. Furthermore, we cannot fully rule out the possibility that selection bias has compromised generalizability of the results. As in many other European countries, the response rates in the German national health surveys have declined over time [64, 65]. The selection bias can be occurred at different levels of recruitment, e.g. selection of sample points, selection of individuals, or participation of individuals. However, we used weighting factors with a common methodology for all three surveys to increase the representativeness of the results [28].

Conclusion

We conclude that prevalence of a lifestyle with at least four out of five healthy behaviors markedly increased between 1990–92 and 2008–11. Another positive development from a public health perspective is that age-related differences in healthy lifestyle have overall decreased during this period. On the negative side, the gender difference in healthy lifestyle prevalence has increased. Moreover, in men, increases in healthy lifestyle prevalence were limited to men aged 35 to 44 and 55 to 69 years and to those with high education, whereas in women, increases were observed in all age groups but limited to those with high or medium education. The awareness of a healthy lifestyle and the fact that each additional healthy behavior factor leads to a further improvement in health should be increased in the population. Additional health promotion interventions are needed to improve the number of combined healthy behavior factors in the general adult population. Special effort should be undertaken to reach men, especially in the age-range 45 to 54 years, and persons with low education level. According to the WHO, multi-component interventions should preferably be used to promote ‘health in all policies’ following a setting-specific approach and focusing on the upstream drivers of healthy behaviors in order to create more healthy environments, systems, societies and people [1, 66, 67].

Proportions (%, 95%-CI) of healthy behaviour factors according to educational level.

(TIFF)

Click here for additional data file.

Proportions of individual healthy behavior factors among men aged 25–69 years according to age.

(PDF)

Click here for additional data file.

Proportions of individual healthy behavior factors among women aged 25–69 years according to age.

(PDF)

Click here for additional data file.

Assessment of alcohol consumption within the German Federal Health Monitoring System.

(PDF)

Click here for additional data file.

5 Jul 2019

PONE-D-19-16785

Time trends in healthy lifestyle among adults in Germany: Results from three national health interview and examination surveys between 1990 and 2011

PLOS ONE

Dear Dr Finger,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by Aug 19 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

David Meyre

Academic Editor

PLOS ONE

Journal Requirements:

1.

When submitting your revision, we need you to address these additional requirements.

Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2.

We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For more information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially sensitive information, data are owned by a third-party organization, etc.) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: I Don't Know

Reviewer #3: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: Yes

Reviewer #3: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Review MS number PONE-D-19-16785

Time trends in healthy lifestyle among adults in Germany: Results from three national

health interview and examination surveys between 1990 and 2011

Abstract

The abstract is well written and clear.

Methods: Please add e.g. sufficient/ xx amount to ‘Daily fruits and vegetables intake’.

Results: ‘Nevertheless, the gender difference in healthy lifestyle has increased’; Please clarify to which direction. Are men healthier compared to women?

Last sentence; start the sentence with referring to education, so it is directly clear that results for differences in education will be presented.

Conclusions: four out of five vs. 4 or 5 mentioned in the method section.

Ethics statement

Did the 1990-92 participants signed an informed consent?

Introduction

Line 73-77: Can authors also relate these numbers to a higher prevalence of health related diseases or premature death?

Materials and Methods

Line 92-94: Make clear that you compare prevalence data measured at 3 different time points in 3 different cohort. This is not always clear.

Line 104: Mentioning the response rate is not in line with informing that response rates for the three surveys were published previously (line 97).

Line 108 (as mentioned at the ethics statement): Why did participants from the 1997-99 wave did not sign an informed consent?

Line 119: ‘Smoking habits were assessed with the questions allowing a distinction between ‘current smokers’ and ‘others’’. This is a strict cut-off value. Please elaborate why this was so strict. Do authors have any information about former smoking?

Line 114 vs 121: clarify that a healthy diet is defined as fruit and vegetables intake.

Line 123-147: The explanation of alcohol consumption is too wordy which makes it difficult to follow. Recommend to rewrite/shorten or put to supplement.

Line 150: Assessment of fruit and vegetables in 1990-92 is not mentioned to which time period is referred to.

Line 148-168: Restructure, so all information is mentioned and provided in the same order.

Line 170 is repetitive to line 119-120.

The outcome section is well written.

Line 186: add that stratification is also done by gender.

Line 195-197: ‘After exclusion of individuals with missing data for at least one healthy behavior factor used for the HLI, the final study sample consisted of 7,382 participants for GNHIES 1990-92, 5,603 for GNHIES 1997-99 and 5,073 for GNHIES 2008-11’ were there any differences between participants with missing data and the included sample?

Line 207: ‘For age it was standardized to the German population structure as of 31 December 2010.’ To me this is unclear; since you will compare the prevalence of health behavior of people with a certain age (which is fixed since within each cohort you do not check changes over time) in 1990, 1997 and 2008. To what extend does age need to be standardized? Do you mean the results are weighted? And why standardized to 31 December 2010, though data has also been collected in 2011. Please clarify what you mean with this sentence.

Results

Authors present the data in the figure stratified by men and women, but why not by age groups or education level. Please elaborate or potentially add visual stratification in the supplement.

Were there any differences between 1990-92 and 1997-99 or between 1997-99 and 2008-11?

Line 235: was the proportion of men with high number of healthy behavior also significantly different.

Discussion

Line 271-274 (and line 376): add that trend also include data from 1997-99. Now the impression is that there were only two data collection points.

Line 287-289: “In line with this assumption, the cardiovascular disease mortality has declined in Germany in the last decades and mortality rates are expected to further decline in Germany until 2025 (35-37).” Do authors expect that this is due to a healthy lifestyle or better health care?

Line 306-307: can authors give a potential explanation for this observation? Different legislation?

Limitations

The response rate for GNHIES 2008-11 was much lower (42%) compared to the other two surveys (70% 1990-92 and 61% 1997-99). Please provide some explanation for the discrepancies.

A healthy diet is considered as daily fruit and vegetables intake. What about fibers, and no consumption of sweets and savory products, or soda?

Fruit and vegetable consumption was assessed by means of a self-administered food-frequency questionnaires. Intake was measured for the past 12 months for 1997-99, while it was measured for the past 4 weeks in 2008-11. Acknowledge the inconsistency and provide any expect differences?

Reviewer #2: This is an interesting study which highlight the importance of lifestyle change during the current epidemic of NCD in World.

Reviewer #3: Thank you for the opportunity to review your article titled, "Time trends in healthy lifestyle among adults in Germany: Results from three national health interview and examination surveys between 1990 and 2011".

Overall, the purpose of this study is interesting, and be well organized paper. However, several issues were still concerned.

Major points:

1) Selection bias: As the authors mentioned in line 370, selection bias (generalizability) would be more important in the present study. The authors must add values of response rates (not proportion of missing of HLI) not only for whole sample in each time point, but also response rates according to sex and age groups if possible. Additionally, the authors should consider generalizability by comparison with complete survey such as German census and the present dataset in each time point (e.g. comparison whether lower education level prevalence are same between German census and the present dataset).

2) Line 367: It was unclear what did the authors concerned as “bringing about inevitably changes in survey and assessment methods”. Please describe the detail which lifestyle assessment was not differed by period.

3) To understand what items mainly contributed to HLI ≥ 4, the authors should add results of cross table about proportion of each items of HLI by binary variable of HLI ≥ 4.

Minor points:

1) Tables: “Relative change 1990-2011” is inappropriate, because it would be 1990-92 vs. 2008-11.

2) Table 3: Tile “Proportions (%) of four or five healthy lifestyle indicators” is inappropriate. For example, “Proportions of adhering to healthy lifestyle combination” may be more appropriate.

3) Results of decrease trend of fruits and vegetables intake in Germany was definitely reported by the previous study (https://www.ncbi.nlm.nih.gov/pubmed/26934826). Reader would concern why fruits and vegetables intake was decreasing among German people. Have the authors checked the decrease trend according to education level?

4) Figures: Font sizes should be bigger.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Ilse Reinders

Reviewer #2: No

Reviewer #3: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

Submitted filename: Review MS Time trends in healthy lifestyle.docx

Click here for additional data file.

21 Aug 2019

Review of MS number PONE-D-19-16785 Time trends in healthy lifestyle among adults in Germany: Results from three national health interview and examination surveys between 1990 and 2011

Response to the reviewer

Reviewer #1:

Abstract

1) The abstract is well written and clear.

Methods: Please add e.g. sufficient/ xx amount to ‘Daily fruits and vegetables intake’.

Answer: For the comparison in this analysis, we could only use the frequency information on fruit and vegetable consumption, so daily amounts were not calculated. The indicator reflects therefore only a daily consumption of both fruits and vegetables. We reformulated this to “A ‘daily intake of both fruits and vegetables’…”.

2) Results: ‘Nevertheless, the gender difference in healthy lifestyle has increased’; Please clarify to which direction. Are men healthier compared to women?

Answer: Thank you for this comment. This sentence is unclear. We revised the sentence in the abstract as well as in the Results to: ‘The RC of increasing healthy lifestyle prevalence between 1990-92 and 2008-11 was stronger albeit on a higher level among women compared to men. Therefore, the gender difference in healthy lifestyle has increased, …’

3) Last sentence; start the sentence with referring to education, so it is directly clear that results for differences in education will be presented.

Answer: Thank you for this comment. We revised the sentence: “Among high educated men the prevalence of a healthy lifestyle increased between 1990-92 and 2008-11 from 10.6% to 16.3% (p=0.01) and among high educated women from 16.4% to 30.3% and also among medium educated women (10.9 to 16.6, p<0.01), but no significant increase in healthy lifestyle prevalence was observed among men with low and medium education and among women with low education level.”

4) Conclusions: four out of five vs. 4 or 5 mentioned in the method section.

Answer: “at least four out of five” is correct; we revised “4 or 5” to “at least four out of five” (line 37)

5) Ethics statement

Did the 1990-92 participants signed an informed consent?

Answer: No, participants did not sign an informed consent, but verbal consent was witnessed and formally recorded. We added the following sentence into the manuscript (line 107): “Verbal consent was witnessed and formally recorded in the surveys 1990-92 and 1997-1999.”

6) Introduction

Line 73-77: Can authors also relate these numbers to a higher prevalence of health related diseases or premature death?

Answer: We added the following sentences: “In addition, an unhealthy lifestyle with a lower number of healthy behavior factors has been linked to a higher risk of developing type 2 diabetes and cardiovascular diseases and to premature death (15-17)” (lines 70-72).

7) Materials and Methods

Line 92-94: Make clear that you compare prevalence data measured at 3 different time points in 3 different cohort. This is not always clear.

Answer: Thank you for this comment. Indeed, the analysis is based on three population-based, cross-sectional survey samples collected at three different time periods. We revised the following sentence in the section ‘Statistical methods’: ‘Age-standardized and weighted prevalence and 95% confidence intervals (CI) were calculated for each healthy behavior factor separately and for the HLI based on three cross-sectional survey samples collected at three different time periods (1990-92, 1997-99 and 2008-11)’ (page 11, lines 224-25).

8) Line 104: Mentioning the response rate is not in line with informing that response rates for the three surveys were published previously (line 97).

Answer: Thank you for this comment. We revised the sentence in line 97 to ‘Detailed information on the study designs and methods for the three surveys were published previously [1-4].’ (lines 95, 96), but did not change the sentence with information on response rate ‘The response rates were 70% for GNHIES 1990-92, 61% for GNHIES 1997-99 and 42% for GNHIES 2008-11 [4].’

9) Line 108 (as mentioned in the ethics statement): Why did participants from the 1997-99 wave did not sign an informed consent?

Answer: Participants in the 1997-1999 signed an informed written consent, but not in 1990-1992. We revised the sentence: “Verbal consent was witnessed and formally recorded in the surveys 1990-92 and 1997-1999. The GNHIES 1997-99 and GNHIES 2008-11 participants signed an informed written consent prior to participation” (page 6, lines 107, 108).

10) Line 119: ‘Smoking habits were assessed with the questions allowing a distinction between ‘current smokers’ and ‘others’’. This is a strict cut-off value. Please elaborate why this was so strict. Do authors have any information about former smoking?

Answer: Thank you for this comment. Information on former smoking was also available in the three surveys assessed with slightly different wordings. We added this information in line 122.

We selected the cut-off “current non-smoking” because it is an established WHO public health indicator. Furthermore, “current non-smoking” was used in healthy lifestyle scores of previous studies demonstrating strong associations between healthy lifestyle and health outcomes and mortality [5, 6]. We added a brief justification why we used this cut-off on page 9, lines 181-83.

11) Line 114 vs 121: clarify that a healthy diet is defined as fruit and vegetables intake.

Answer: Thank you for this comment. We inserted the sentence: “A daily consumption of both fruits and vegetables was used as an indicator for a healthy dietary pattern” (page 7, line 151).

12) Line 123-147: The explanation of alcohol consumption is too wordy which makes it difficult to follow. Recommend to rewrite/shorten or put to supplement.

Answer: Thank you for this comment. We put the explanation of alcohol consumption into an extra file (S1 File) and inserted the sentence “A detailed description of the alcohol assessment is given in the supplement (S1 File)” (line 149).

13) Line 150: Assessment of fruit and vegetables in 1990-92 is not mentioned to which time period is referred to.

Answer: Thank you for this advice. We changed the sentence into: “Information on frequency of consumption of fruit and vegetables was assessed in GNHIES 1990-92 with the question ‘How often do you consume these particular foods?’ without a specific reference period” (page 7, lines 153-156).

14) Line 148-168: Restructure, so all information is mentioned and provided in the same order.

Answer: We revised the text according to the reviewer’s suggestion.

15) Line 170 is repetitive to line 119-120.

Answer: Thank you for this comment. See also answer on comment no. 10. We revised the text in both paragraphs. In the first paragraph we now focus on the measurement of smoking habits in the three surveys (line 122). In the second paragraph we focus on the variable definition and the justification of the cut-off used (lines 181-83).

16) The outcome section is well written.

Line 186: add that stratification is also done by gender.

Answer: We revised the sentence according to the reviewers suggestion: “We stratified the analyses by gender and used the following age strata….” (page 9, line 200).

17) Line 195-197: ‘After exclusion of individuals with missing data for at least one healthy behavior factor used for the HLI, the final study sample consisted of 7,382 participants for GNHIES 1990-92, 5,603 for GNHIES 1997-99 and 5,073 for GNHIES 2008-11’ were there any differences between participants with missing data and the included sample?

Answer: Thank you for this advice: in GNHIES 1990-92 98.9 % of participants were included in the study sample, in GNHIES 1997-99 96.2 % and in GNHIES 2008-11 94.4 %. We analysed the sample with and without missings and compared the deviation of the weighting factors. As there were no differences in these weighting factors between both samples, we used the complete case sample.

18) Line 207: ‘For age it was standardized to the German population structure as of 31 December 2010.’ To me this is unclear; since you will compare the prevalence of health behavior of people with a certain age (which is fixed since within each cohort you do not check changes over time) in 1990, 1997 and 2008. To what extend does age need to be standardized? Do you mean the results are weighted? And why standardized to 31 December 2010, though data has also been collected in 2011. Please clarify what you mean with this sentence.

Answer: Both is correct, the results are weighted with adaptive weights and age standardized. Adaptive weighting by age, sex, region and educational level was applied in a methodological consistent manner for all three surveys. In addition, the results were age-standardized to the German population structure as of 31 December 2010. The analysis is based on three population-based cross-sectional survey samples collected at three different time periods. As the demographic structure of the population in Germany has changed between 1990-92 and 2008-11 which is also reflected in the sample distributions, in terms that the population became slightly older over time, it is necessary to age-standardise the prevalence of healthy lifestyle across the three surveys according to a reference population. This allows investigating whether healthy lifestyle style prevalence has changed over time independently of the changing age structure of the population. We decided to use the population structure of the last survey (GNHIES 2008-2011) as the reference population to perform the age standardization. The population projection of the Federal Statistical Office of 31 December 2010 was also used for constructing the cross-sectional adaptive weighting factor for GNHIES 2008-11 because this date comes closed to the end of the data collection period of GNHIES 2008-11. We now clarify this in more detail in the ‘Statistical methods’ section (page 11, lines 220-22).

19) Results

Authors present the data in the figure stratified by men and women, but why not by age groups or education level. Please elaborate or potentially add visual stratification in the supplement.

Answer: Figure 1 presents the trend of five health behaviour factors among men and women. Due to differences in some health behaviour factors we presented information by age groups for men and women separately in table S1 and S2. An additional stratification into education level was not valid due to small n per strata. Nevertheless, as suggested we added proportions of healthy behaviour factors according to educational level for the total sample in a Figure in the supplement (S1 Fig). Further stratification was not done as a detailed analysis is part of a next project which investigates absolute and relative social inequalities in health behaviour over time. We inserted the following sentence in the result section: “The increasing trend in sufficient physical exercise and no current risk drinking and the decreasing trend in daily fruits and vegetables intake over time can be observed in all educational groups albeit on a different level (S1 Figure). No significant changes in the prevalence of normal weight over time between different educated groups can be observed” (page 12, lines 248-252).

20) Were there any differences between 1990-92 and 1997-99 or between 1997-99 and 2008-11?

Answer: Yes, we added the following sentence: “More pronounced differences were observed for no current risk drinking with higher increases in prevalences between 1990-92 and 1997-99 compared to 1997-99 and 2008-11. The decrease in daily fruits and vegetables intake was more distinct between 1997-99 and 2008-11 in comparison to 1990-92 and 1997-99” (page 12, lines 237-40).

21) Line 235: was the proportion of men with high number of healthy behavior also significantly different.

Answer: The written sentence is unclear and incomplete. We revised the sentence to “The proportions of men with none or only one healthy behavior factor have declined (relative change: -62.9% and -19.2%, both p < 0.0001), while the proportions of those with three (RC: +25.8%, p = 0.10), four (RC: +42.4%, p = 0.009) and five factors (RC: +66.7%, p < 0.0001) have increased” (page 13, line 258-61).

22) Discussion

Line 271-274 (and line 376): add that trend also include data from 1997-99. Now the impression is that there were only two data collection points.

Answer: We revised the sentence according to the reviewers suggestion: “…overall increased explicitly among adults in the period between 1990-92 and 1997-99 and further slightly between 1997-99 and 2008-11” (page 16, lines 312-13).

23) Line 287-289: “In line with this assumption, the cardiovascular disease mortality has declined in Germany in the last decades and mortality rates are expected to further decline in Germany until 2025 (35-37).” Do authors expect that this is due to a healthy lifestyle or better health care?

Answer: We assume that it is due to both, a healthy lifestyle and better health care and added the following sentence: “This is on the one hand due to a better health care with improved detection and treatment of cardiovascular and metabolic diseases and on the other hand due to an improved health-related behaviour (28)” (page 17, lines 329-331).

24) Line 306-307: can authors give a potential explanation for this observation? Different legislation?

Answer: We insert the following two sentences with a potential explanation: “In Germany in the 1990s there was a discussion on a stronger legal alcohol limit when driving, which resulted in a revised law in 2001 (45). Furthermore, the national legislation has developed the banning of alcohol use at work within the labor protection law and employer agreements at the workplace (46)“ (page 18, lines 351-54).

25) Limitations

The response rate for GNHIES 2008-11 was much lower (42%) compared to the other two surveys (70% 1990-92 and 61% 1997-99). Please provide some explanation for the discrepancies.

Answer: There is evidence that response rates have declined in many studies in many countries in recent years and this is also the case in national health surveys [7, 8]. Nevertheless, declining response rates does not necessary mean that the samples are less representative. It is important that the composition of the realised sample is unbiased as possible. Selection bias can occur at different levels of recruitment (selection of sample points, selection of individuals, participation of individuals).

In the current study weighting factors were applied to adjust for deviations of the sample compared to the general population structure to increase the generalizability of the results. We added a brief paragraph in the limitations section and elaborate more in detail on this issue (pages 21/22, lines 442-47).

26) A healthy diet is considered as daily fruit and vegetables intake. What about fibers, and no consumption of sweets and savory products, or soda?

Answer: Indeed, daily fruit and vegetables intake is only a rough proxy indicator of a healthy dietary behaviour. We added the following sentences in the “Limitations” section: “A consumption of both fruits and vegetables each day is a very rough indicator for a general healthy diet which may include much more aspects e.g. intake of highly processed foods, saturated fat, fiber, sugar intake. However, comprehensive information on the diet was not available in all three surveys”, (page 21, lines 435-38).

27) Fruit and vegetable consumption was assessed by means of a self-administered food-frequency questionnaires. Intake was measured for the past 12 months for 1997-99, while it was measured for the past 4 weeks in 2008-11. Acknowledge the inconsistency and provide any expect differences?

Answer: Thank you for this comment. We added the following sentences: “Consumption was assessed with self-administered food-frequency questionnaires with inconsistent reference periods and slightly different food items and answering categories for the three surveys. Although we tried to standardize this by condensing this information, this may have resulted in systematic differences in estimates for the surveys”, (page 21, lines 438-41).

Reviewer #2:

This is an interesting study which highlight the importance of lifestyle change during the current epidemic of NCD in World.

Reviewer #3:

Thank you for the opportunity to review your article titled, "Time trends in healthy lifestyle among adults in Germany: Results from three national health interview and examination surveys between 1990 and 2011".

Overall, the purpose of this study is interesting, and be well organized paper. However, several issues were still concerned.

Major points:

28) Selection bias: As the authors mentioned in line 370, selection bias (generalizability) would be more important in the present study. The authors must add values of response rates (not proportion of missing of HLI) not only for whole sample in each time point, but also response rates according to sex and age groups if possible. Additionally, the authors should consider generalizability by comparison with complete survey such as German census and the present dataset in each time point (e.g. comparison whether lower education level prevalence are same between German census and the present dataset).

Answer: Thank you for this comment. See also answers on comments of Reviewer 1 no. 17 and 25. Unfortunately, the authors do not have information on response rates according to sex and age groups. Response rates in the German national health surveys were calculated as the number of participants divided by the number of invited sample members reduced by quality neutral losses. Nevertheless, all analyses were weighted. Weighting at the design level affects two probabilities: the selection of a particular sample point and selection of participants within the sample point. Weighting factors were applied to compensate for differences in willingness to participate with regard to the total German population structure according to age, gender, region and educational level. To compensate for a selection bias of the sample, the weighting factors were used to calculate prevalences. Using weighting factors, participants with lower educational level show higher weighting factors compared to participants with medium educational level. Therefore, prevalences in the survey can be generalized to the German population at a defined time point.

From all participants in GNHIES 1990-92 98.9 % of them were included in the analysis, in GNHIES 1997-99 96.2 % and in GNHIES 2008-11 94.4 %. We analysed the sample with and without missings and compared the deviation of the weighting factors. As there were no differences in these weighting factors between both samples, we used the complete case sample.

29) Line 367: It was unclear what did the authors concerned as “bringing about inevitably changes in survey and assessment methods”. Please describe the detail which lifestyle assessment was not differed by period.

Answer: Unfortunately, the questions on smoking, physical activity, alcohol as well as fruit and vegetable consumption differed slightly in the wording between the surveys, e.g. in 1990-92 and 1997-99 smoking was assessed with the question “Have you ever smoked or do you smoke at the moment?” and in 2008-11 “Do you currently smoke – even if only occasionally?”. This was meant with “bringing about inevitably changes in survey and assessment methods“ and these changes should be considered when interpreting the results as we cannot rule out that this fact has been affected the results. According to the reviewer’s suggestion we added in the limitation section the following sentence to describe which assessment was not changed as requested:

“The initial physical exercise question has not changed across the three surveys, but the answer categories have been adapted. The initial smoking question was the same in the first two surveys but was adapted in the third survey. The BMI assessment method based on physical examination data on body weight and height has remained consistent across the three surveys (28).” (page 21, lines 431-35) .

30) To understand what items mainly contributed to HLI ≥ 4, the authors should add results of cross table about proportion of each items of HLI by binary variable of HLI ≥ 4.

Answer: Thank you for this comment, we added a table (top of page 14):

The proportion of sufficient physical exercise, daily fruits and daily vegetables, no current smoking, and normal weight as proportion of HLI increased over time. No current risk drinking did not show any changes over time (page 13, lines 277-79).

Minor points:

31) Tables: “Relative change 1990-2011” is inappropriate, because it would be 1990-92 vs. 2008-11.

Answer: Thank you for the comment! We changed the table description according to the reviewer’s suggestion.

32) Table 3: Title “Proportions (%) of four or five healthy lifestyle indicators” is inappropriate. For example, “Proportions of adhering to healthy lifestyle combination” may be more appropriate.

Answer: Thank you for this comment. We changed the title of the table into “Proportions (%) of adhering to a healthy lifestyle”.

33) Results of decrease trend of fruits and vegetables intake in Germany was definitely reported by the previous study (https://www.ncbi.nlm.nih.gov/pubmed/26934826). Reader would concern why fruits and vegetables intake was decreasing among German people. Have the authors checked the decrease trend according to education level?

Answer: Thank you for the comment! According to the suggestion of Reviewer 1 (comment no. 19) we added an additional Figure in the supplement (S1 Figure) on proportions of healthy behaviour factors according to educational level. Furthermore, the following sentence was inserted in the text: “The increasing trend in sufficient physical exercise and no current risk drinking and the decreasing trend in daily fruits and vegetables intake over time can be observed in all educational groups albeit on a different level (S1 Figure)”, page 12, lines 248-51. Furthermore, we included the suggested reference on trends of fruits and vegetables from Gose et al. 2016 in the discussion on page 17, line 349.

34) Figures: Font sizes should be bigger.

Answer: Font sizes were adapted.

References

1. Hoffmeister H, Bellach BM (1995) [Health of the Germans. A East-West comparison of health data] Robert Koch Institute, Berlin

2. Bellach BM, Knopf H, Thefeld W (1998) [The German Federal Health Survey 1997/98]. Gesundheitswesen 60 Suppl 2:S59-68

3. Scheidt-Nave C, Kamtsiuris P, Gosswald A et al. (2012) German health interview and examination survey for adults (DEGS) - design, objectives and implementation of the first data collection wave. BMC Public Health 12:730

4. Finger JD, Busch MA, Du Y et al. (2016) Time Trends in Cardiometabolic Risk Factors in Adults. Dtsch Arztebl International 113(42):712-719

5. Khaw K-T, Wareham N, Bingham S et al. (2008) Combined Impact of Health Behaviours and Mortality in Men and Women: The EPIC-Norfolk Prospective Population Study. PLOS Medicine 5(1):e12

6. Myint PK, Luben RN, Wareham NJ et al. (2009) Combined effect of health behaviours and risk of first ever stroke in 20,040 men and women over 11 years’ follow-up in Norfolk cohort of European Prospective Investigation of Cancer (EPIC Norfolk): prospective population study. Bmj 338:b349

7. Beullens K, Loosveldt G, Vandenplas C et al. (2018) Response Rates in the European Social Survey: Increasing, Decreasing, or a Matter of Fieldwork Efforts? Survey Methods: Insights from the Field DOI:10.13094/SMIF-2018-00003

8. Tolonen H, Helakorpi S, Talala K et al. (2006) 25-Year Trends and Socio-Demographic Differences in Response Rates: Finnish Adult Health Behaviour Survey. European Journal of Epidemiology 21(6):409-415

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

26 Aug 2019

Time trends in healthy lifestyle among adults in Germany: Results from three national health interview and examination surveys between 1990 and 2011

PONE-D-19-16785R1

Dear Dr. Finger,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

David Meyre

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

29 Aug 2019

PONE-D-19-16785R1

Time trends in healthy lifestyle among adults in Germany: Results from three national health interview and examination surveys between 1990 and 2011

Dear Dr. Finger:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr David Meyre

Academic Editor

PLOS ONE