We are testing a new beta website for gov.scot go to new site

The Scottish Health Survey: Topic Report: Older People's Health



Jennifer Mindell


This chapter considers the findings across the whole report and brings them together. It also compares the findings of logistic regression models among older people in the Scottish Health Surveys with findings from analyses reported in previous years for adults aged 16 and over.


Although the prevalence of self-reported very good health fell with age among older people and the proportion reporting their health as 'fair' increased, the proportion of older adults reporting good, poor and very poor health changed little across the older age groups. Logistic regression models found that in both men and women aged 65 and over, deprivation was very significantly related to the odds of reporting poor or very poor health, with the effect stronger in women. The analyses conducted in 2008 for men and women aged 16 and over 1 had found a similar but stronger trend towards poorer health both in men living in the more deprived quintiles and, to a significant but lesser extent, in women.

While it is unsurprising that ex-smokers (many of whom stopped smoking because of developing a smoking-related disease) and current smokers were more likely to report poor or very poor health, the reduced odds of poor/very poor health in those drinking above the recommended weekly limits appears counter-intuitive. A similar, though less pronounced effect, was seen in men and women aged 16 and over in SHeS 2008. 1 It is likely that this is because the reference category includes not only those who drank alcohol within the weekly guidelines and those who have chosen not to drink alcohol but also those who had stopped drinking alcohol due to disease, whether caused by alcohol or because advised to avoid alcohol. One example of this last category is people with diabetes, where alcohol both contributes calories to those on weight-reducing diets and can precipitate hypoglycaemic episodes. Thus it is likely that this represents, at least in part, confounding by other disease states and should not be interpreted as indicating that exceeding the weekly alcohol limits is to be recommended to reduce the risk of poor/very poor health.

Unlike the SHeS 2008 results for adults aged 16 and over, 1 neither age nor quintile of equivalised income were significantly associated with reporting poor or very poor health among men and women aged 65 and over. Marital status was also not associated with poor/very poor self-rated health. GHQ12 score was significantly related to reporting poor/very poor health but was omitted from the final model as the direction of the association could not be determined ( i.e. whether having symptoms of possible psychiatric disease led to poor/very poor self-rated health, or whether having such poor health led to an increased likelihood of developing symptoms of poor mental health). Unsurprisingly, the presence of a long-term condition, particularly if limiting, was significantly related to reporting poor/very poor health but was also omitted from the final model, as the confidence limits were very wide and the model presented in table 1.2 was limited to demographic, socio-economic, and lifestyle factors, as in the 2008 analyses for all adults.


The results for GHQ12 were not straightforward. Adults aged 65 and over were less likely than younger adults to have a raised GHQ12 score of 4 or more, indicative of possible psychiatric disorder; women were more likely to have a raised score than men. The finding that prevalence was higher among women than men at each age concurs with the results of previous studies of both common mental disorders and some specific psychiatric diagnoses (see section 1.1.1). 2, 3 Those studies have also found that the prevalence of common mental disorders falls with age. 2 While this was true in the current report when comparing those aged 65 and over with younger adults, it was not the case among older women, in whom prevalence increased from the young old to the oldest old.

Positive aspects of mental health (assessed using the Warwick Edinburgh Mental Wellbeing Score ( WEMWBS)) were higher in adults aged 65 and over than in those aged 16-64, but fell with age among adults aged 75 and over. Comparing the results of the logistic regression reported in Table 1.7 for people aged 65 and over in this report with that reported for those aged 16 and over in the SHeS 2008 report, marital status was important in both sexes and SIMD in women but not men in both sets of analyses: the SHeS 2008 results for adults aged 16 and over had found a significant effect in women and a similar but again non-significant trend towards poorer wellbeing in men living in more deprived quintiles. However, income was not significant in older people, and weekly alcohol consumption was not significant in women aged 16 and over. Employment status was not included as an explanatory variable in the analyses of people aged 65 and over, as most older participants had retired and there were few in other categories.

The regression analysis of poor wellbeing also showed that self-reported long-term conditions and physical activity had the strongest associations with poor wellbeing. Physical activity is part of the causal pathway between long-term conditions and poor wellbeing, as current limiting long-term conditions affect the ability to be physically active. In the longer term, physical activity also affects the likelihood of developing a long-term condition but that is less relevant in this cross-sectional analysis. Because of this causal link, the analyses were initially conducted including only one of these two factors. The results shown in Table 1.7 have slightly lower odds ratios for these two variables than when only one of the two is included. Nonetheless, both are strongly related to mental wellbeing even when mutually adjusted for each other and the other significant variables. This is important as physical activity interventions can be designed for those with chronic diseases and may improve their wellbeing, as discussed in sections 6.4 and 6.8 below.

When interpreting the finding that women drinking above the weekly guidance had lower odds of poor wellbeing, it should be remembered that this was in comparison with women who never drank alcohol, no longer drank alcohol, had not drunk in the past week, or had drunk within the guidance. It may be that those drinking but drinking less had different odds from the non-drinkers, but it was decided to use the same variable in the regression as in the models for other outcomes in this report.


Alcohol consumption above the weekly or daily recommended limits was notably lower among older adults than those aged 16-64. However, a recent report from the Royal College of Psychiatrists recommends that the weekly limits and binge drinking definitions for older people should be lower due to physiological and metabolic changes associated with ageing and evidence that older people have higher concentrations of blood-alcohol than younger people after drinking the same volume of alcohol. 4 The report instead recommends a maximum of 11 units per week (and no more than 1.5 units per day) for men and women aged 65 and over, with binge drinking defined as more than 4.5 units for men and 3 for women in a single session. Men aged 65 and over consumed an average of 12.5 units per week, exceeding the 11 units per week advice, whereas older women were below this level by some margin (4.0 weekly units). Older men consumed 2.9 units on their heaviest drinking day in the previous week (almost twice the recommended maximum of 1.5 units). Again, women's drinking was below this limit (1.1 units). The prevalence of binge drinking in older adults clearly increases as the threshold reduces, for example 11% of women had drunk more than 6 units (the current definition for all women), whereas more than twice as many (24%) had drunk more than 3 units (the suggested definition for older women).

It was particularly notable that the proportion of women reporting that they had never drunk rose from one in ten aged 65-69 to over one-quarter of women aged 85 and over, much greater than the rise in men (from one in ten to more than one in seven). This is likely to be the main reason for the increasing difference in mean weekly consumption between men and women with age, as non-drinkers are included in the denominator. Assuming this is a cohort effect, not an effect of age (as the proportion of ex-drinkers changed little), it is likely that alcohol consumption among older women will increase over the coming decades as young women who drink reach older ages.

Among older adults, there were two to three times as many ex-smokers as current smokers, among each age and sex group. The prevalence of current smoking was markedly lower among older than younger adults, reflecting a cohort effect (with older women less likely than younger women to have ever smoked), more time to have quit smoking, and higher mortality rates among smokers. Current smoking among older adults has also dropped in the past decade, from one in four adults aged 65-74 in 1998 to fewer than one in five in 2008/2009/2010. 5 A positive sign for better health in future generations of older men is the higher prevalence of never smokers among young men, matching that seen in both older and younger women. Of concern, however, are the continuing high smoking rates among younger men and women.

Fruit and vegetable consumption remains worryingly low among adults in Scotland of all ages, with only one in five adults eating five or more portions a day. As expected, physical activity declined markedly with age. Overall, more than one in two older adults reported at least moderate intensity activity for 30 minutes less than once a week, with this being three-quarters of adults aged 80 and over. The equivalent figures in the 1998 SHeS were 62% of men and 67% of women (though note that the 1998 estimate was based on activities of at least 15 minutes' duration whereas activities of 10-14 minutes have also been included since 2008). 6 Physical activity remains a very important aspect of aging well, preventing onset and exacerbation of disease; improving functional capacity; and being important for quality of life 7 and mental wellbeing (section 6.3). Although one intervention study found only a weak association between increased physical activity and improved wellbeing (and a weak negative association between sedentary behaviour and wellbeing), 8 a longitudinal study of GP patients aged 60 and over found that habitual activity (including walking and housework) at baseline was positively, though modestly, associated with changes in life satisfaction at follow-up. 9 Importantly, a systematic review showed that exercise interventions can improve mental wellbeing among older, primarily sedentary, people in a community setting. Effective interventions were: designed for such individuals, were delivered to a group by trained leaders, and comprised at least two 45min sessions per week. Such sessions can also improve the mental wellbeing of frail older people. 10 Walking is beneficial for health (see section 2.1.4) although generally older people walk at too slow a speed for it to be included as 'moderate intensity activity'. A study currently underway in Scotland is assessing the effectiveness of encouraging walking among older people. 11

As explained in section 3.1 of this report, the significance of a raised BMI in older people is less clearcut than for younger adults. General obesity was markedly lower in the oldest old, particularly those aged 85 and over and in particular among men. Being cross-sectional data, it is not possible to determine the relative contributions of four factors: a cohort effect, with those born in the early years of the 20th century affected much less by the recent obesity epidemic; a survivor effect, with higher mortality among obese individuals; a greater degree of sarcopenia (loss of muscle mass) among the older old, leading to a lower BMI; and a higher prevalence among the oldest group of diseases that lead to weight loss. Of greater concern is the very high prevalence of abdominal obesity among older adults, regardless of age, affecting almost two in three men and women. This is of particular concern because of the much greater risk of diabetes, metabolic syndrome, and circulatory diseases among those with central adiposity.


The changing prevalence of disease with age among older adults (section 1.3.1) reflects the net effect of two opposing trends. Most diseases increase with age, due to the length of time people have been exposed to factors that can lead to disease (environmental factors and particularly personal lifestyles, interacting with genetic susceptibility). But the most susceptible die prematurely. This balance of two trends is seen most clearly for diabetes, the main disease in the endocrine and metabolic disease category in Table 1.3. Although the incidence of type 2 diabetes increases exponentially with age, diabetes increases the risk of developing and dying from circulatory diseases 12 ; the prevalence of endocrine and metabolic disease falls with age among the oldest women and all but the youngest men aged 65 and over. This probably also explains why for many potentially fatal conditions, the prevalence in men aged 85 and over is lower than in men aged 80-84, while for some non-fatal conditions (ear and eye problems) this was not the case. The lower prevalence for respiratory conditions in older women may represent a cohort effect, with the oldest women the least likely to have ever smoked. 13

A fourth factor is the likelihood of participating in the survey. Thus the prevalence of mental disorders decreases with age partially because people with mental illness have greater mortality from physical illness, particularly circulatory and respiratory diseases, associated with smoking. 14 However, equally importantly, those with significant cognitive impairment ( i.e. all but the mildest forms of dementia) are unable to give informed consent and/or to understand and respond to questions, so are excluded from the survey. This report therefore underestimates the prevalence of dementia and therefore all mental disorders.

With the exception of women aged 65-69, the prevalence of any doctor-diagnosed CVD reported in response to direct questions ( Table 4.1) was substantially higher than the prevalence of long-term conditions of the heart and circulatory system ( Table 1.3), which was limited to those who reported having a longstanding condition, whether or not it limited their activities. Thus a considerable proportion of older adults with cardiovascular disease do not consider themselves to have a long-term condition (assuming there were few who chose to mention six other co-existing conditions first - the maximum permissible in the questionnaire). A similar pattern was seen for men with diabetes, which was more prevalent among those aged 70 and over than the proportion reporting an endocrine or metabolic condition as a long-term condition. That was not the case, however, for women, but there are a number of other conditions apart from diabetes, such as thyroid disease, which fall into that latter category and are more common in women.


6.6.1 Health-seeking behaviour

Despite the high prevalence of longstanding conditions, four-fifths of men and more than three-quarters of women aged 65 and over had not seen a GP within the previous two weeks, with little variation by age except for younger men. A two-week framework provides more accurate recall than asking about the previous three months, and enables average number of visits per year to be estimated. However, it is not possible to ascertain what proportion of survey participants with longstanding conditions see a GP or practice nurse every three or six months for regular monitoring of their condition and treatment.

The proportion of people aged 65 and over in Scotland with two or more emergency hospital admissions in one year has increased over the past decade from 4.2% to 4.9%. 15 Whether this reflects increased prevalence of disease, better access to emergency and secondary healthcare, or worse access to primary healthcare is unclear.

6.6.2 Detection and treatment of disease in older people

Detection of disease is a prerequisite for adequate treatment and control. The proportions of older people with survey-defined hypertension who were untreated, treated but not controlled, or controlled were roughly equal among adults aged 65 and over. This contrasts with younger adults, in whom most of the cases were untreated, suggesting that detection and treatment are better in older people, although both remain sub-optimal. It should, however, be noted that a diagnosis of hypertension requires more than a single measurement, thus 'survey-defined hypertension' may be an overestimate of the true prevalence of hypertension. Additionally, the indications for treating hypertension are a blood pressure of at least 160/100; existing cardiovascular disease or diabetes; or evidence of end-organ damage. 16 Thus drug treatment is not indicated for some adults with hypertension, and this will be more true for younger adults.

A similar pattern was seen for diabetes in men, but not women. A higher proportion of cases of diabetes were diagnosed in older than younger men, but the proportion of undiagnosed cases was twice as high in older than younger women. It is probable that younger women are tested because of healthcare for contraception and pregnancy, as well as greater health-seeking behaviour (see Table 5.3). In contrast, men aged 65 and over, who are considered to be at greater CVD risk, are probably more likely to be tested than women of that age. Additionally, there are more 'older old' among the women, and there can be some reluctance among both patients and healthcare staff to investigate and treat older people for conditions that may have primarily long-term benefits if the shorter term side-effects of diagnosis and treatment (including lifestyle changes) are deemed to affect quality of life adversely, in relation to the long-term benefits. However, nine out of ten older people were taking prescribed medication, regardless of age or area deprivation.


It is estimated that 14% of Scottish households contain someone requiring care. This number is estimated to rise as the elderly population grows (see the Introduction to this report). In 2008/2009/2010, 9% of younger men, 15% of younger women, and 12% of older adults of each sex provided regular care to someone within or outside the home. Apart from the oldest women, these figures changed little by age ( Table 5.1). By 2037 it is estimated that three in five people will act as carers at some point in their lives. 17

Although figures for sheltered accommodation in Scotland have changed little in the past 10 years (about 33,000-35,000 units), the amount of very sheltered accommodation has risen every year, from around 700 units in 1996 to 5,300 in 2010. There has been a drop in less specialised medium dependency housing, from 17,600 units in 1996 to 12,800. 18


The World Health Organisation defines active ageing as:

"the process of optimising opportunities for health, participation and security in order to enhance quality of life as people age."19

Maintaining good quality of life in older people is aided by avoiding depression, maintaining good physical function, and having good neighbourhood standards, financial circumstances, and family relationships. 20

Some factors associated with ageing well apply primarily to older people; others need a lifelong approach and if implemented now, will affect future cohorts of older people. For example, preclinical mobility disability (decreased ability to walk half a mile, climb up steps, do heavy housework, or get in or out of a bed or chair) occurs more commonly in older women with low educational attainment, even after adjusting for income, marital status, number of diseases, and high self-efficacy. 21 The authors of that American study commented that this association is important for primary and secondary prevention and can be assessed easily in healthcare encounters. It is also an additional reason for continued efforts to improve educational attainment among young people. Hence the importance of the Scottish government's recognition in Equally Well that:

"Tackling health inequalities requires action from national and local government and from other agencies including the NHS, schools, employers and Third Sector."22

Living in a deprived area and cigarette smoking, itself socially patterned, were two of the three factors significantly associated with older people in 2008/2009/2010 reporting poor or very poor health. However, community-based walking and exercise programmes can be very cost-effective, with estimates of £7,300 and £12,100 per quality adjusted life year (compared with minimal intervention). 10

In the industrialized countries, premature death has fallen so much that the age at which death is deemed to be 'premature' is revised upwards at regular intervals, with 75 years being the current threshold in Scotland. Although this report has focused on aspects of poor health, it is important to note that more than one in three adults aged 65 and over did not report any long-term condition (including more than one in four aged 85 and over), and more than half described their health as good or very good, including men and women aged 85 and over. With the exception of women aged 85 and over, more than half of older adults in every age and sex group had a GHQ12 score of zero.

Declining rates of disability, cardiovascular mortality, and prevalence of non-communicable disease, and improvements in self-rated health are seen in many countries, resulting in a compression of morbidity and disability. 23 Such improvements are thought to be not only through advances in healthcare but primarily through improved environmental circumstances in childhood and educational and financial resources throughout life. But the higher prevalence of smoking in adults under 65, increasing obesity in middle-aged adults, and poor diet and insufficient activity across all age groups in Scotland in 2008/2009/2010 could, if not tackled effectively, undo some of these advances.