Physical Health of Young, Australian Women: A Comparison of Two National Cohorts Surveyed 17 Years Apart

Introduction Very little is known about the extent of physical health issues among young women in early adulthood and whether this is changing over time. Methods We used data from two national samples of young women aged 18–23 years, surveyed 17 years apart, who participated in the Australian Longitudinal Study on Women’s Health. We used multinomial logistic regression to compare the women’s physical health (i.e., self-rated health, common symptoms and conditions) and identify whether sociodemographic factors, health behaviours and stress explained any physical health differences between the samples. Results Women aged 18–23 years in 2013 (N = 17,069) were more likely to report poor self-rated health and physical symptoms (particularly urogenital and bowel symptoms) than women aged 18–23 years in 1996 (N = 14,247). Stress accounted for a large proportion of the physical health differences between the cohorts, particularly for allergies, headaches, self-rated health, severe tiredness, skin problems, severe period pain and hypertension. Conclusions Women’s health appears to be changing, with young women born in more recent decades reporting greater physical symptom levels. Changing socio-cultural and economic conditions may place pressure on young adults, negatively affecting their health and wellbeing. Assessing the extent to which social structures and health care policies are offering adequate support to young women may offer avenues for promoting positive health and wellbeing.


Introduction
Although people in developed countries are living longer, there is evidence to suggest that their health and wellbeing is becoming worse, not better [1,2]. Much of the research focuses on adult health, however, studies from several countries suggest that the decline in health and wellbeing is most apparent among adolescents and young adults. Adolescent girls and young women appear to be at greater risk of poor health and wellbeing than boys and young men [3][4][5], with studies showing that physical symptoms (e.g. headaches, tiredness) and emotional distress are increasing [6][7][8]. The reasons for the apparent health "decline" are complex but are often attributed to individual health (e.g., obesity) and health behaviours (e.g., smoking) and socioecominic factors (e.g., unemployment, education). Sociocultural factors are less often considered [2,9], but Western cultural values that emphasise wealth, materialism and individualism may place pressure on young adults, negatively affecting their health and wellbeing [2,10].
Although some stress is normal, and often expected, for young people as they transition into adulthood [11,12], increasing social pressures may negatively impact on adolescent and adult health [13]. Adolescent worries about school achievement and family conflict have been associated with increasing rates of psychological distress and stress [5,14]. Similarly, adults' concerns about work and family are associated with physical symptoms [15]. Current understandings and perceptions of health and illness are also shaped by social context. For example, Western culture has been criticised for "medicalising" and "pathologising" "everyday problems" [16], particularly among women, leading to an overdiagnosis of physical and mental health conditions [17]. At the same time, there will be young women who have significant physical and mental health issues [17] and early intervention is a priority. However, there is very little evidence regarding the extent of physical health issues among young women in early adulthood and whether this is in fact changing over time.
In this paper we compare the physical health (i.e., self-rated health, common symptoms and conditions) of two national samples of young, Australian women (aged 18-23 years) who were surveyed 17 years apart. We aim to identify whether sociodemographic factors, health behaviours and stress may explain any physical health differences between the cohorts.

Methods
The Australian Longitudinal Study on Women's Health (ALSWH) is a national study focusing on the biological, psychological, social and economic factors relevant to women's health [18]. The original three cohorts of Australian women, who were aged 18-23 years, 45-50 years, and 70-75 years when the project began in 1996, are sent self-report surveys on a triennial basis to explore their health and wellbeing [18]. The 40,000 participants were randomly selected using the national health insurance database (Medicare), which includes all permanent residents of Australia. Women in rural and remote area were oversampled to ensure that their health and health service needs could be adequately assessed. Comparisons with Australian census data show that the three cohorts of women are broadly representative of the Australian population in these age groups [19].
In 2012-13, ALSWH recruited a new cohort of young women born 1989-95 and aged 18-23 when they were first surveyed. Participants were recruited through conventional (e.g., magazine advertising) and online social media avenues (e.g., Facebook) and completed a web-based survey similar to previous ALSWH surveys sent to the young women (full details reported elsewhere) [20,21]. Women were eligible if they lived in Australia, had a valid Medicare number and if they consented to linking survey data with administrative health data. Further details of the survey methodology are available from the study website [22].

Measures
Outcomes. In both surveys, women were asked to rate their general health (excellent, very good, good, fair or poor) using an item from the 36-item Short Form health survey (SF-36) [23], and to report if they had ever been diagnosed with a chronic health condition (e.g. asthma, hypertension, low iron), classified as "yes" vs. "no". Women were also asked about the frequency of 15 symptoms (never, rarely, sometimes, often) in the last 12 months: Somatic symptoms (e.g. headaches, severe tiredness, back pain), urogenital and menstrual symptoms (e.g., urine that burns or stings, leaking urine, heavy periods, severe period pain), digestive and intestinal symptoms (e.g., constipation, hemorrhoids, other bowel problems).
Sociodemographics. Sociodemographic information was also collected from both cohorts and included age (in years); area of residence based on an index of distance to the nearest urban centre (major cities, inner regional, outer regional, remote/very remote) [24]; highest level of education (less than year 12, year 12 or equivalent, certificate/diploma, university degree); current relationship status (never married, never married but in a relationship, married/engaged, separated/divorced/widowed) and parity (yes, no).
Health behaviours. Both cohorts of women were asked about smoking (current smokers, ex-smokers or never smokers) and alcohol consumption (categorised according to the National Health and Medical Research Council Australian Alcohol Guidelines) [25]. Women reported their height (cm) and body weight (kg), which was used to calculate body mass index (BMI) and classified according to the World Health Organization's guidelines: underweight (<18.5 kg/m 2 ), normal weight (18.5 to <25 kg/m 2 ), overweight (25 to <30 kg/m 2 ) and obese (30 kg/ m 2 ) [26].
Stress. Women's level of perceived stress over the last 12 months in specific areas of their life including study, relationships and their own health was measured using the Perceived Stress Questionnaire for Young Women [27]. The questionnaire is used to derive an overall mean stress score, ranging in from 0 (no stress) to 4 (extreme stress). The Perceived Stress Questionnaire is correlated with other psychological and physical health measures and has moderate internal reliability with Cronbach's alpha being 0.75 [28,29].

Statistical Analysis
Multinomial logistic regression was used to compare the physical health of the 1989-95 cohort (18-23 years in 2013) to 1973-78 cohort (18-23 years in 1996). Potential confounding variables were added to the models in blocks to assess their association with physical health outcomes: Sociodemographic variables (age, marital status, education, area of residence, parity), followed by health behaviours (BMI, alcohol consumption, smoking) and then stress. Age (in single years), area of residence and education were included to account for design and response differences between the samples. An interaction term between education and cohort was also included in the models because the association between cohort and physical health differed by education.
Data analysis was conducted using SAS software, version 9.4 (TS1M0) of the SAS System for Windows [30].

Results
The 1989-95 cohort of women were similar to the 1973-78 cohort regarding sociodemographic characteristics: Most were not married, residing in a major city and had no children. There was slight over-representation of better-educated and non-smoking women in the 1989-95 cohort, and a greater proportion of women were obese and reported higher levels of stress (see Table 1).
Compared to the 1973-78 cohort, the 1989-95 cohort appeared to be in poorer physical health. They were more likely to report their self-rated general health as "fair or poor"; and to report allergies, headaches, severe tiredness, back pain and skin problems "often" (see Table 2). Tables 3 and 4 show that women in the 1989-95 cohort also had higher odds of urogenital and menstrual symptoms-particularly leaking urine, vaginal discharge and irregular periods-and constipation, hemorrhoids, other bowel problems. They also had higher odds of low iron but lower odds of hypertension and cancer (see Table 5). With the exception of premenstrual tension and leaking urine, adjustment for sociodemographic characteristics and health behaviours did not make substantive changes to the results.
Adding stress to the models did account for a large proportion of the physical health differences between the cohorts. The estimates for allergies and headaches were not significant when adjusting for stress. In addition, stress attenuated the differences between the cohorts for selfrated health, severe tiredness, skin problems, severe period pain and hypertension, and to a lesser extent, asthma and other bowel problems. The odds of having back pain, urine that burns or stings, vaginal discharge, heavy periods and irregular periods "often" were also lower for women in the 1989-95 cohort after adjustment for stress. However, adjustment for stress did not alter the estimates for cancer, low iron and vaginal discharge.

Discussion
The results of this study, comparing two national cohorts of young Australian women when they were both aged 18-23 years, suggest that young women's health is declining. Women born in 1989-95 were more likely to reported poor self-rated health and physical symptoms (particularly urogenital and bowel symptoms) than women born in 1973-78. Of the young women who were born in 1989-95, less than half rated their health as "excellent" or "good", which is notably lower than what has been reported for [15][16][17][18][19][20][21][22][23][24] year olds in the 2011-12 Australian National Health Survey (63% rate health as "excellent" or "good") [31]. The decline in health may occur early, with a meta-analysis of studies examining the health of children and adolescents reporting an "excess" of physical health symptoms among girls compared to boys, largely for headache, abdominal pain, tiredness, migraine and self-rated health [4].
Poor health among young women may increase the risk of chronic illness in late adulthood and, thus, identifying potential "causes" is necessary for prevention and intervention. While sociodemographic characteristics and health behaviors are associated with physical health outcomes [32], health behaviors explained relatively little of the difference between the health of the cohorts of women in this study. The transition from adolescence to adulthood is marked by engagement in "risky" behaviors (e.g., smoking, alcohol and illicit drug use) [11,33], and with the exception of the decline in smoking, many of these behaviors have not changed over Researchers suggest that changing socio-cultural and economic conditions are part of the complex web of factors influencing young adults' health [2,10]. Young women are more stressed than men [34], which may be the product of "conflicts and ambiguities" for women that are not socially imposed on young men [12]. The decline in young women's health is not unique to Australia, with several other international studies reporting poor health and  wellbeing among young women. Many of these studies are from Sweden [5,7,8,10], which is highly regarded for its health care system and policies [35]. This evidence perhaps points to societal rather than health care deficits explaining the decline in young women's health.
Socio-cultural and medical understandings of health and illness and individual factors (e.g., age, health experiences and expectations) contribute to a person's perception of their health and also influence with whom (e.g. peers, family, general population) people use to compare their health status [36][37][38]. In addition, the rapidly changing medical landscape in Australia, including increased screening, diagnosis and availability of medication may make people more aware of, and focused on, their health. Together, these factors may increase social expectations about what constitutes 'good' health and this may explain some of the apparent self-reported decline in health in the 1989-95 cohort.
Although women born in 1989-95 did not have higher odds of chronic health conditions, they were more likely to report high symptom levels. Because many of the common symptoms examined in this study are also those that are frequently experienced by people in times of heightened stress, disentangling cause and effect can be difficult [13]. However, viewing stress as part of, rather than separate from, poor physical health perhaps offers a way to understand these findings. Similarly, broad definitions and measures of health that do not that assume "health" only in the absence of major disease, will be more effective at capturing the extent of health issues relevant to young people [2,39].
In this study we were able to compare the health of two, national cohorts of women within the same age group, using similar measures, which offers an initial step towards understanding if, and why, young Australian women's health is changing. While it is possible that recent social, economic and health service changes in Australian society (i.e., period effects) explain the differences in physical symptoms between the two samples, we are unable to disentangle period effects from cohort effects in this study due to the absence of longitudinal data for the 1989-95 cohort [40]. In addition, some measures between our surveys were not compatible (e.g. physical activity, psychological distress; diabetes), and others were not available (e.g., social support) [41], and these variables may assist further in explaining the differences between the cohorts. Although the survey administration also differed between the cohorts (paper versus Internet), there is evidence that the reliability and validity of measures of health and wellbeing do not differ between these survey administration modes [42,43]. Women's health appears to be changing, with young women born in more recent decades reporting greater physical symptom levels. Some of the "excess" may be explained by higher levels of perceived stress among our recent cohort of young women, and this may have long-term consequences on women's mental health as well as major life transitions [44]. Measuring societal changes in epidemiologic surveys is challenging but feasible with cross-cohort comparisons such as the present study. Assessing the extent to which social structures and health care policies are offering adequate support to young women may offer avenues for promoting positive health and wellbeing.