Factors influencing Internet usage in older adults (65 years and above) living in rural and urban Sweden

Abstract

Older adults living in rural and urban areas have shown to distinguish themselves in technology adoption; a clearer profile of their Internet use is important in order to provide better technological and health-care solutions. Older adults’ Internet use was investigated across large to midsize cities and rural Sweden. The sample consisted of 7181 older adults ranging from 59 to 100 years old. Internet use was investigated with age, education, gender, household economy, cognition, living alone/or with someone and rural/urban living. Logistic regression was used. Those living in rural areas used the Internet less than their urban counterparts. Being younger and higher educated influenced Internet use; for older urban adults, these factors as well as living with someone and having good cognitive functioning were influential. Solutions are needed to avoid the exclusion of some older adults by a society that is today being shaped by the Internet.

Keywords

Internet use older adults rural and urban living

Introduction

Globalisation and the Internet have created easier connectivity between people, where we no longer think of location in the same way as 20 years ago. However, there are still rural and urban divides; noticeably, trade and economic development have contributed to a considerable gap in economic conditions between rural and urban living.^1,2

Depending on geographical location, there is also a divide in the use of the Internet.³ What was noted from the beginning of the Internet era 1995⁴ was that the Internet is a valuable source of communication that is quick, affordable and where living location no longer matters.⁵ Ironically it still does. American studies claim that rural living means less extensive use of the Internet,⁶ and also different types of uses.⁷ Some studies conducted showed that older adults (65 years of age and above) living in rural areas contribute to the digital divide, as there is a wide gap in adoption of technology between older adults living in rural versus urban areas.^8,9 Urban living has also been positively related to more computer usage.⁸ Few studies have focused on which variables explain the differences in Internet use and technology adoption by older adults who live in rural or urban settings.

It is important to highlight where the equity of access is less evident, as the Internet can be a valuable resource for older adults; for example, to look up information, maintain social contact and take advantage of possibilities such as online shopping and health status monitoring.¹⁰ In addition, online services can contribute to easier access to health care, where, for example, distance may no longer be a hindrance to receiving adequate care. This can already be seen today with online health-care planning, which can provide more efficient access to health care.¹¹ Living in remote areas or far from a hospital, health-care centre and town can thereby become less of a problem.

The Internet may function as an alternative to face-to-face health-care services for older adults. Health-promoting functions or technological and Internet applications could be a good way for older adults to maintain independence and to live at home as long as possible. This is often a preferred choice by older adults in developed countries.¹² Slowly, the industrialised world is implementing mobile solutions with functions such as reminders to take medications, sensors and audio-visual connections with a nurse.¹³ These possibilities are a solution for the increasing number of older adults who are frail and need special support. There are, however, serious hindrances to the aforementioned technologies. One is that they are often developed without the specific needs and consultation of the elderly or the clinician.¹⁴ Second, they rely on the older adult to be somewhat technologically knowledgeable.¹⁵

It is possible that what we are noticing today is in fact a cohort effect. Today’s older adult is not an evident technology and Internet user, but perhaps the next generation will be. In the last 6 years, there has been an increase in the Internet use among older adults.^16,17 Yet, they are still slower than younger adults to adopt the Internet.^16,18,19 In Europe, for example, there is much less Internet use by senior citizens as compared to younger adults.²⁰ A group that has been especially marginalised are the oldest old – people in their 80s and 90s (octogenarians and nonagenarians) – who are not adapting at all compared to the younger older adults (65–80 years of age).¹⁸

Background (factors affecting Internet usage of older adults)

Age and education are both strong predictors to whether the older adults adapt to the Internet.^21,22 They have been linked directly to computer anxiety, which has been shown to be positively correlated with age; similarly, higher education was linked to interest and computer use for older people.²³ Other studies have shown that older adults will refrain from using the Internet because it is too expensive,¹⁸ which in some studies has been correlated with lower income, household economy and socio-economic status.²⁴ Being male is a factor affecting older adults’ computer and Internet usage – even still in the last decade.¹⁶

If the interface of computers and the Internet do not take into account the cognitive functioning of younger older adults (65–80 years of age) and oldest older adults (80 years and above), as well as normal age-related physical declines, such as eyesight,^19,25 there will be less Internet usage. Studies have even suggested more awareness to create senior friendly websites, taking into account cognitive decline.²⁶ Lower cognitive functioning has been related to not starting to use the Internet in old age.¹⁶ Some studies have highlighted that cognitive aspects are less understood than other age-related declines when it comes to Internet use by older adults.²⁷ Some authors have distinguished between what type of cognitive abilities affect what use, such as problem solving and short-term memory (fluid intelligence) which are strong predictors of web use.²⁸

Living in a rural or urban area may have implications on technology adoption. Boase⁶ has indicated that there is an element of Internet access, which influences the use in rural areas; however, other demographic factors also have a significant relationship on rurality and high-speed Internet adoption. Previous research indicated that Internet access is mainly adopted by older adults if the broadband access is good and there is a perceived usefulness and usability.²⁷

Sweden has a very good Internet connection partly due to its high fibre networks. In 2010, Sweden rated second in most broadband subscriptions per inhabitant in the Nordic countries (where Denmark was first).²⁹ Since 1995, Internet access has increased annually from 2 per cent to 89 per cent in 2012.³⁰ Sweden has been striving for an equal level of access to the total Swedish population, in an effort to work towards the EU legislation on Internet access.³¹ In the last few years, the increase of access has mainly been due to mobile phones and tablets.

If there is less Internet use in rural areas, this can in turn have implications for the provision of care or support for older adults. Studies have shown that many who live in rural areas have different constellations than their urban counterparts in how they are taken care of, relying both on neighbour and family structures more than community services.^32,33 A previous study has shown that older women who live in rural areas require more support, where in some communities, they are struggling to stay living independently and sustain their well-being.³⁴ There seem to be differences in social support, where urban older adults tend to feel more depressed due to anonymity in cities, suggesting they are more lonely, contrary to rural older adults who have stronger social connections that they can trust and depend on.³⁵

Irrespective of whether older adults live in rural or urban areas, research has shown that older adults living with someone have better physical and mental health than older adults living alone.³⁶ Few studies have, however, looked at how older adults who live with someone influences Internet usage per se. Some recent studies demonstrated that older adults were more likely to go online and join social networking websites if they are shown how to, see the perceived usefulness and receive help in the use and process.^37,38 Such support could be facilitated if the older adult lived with someone.

Many factors influence older adults’ Internet use. What this study wanted to shed light on was how these factors varied (if at all) with Internet use and living rural or urban. Thus, the aim of this study was to explore individual responses to whether older adults were using the Internet and why across rural and urban Sweden between the years 2001 and 2004. Data on the spread of the Internet access across Sweden are not available before 2006 for this older adult population. Therefore, the access was not noted.

Two research questions were investigated:

How is Internet usage distributed in rural and urban areas throughout Sweden?

Which factors influence Internet use in these areas, controlling for the variables age, gender, education, living alone or with someone, cognition, household economy and rural or urban lifestyle?

Method

This study has a cross-sectional sample which was taken from the longitudinal Swedish National Study on Aging and Care (SNAC). SNAC is a study conducted since 2001, in which central and local governments cooperate. The total sample consists of sub-samples from four regions in Sweden: Skåne, Kungsholmen, Nordanstig and Blekinge. The sample is considered representative of the Swedish older adults. Different questionnaires and interviews are used to investigate the mental and physical health, distribution of age, gender, perceived quality of life and social and well-being statuses. In effect, the study aims to extend the knowledge of what the medical and functional statuses of older adults are, as well as their needs for social and medical services and identify potential areas of lack of care. The full outline of SNAC can be found in Lagergren et al.³⁹

The population in this study consists of 7181 people in total, from the age of 59 years through to 100 years. The mean age in the sample is 73 years, with a mode of 66 years and a median of 72 years. All information used is from the baseline data (2001–2004) from all four participating regions. The regions had high and equal participation rates, where in the baseline data, there is a non-response rate of 33.6 per cent. Those who do not respond tend to be women, oldest older adults and the frail elderly.

The participants were randomly selected and invited by mail, with one reminder sent. Both written and verbal consent were obtained for all participants. Trained health professionals would thereafter conduct the medical examinations and interviews and also help the older adults to fill in questionnaires when needed. Where there were cases of non-participation, the reason was registered.

Ethics

The research carried out in SNAC complies with the Helsinki declaration; ethical permissions were obtained from the ethical committees of the Karolinska Institute (KI dnr 01-114) and University of Lund (LU dnr 650-00 and LU 744-00).

The definition of urban and rural population

In Sweden, the population consists of close to 9.5 million people, of which the older adult population (65 years old and more) consists of 1.6 million people.^18,40

Sweden does not have a proper definition of what a city is, in physical aspects and in population size; therefore, some choose at times to define regions and make delimitations by closeness to, or easy access to, a place.⁴⁰ The municipalities in this study were categorised according to the Swedish regions by the Bureau of Statistics Sweden:⁴¹

Big city urban – the city of Malmö and Stockholm

Midsize urban – Karlskrona, Eslöv, Ystad and Hässleholm

Rural – Osby and Nordanstig

The data in this study consist of older adults living throughout many regions in Sweden. Two delimitations were chosen to analyse rural and urban living. First, the aforementioned division was used. Second, in order to have a better grasp of the location of the population sample, the authors decided to use a Swedish living conditions’ delimitation of rural or urban.⁴² This is often used in a Swedish context, where if a person has more than 2 km distance to a nearby health-care centre, then he or she is considered to be living rurally. Otherwise, (if they have less than 2 km) they are considered to be living in a closer agglomeration and in a city.

Variables and instruments

Internet use was a question given at baseline (2001–2004) with a yes/no answer possible.

Age was categorised into four groups: group 1 – 59–74 years; group 2 – 75–84 years; group 3 – 85–90 years; and group 4 – 91–100 years.

Education was categorised into three groups according to the previous Swedish education system, relevant for the age groups in this study: level 1, those who did not finish secondary school; level 2, those who finished secondary school; and level 3, those with some form of higher education. Yet for the logistic regressions, the variable was dichotomised, in order to have those higher educated (level 3) and those who were not (levels 1 and 2).

Living alone or not was categorised as either the older adult was living alone or with someone (next of kin, partner), where 1 = living alone and 2 = living with someone.

Cognition was measured using the Swedish translation of the ‘Mini-Mental State’ test (MMT), where maximum score of 30 means that one has normal cognitive functioning. Many studies use a cut-off point between 24 and 26, signifying that under that score, there is presence of cognitive difficulty.⁴³ In this study, each individual result was categorised at the mean (score 27), where 1 = above mean and 2 = below mean.

Household economy was measured based on a Swedish survey on income and living conditions.⁴⁴ The question asks whether one can raise within 1 week a sum of 14,000 SEK (1533 Euro) in 2001–2004 (baseline data).

Statistical procedures

The SPSS Statistics package 20 was used to conduct the analyses below.

A descriptive cross-tabulation analysis was done to investigate the dispersion of the rural and urban divide of 2 km with the following variables: Internet use, age, gender, education, living alone or with someone, cognition and household economy (Table 1).

Table 1.

Cross-tabulation with the dependent variable rural or urban living, with the variables Internet use, age, gender, education, living alone or with someone, cognition and household economy.

	Urban (2 km)	Rural (>2 km)	Total (%)
Internet use, yes	1896 (27.4%) [30%]	129 (1.9%) [19%]	2025 (29.3%)
Internet use, no	4354 (62.9%) [70%]	539 (7.8%) [81%]	4893 (70.7%)
Age: 59–74 years	3626 (52.4%) [58%]	370 (5.3%) [46%]	3996 (57.8%)
Age: 75–84 years	1734 (25.1%) [28%]	209 (3%) [3%]	1943 (28.1%)
Age: 85–90 years	673 (9.7%) [11%]	74 (1.1%) [11%]	747 (10.8%)
Age: 91–100 years	217 (3.1%) [3%]	15 (0.2%) [2%]	232 (3.4%)
Males	3222 (46.6%) [51.5%]	342 (4.9%) [51%]	3564 (51.5%)
Females	3028 (43.8%) [48%]	326 (4.7%) [49%]	3354 (48.5%)
Education, low	2106 (30.7%) [34%]	477 (6.9%) [72%]	2583 (37.6%)
Education, middle	2276 (33.2%) [37%]	82 (1.2%) [12%]	2358 (34.4%)
Education, high	1824 (26.6%) [30%]	99 (1.4%) [15%]	1923 (28%)
Living alone	2913 (42.3 %) [47%]	230 (3.3%) [35%]	3143 (45.6%)
Living with someone	3323 (48.2%) [53%]	428 (6.2%) [65%]	3751 (54.4%)
Cognitive functioning, normal	4083 (60.3%) [67%]	426 (6.3%) [66%]	4509 (66.6%)
Cognitive functioning, poor	2041 (30.1%) [33%]	220 (3.2%) [34%]	2261 (33.4%)
Household economy, good	3089 (44.7%) [49%]	413 (6%) [62%]	3502 (50.6%)
Household economy, poor	3161 (45.7%) [50%]	255 (3.7%) [38%]	3416 (49.4%)

N indicates the number of people; (%) indicates as the percentage of the population sample; [%] indicates the percentage of the rural/urban population.

Urban: having less than 2 km to a health-care centre; rural: having more than 2 km to a health-care centre; education, low: did not finish secondary school; education, middle: finished secondary school; education, high: some form of higher education; cognition, normal: 27 ≤ score ≤ 30; cognition, poor: score > 27; household economy, good: can obtain 14,000 SEK in 1 week (1533 Euro); household economy, poor: cannot obtain 14,000 SEK in 1 week.

Internet usage was compared throughout Sweden using the division big city (Malmö and Stockholm), midsize urban (Karlskrona, Eslöv Ystad and Hässleholm) and rural Sweden (Osby and Nordanstig). A logistic regression was conducted and Internet use was the dependent variable (DV); two dummy variables were created for big city and midsize city, which can be found in Table 2.

Table 2.

Backward logistic regression of Internet use of older adults throughout Sweden.

Model 1
DV	IV	OR	95% CI	p value	Tolerance	VIF	c-statistic
Internet use	Big city	2.500	(2.029–3.081)	0.001	0.33	3.03	0.61
	Midsize city	1.083	(0.861–1.362)	NS
	Rural	1.00	–
Model 2
DV	IV	OR	95 % CI	p value	c-statistic
Internet use	Big city	2.656	(2.366–2.982)	0.001	0.61

DV: dependent variable. IV: independent variable. OR: odds ratio; CI: confidence interval; NS: not significant; VIF: variance inflation factor; c-statistic: concordance statistic.

The big cities (Malmö and Stockholm) were compared with midsize cities (Karlskrona, Eslöv and Ystad) and rural Sweden (Osby and Nordanstig). N = 7181 (included in analysis).

Tolerance and VIF statistic: for multicollinearity; c-statistic: ranging from 0.5 (no discrimination) to 1 (outstanding discrimination).

A second logistic regression (backward) was used to investigate whether Internet use was influenced by the following variables: gender, education, living alone/with someone, cognition, household economy, age and living rural/urban (Table 3).

Table 3.

Backward logistic regression with Internet use as a dependent variable and six independent variables: gender, education, living with someone or alone, normal cognitive functioning, household economy, age and living urban or rural.

Dependent variable	Independent variables	OR	95% CI	p	Tolerance	VIF	c-statistic
Internet use	Being male	1.143	(1.010–1.295)	p < 0.05	0.996	1.00	0.834
	More highly educated	1.762	(1.623–1.914)	p < 0.001	0.88	1.14
	Living with someone	1.784	(1.565–2.033)	p < 0.001	0.91	1.10
	Good/normal cognition	1.222	(1.177–1.268)	p < 0.001	0.84	1.19
	Lower in age	1.118	(1.108–1.127)	p < 0.001	0.78	1.28
	Living urban	1.576	(1.242–2.001)	p < 0.001	0.96	1.04

OR: odds ratios; CI: confidence interval; VIF: variance inflation factor; c-statistic: concordance statistic.

N = 6707 (included in analysis). More highly educated: at least a college or university degree; living with someone: next of kin/partner/friend; normal cognitive functioning between mean 27 ≤ score ≤ 30 (top score); lower in age: continuous variable (59–100 years old); living urban: within 2 km of a health-care centre or shop; tolerance and VIF statistic: for multicollinearity; c-statistic: ranging from 0.5 (no discrimination) to 1 (outstanding discrimination).

A new variable was created with Internet usage and urban or rural lifestyle. Third and fourth backward logistic regressions were conducted to investigate rural Internet use and urban Internet use as DVs with the six independent variables (IVs): gender, education, living alone/with someone, cognition, household economy and age (Tables 4 and 5).

Table 4.

Backward logistic regression: rural Internet usage as dependent variable and gender, education, living with someone or alone, normal cognitive functioning, household economy and age as independent variables.

Dependent variable	Independent variables	OR	95% CI	p	Tolerance	VIF	c-statistic
Rural Internet use	More highly educated	1.368	(1.052–1.780)	p < 0.05	0.95	1.06	0.786
	Lower in age	1.108	(1.075–1.142)	p < 0.001	0.95	1.06	0.786

OR: odds ratios; CI: confidence interval; VIF: variance inflation factor; c-statistic: concordance statistic.

N = 628 (included in analysis). More highly educated: at least a college or university degree; lower in age: continuous variable (59–100 years old); tolerance and VIF statistic: for multicollinearity; c-statistic: ranging from 0.5 (no discrimination) to 1 (outstanding discrimination).

Table 5.

Backward logistic regression: urban Internet usage as dependent variable and gender, education, living with someone/alone, normal cognitive functioning, household economy and age as independent variables.

Dependent variable	Independent variables	OR	95% CI	p	Tolerance	VIF	c-statistic
Urban Internet use	More highly educated	1.813	(1.661–1.979)	p < 0.001	0.90	1.10	0.84
	Living with someone	1.860	(1.624–2.129)	p < 0.001	0.91	1.09
	Normal cognition	1.233	(1.186–1.282)	p < 0.001	0.85	1.18
	Lower in age	1.119	(1.109–1.129)	p < 0.001	0.91	1.09

OR: odds ratios; CI: confidence interval; VIF: variance inflation factor; c-statistic: concordance statistic.

N = 6069 (included in the analysis). More highly educated: at least a college or university degree; living with someone: next of kin/partner/friend; normal cognitive functioning between mean 27 ≤ score ≤ 30 (top score); lower in age: continuous variable (59–100 years old); tolerance and VIF statistic: for multicollinearity; c-statistic: ranging from 0.5 (no discrimination) to 1 (outstanding discrimination).

The variance inflation factor (VIF) and tolerance statistic were calculated for each variable in each regression. The VIF indicates whether a predictor has a strong linear relationship with other predictors, where a VIF higher than 10 and a tolerance (1/VIF) lower than 0.10 indicate multicollinearity may be biasing the model.⁴⁵ This was calculated by running a linear regression analysis with the outcome (DV) and predictors (IV) from the logistic regression. The collinearity of the data was obtained with a VIF and tolerance statistic for each variable. A concordance (c-) statistic was used to measure how well the model discriminated between observations at different levels of the outcome. The range is from 0.5 (no discrimination) to 1.0 (outstanding discrimination). The c-statistic values of 0.7–0.8 are considered to show acceptable discrimination, values of 0.8–0.9 indicate very good discrimination and values of ≥0.9 show outstanding discrimination.⁴⁶ Previous research has indicated that using the c-statistic strengthens the results of the logistic regression in cases where there is strong heterogeneity in the population. The odds ratio (OR) alone is then not enough, especially when having continuous explanatory variables.⁴⁷

Results

Table 1 indicates that there are more older adults living in an urban setting and more younger older adults in this sample. There are almost as many males and females in this sample. More older adults who are highly educated live in an urban setting, yet they have a slightly worse household economy than those living rurally. In this sample, there are more people with normal cognitive functioning than not, in both urban and rural settings.

In the first logistic regression model, it was demonstrated that there are more people using the Internet in big city/urban Sweden compared to midsize and rural. There were significantly more older adults using the Internet in more densely populated regions, compared to those living rurally (OR = 2.500; p < 0.001). VIF and tolerance statistic were calculated afterwards, showing no multicollinearity. The model, however, is not so strong with a low c-statistic of 0.61 indicating that the influence on Internet usage does not just come from the location big city, midsize or rural.

The second model demonstrated that when including all variables into one equation with Internet usage, the household economy variable fell out as non-significant. All other variables significantly increased Internet usage: being male (OR = 1.143; p < 0.05), more highly educated (OR = 1.762; p < 0.001), living with someone (OR = 1.784; p < 0.001), normal cognitive functioning (OR = 1.222; p < 0.001), lower in age (OR = 1.118; p < 0.001) and living in an urban environment – living within 2 km to a health-care centre/shop (OR = 1.576; p < 0.001). The model has a c-statistic of 0.834, showing strong discrimination. The VIF is around 1 and the tolerance is above 0.10 for each variable.

The third logistic regression analysis conducted had the DV rural Internet use. This was based on the rural variable created (living more than 2 km to a health-care centre/shop), and those who were using the Internet contrary to those who were not. All six IVs were investigated. The final model showed that higher education (OR = 1.368; p < 0.05) and lower age (OR = 1.108; p < 0.001) were factors which made rural older adults use the Internet. The c-statistic was 0.786, indicating acceptable discrimination, and there was no multicollinearity between variables as seen with the VIF close to 1 and the tolerance above 0.10.

The fourth logistic regression analysis had as a DV urban Internet use. This was based on the urban variable created (living within 2 km to a shop/health-care centre). Similarly all six IVs were investigated with urban Internet use. The final model showed that more highly educated (OR = 1.813; p < 0.001), living with someone (OR = 1.860; p < 0.001), normal cognition (OR = 1.233; p < 0.001) and lower age (OR = 1.119; p < 0.001) were significant for the final logistic equation. Higher education and living with someone are almost twice as strong predictors of Internet use among older adults in an urban environment. In addition, the c-statistic was above 0.8, indicating strong discrimination between observations in this last model.

Discussion

The main results of this study are that older adults living in urban areas are more frequent Internet users. However, urban living was not enough to explain Internet use of these older adults. Factors that were related to Internet use in general were being male, high education, living with someone, normal cognitive functioning, the younger age groups and urban living. When looking specifically at rural localities, older adults used the Internet if they were more highly educated and if they were younger. Urban older adults used the Internet if they were more highly educated, living with someone, had normal cognitive functions and were lower in age.

When comparing regions in Sweden, big cities showed greater likelihood of Internet use among older adults than their rural counterparts. This is a finding which is supported by other research.^8,17 Even in the second model, when investigating Internet use in general, urban living meant that older adults were more likely to use the Internet (OR = 1.576; p < 0.001). Urban living may entail being surrounded by adverts and bustle, which could make the Internet more attractive. Another reason could be that cities may have quicker bandwidth access compared to their rural counterparts.

Internet use by the older adults was significantly more prominent for males and more highly educated individuals. Whether more education suggests the type of job before retirement, which allowed the older adults to use a computer, or whether it suggests curiosity to pick up new skills is not answered in this study. Learning, however, was linked with higher education in previous studies, where older adults with higher education more easily pick up and try the Internet.⁴⁸

Living alone or with someone stood out in the older adult Internet user as being an influential factor. Those who were not using the Internet were more likely to live alone. There is probably a strong link when living together with someone that makes it easier to use the Internet. Seeing, being exposed to the Internet and having the possibility to ask someone if it goes wrong are all very probable contributory factors. Other studies have indicated that living alone was linked with a lower income, and therefore less Internet use.⁴⁹ Interestingly in this study, there were almost equally as many people living with someone or alone in an urban setting. In contrast to this, there were almost twice as many older adults living together rather than alone.

The younger older adults (65–80 years of age) are using the Internet more frequently than the older age groups. This study shows that age strongly influences Internet use, as age was entered as a continuous variable in the logistic regressions. This means that for every ageing year, there is less Internet usage; just 1 year can make a very big difference. It could be that with increasing age, people are not socially exposed in the same way as each year passes. A study by Cresci et al.¹⁰ found that those who were much older did not see the need to use the Internet. There could be a loss of interest as the years pass. There may also be more physical hindrances. Chronic illnesses become more common, especially as the person passes 80 years of age, in addition to a decline in vision and motor coordination skills which are most likely to change. Studies have shown that cognitive impairment caused people to use technology less.⁸ This study showed that it was also a factor in Internet use among older adults.

Household economy is not influencing Internet usage in this study. This is contrary to other studies.¹⁸ Some investigated ‘late-adopters’ of Internet and broadband, where social networks, the perceived utility, knowing how to use the Internet and how safe it is, but also the price of a subscription, were influential whether ‘late-adopters’ like older adults would want to use the Internet.⁵⁰ In this study’s sample, the purchasing power is less of an influence with the other selected variables.

When further exploring the differences between urban and rural Internet users, a clear distinction could be found between the two. Both rural and urban older adults were influenced by age and higher education if using the Internet. Specifically to urban older adults, living with someone and normal cognitive functioning were the factors influencing their Internet use. The setting where the older adult lives therefore affects the use of the Internet. This may be important within a health-care setting; if the persons in this cohort do not use the Internet, they will also not receive health solutions on the Internet, which are promising better accessibility, information and efficiency to health care.⁵¹

Equal care should be provided irrespective of geographical location. Discrepancies have previously been demonstrated between urban and rural living, especially in predictors of health outcomes for older adults. Access to health information in rural areas can easily be reached via the Internet, possibly reducing health inequalities in the long run.⁵² However, according to the results in this study, those who will not benefit are those who are older, of less education and specifically urban older adults who live alone and do not have normal cognitive functioning.

If older adults do not start to use the Internet via a computer, there will be a lack of interaction with Internet-based health-care services, which will need to be replaced by other ways of communication. For example, by trying other technological devices, such as tablets which may be easier to use than computers, or catering specifically to older adults by providing case managers who can function as a link between health-care centres and the older adult. As claimed by other studies,⁵³ the innovative step is to understand the social roles that are in place today. Older adults have a new and different social role in that they are living longer, engaged longer and are present in society. Older adults who live without the Internet may not want to engage and start using it, which is something further studies should investigate by interview studies. Those not adapting, for different reasons may feel that they are overwhelmed by the information society. It could be that older adults are pushed (forced to use the Internet, an external push) into using the Internet contrary to being pulled (enticed to use the Internet, user-driven) into using it.⁵⁴ It may also be important to take into consideration the impact of the changing constellations of face-to-face contact (online shopping, bank errands and health-care services). A reduction of face-to-face social contact may be felt more with older people, which may be a cause for depression⁵⁵ and ostracism. This, in turn, can also lead to not wanting to start to use the Internet.

Strengths and limitations of the study

Three strengths can be found in this study. First, the study has a large sample of over 7000 people, allowing for generalisations to be made. A second strength is that big city, midsize and rural Sweden are covered in the analyses, giving an overview of older adults’ Internet use in Sweden. Moreover, the chosen delimitation of rural and urban is a workable definition which could leave the research easily transferable. Other definitions of rural and urban may be more country specific. Therefore, a more than or less than 2 km distance to a health-care centre puts rural and urban into a concrete context.

Some limitations of this study are that many factors could be influencing Internet use, which were not investigated here, like health, for example; it may be a strong factor in the rural and urban divide and may influence Internet use. The response rate in the study is 33.6 per cent, which means many are not represented, namely, frail older adults, some women and the oldest old. Furthermore, the data are cross-sectional, so causality cannot be expressed.

Conclusion

There are differences in the Swedish older adults who live in an urban or rural setting and their Internet use. This study confirmed the results from other research that Internet use by the older adults varies across age, gender, levels of education, living situations, normal cognitive functioning, and an urban or rural lifestyle. Specific to this study are the findings that age is a very strong factor influencing Internet use and that the older adults who live in an urban environment, together with someone who has normal cognitive functioning, are more likely to use the Internet.

Footnotes

Acknowledgements

We are grateful to the participants, the counties and municipalities that participated.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of following financial support for the research, authorship, and/or publication of this article: The Swedish National Study on Aging and Care (SNAC; ) is supported financially by the Swedish Ministry of Health and Social Affairs and by the four regions’ county councils.

References

Deller

. Globalization and the rural-urban divide. Am J Agric Econ 2011; 93(3): 912–914.

Sheldon

. Globalization and the rural-urban divide. J Agric Econ 2012; 63(2): 487–489.

Hampton

. Internet use and the concentration of disadvantage: glocalization and the urban underclass. Am Behav Sci 2010; 53(8): 1111–1132.

Giles

. Media psychology. Hillsdale, NJ; London: Lawrence Erlbaum Associates, 2003.

Castells

. The information age: economy, society and culture, volume 1. The rise of the network society. Malden, MA: Blackwell, 1996.

Boase

. The consequences of personal networks for Internet use in rural areas. Am Behav Sci 2010; 53(9): 1257–1267.

Hale

Cotten

Drentea

. Rural-urban differences in general and health-related Internet use. Am Behav Sci 2010; 53(9): 1304–1325.

Calvert

Kaye

Leahy

. Technology use by rural and urban oldest old. Technol Health Care 2009; 17: 1–11.

Marcellini

Giuli

Gagliardi

. Aging in Italy: urban-rural differences. Arch Gerontol Geriat 2007; 44(3): 243–260.

10.

Cresci

Yarandi

Morrell

. Pro-nets versus no-nets: differences in urban older adults’ predilections for Internet use. Educ Gerontol 2010; 36: 500–510.

11.

Hofflander

Nilsson

Eriksén

. Discharge planning: narrated by nursing staff in primary healthcare and their concerns about using video conferencing in the planning session – an interview study. J Nurs Educ Pr 2013; 3(1): 88–98.

12.

Meinow

Kårenholt

Lagergren

. According to need? Predicting the amount of municipal home help allocated to elderly recipients in an urban area of Sweden. Health Soc Care Comm 2005; 13(4): 366–377.

13.

Cheek

Nikpour

Nowlin

. Aging well with smart technology. Nurse Admin Q 2005; 29(4): 329–338.

14.

Kang

Mahoney

Hoenig

. In situ monitoring of health in older adults: technologies and issues. J Am Geriatr Soc 2010; 58(8): 1579–1586.

15.

Ruppe

. Active ageing and prevention in the context of long-term care. Vienna: European Centre for Social Welfare Policy and Research, 2011.

16.

Berner

Rennemark

Jogréus

. Factors associated with change in Internet use by Swedish older adults (2004–2010). Health Inf J 2013; 19(2): 152–162.

17.

Cresci

Yarandi

Morrell

. The digital divide and urban older adults. Comput Informat Nurs 2010; 28(2): 88–94.

18.

Findahl

. Äldre svenskar och Internet 2010 (The Swedish older adults and Internet). Stockholm: Danagårds Grafiska, 2011.

19.

Saunders

. Maximizing computer use among the elderly in rural senior centers. Educ Gerontol 2004; 30: 573–585.

20.

Europe’s digital competitive report 2010 (Internet) . Luxembourg: The European Union, 2010, p. 202, http://ec.europa.eu/information_society/digital-agenda/documents/edcr.pdf

21.

Czaja

Lee

Nair

. Older adults and technology adoption. Proc Hum Factors Ergon Soc Annu Meet 2008; 52(2): 139–143.

22.

Berner

Rennemark

Jogréus

. Distribution of personality, individual characteristics and Internet usage in the Swedish older adult. Aging Ment Health 2012; 16: 119–126.

23.

Ellis

Allaire

. Modeling computer interest in older adults: the role of age, education, computer knowledge, and computer anxiety. Hum Factors 1999; 41(3): 345–355.

24.

Morrell

Mayhorn

Bennett

. A survey of World Wide Web use in middle-aged and older adults. Hum Factors 2000; 42(2): 175–182.

25.

Hawthorn

. Possible implications of aging for interface designers. Interact Comput 2000; 12(5): 507–528.

26.

Beckers

. A study of web usability for older adults seeking online health resources. ACM T Comput-Hum Int 2004; 11(4): 387–406.

27.

Hanson

. Influencing technology adoption by older adults. Interact Comput 2010; 22(6): 502–509.

28.

Czaja

Charness

Fisk

. Factors predicting the use of technology: findings from the center for research and education on aging and technology enhancement (CREATE). Psychol Aging 2006; 21(2): 333–352.

29.

Swedes – the swiftest surfers in the Nordic countries (Internet), http://www.pts.se/en-GB/News/Press-releases/2010/Swedes–the-swiftest-surfers-in-the-Nordic-countries/ (accessed 2 May 2013).

30.

Findahl

. Swedes and the Internet 2012 (Internet). Stockholm: .SE (internetstatistik), 2012, pp. 1–70, http://www.internetstatistik.se

31.

Everyone should be able to use vital public services on the Internet (Internet), http://www.pts.se/en-GB/News/Press-releases/2009/Everyone-should-be-able-to-use-vital-public-services-on-the-Internet-/ (accessed 2 May 2013).

32.

Jett

. Nursing and aging in rural and frontier settings: heal aging. St. Louis, MO: Mosby/Elsevier, 2008, pp. 532–547.

33.

Zulkowski

Coon

. Comparison of nutritional risk between urban and rural elderly. Ostomy Wound Manage 2004; 50(5): 46–48, 50, 52 passim.

34.

Arbuthnot

Dawson

Hansen-Ketchum

. Senior women and rural living. Online J Rural Nurs Health Care 2012; 7(1): 35–46.

35.

Evans

. A comparison of rural and urban older adults in Iowa on specific markers of successful aging. J Gerontol Soc Work 2009; 52(4): 423–438.

36.

You

Lee

. The physical, mental, and emotional health of older people who are living alone or with relatives. Arch Psychiat Nurs 2006; 20(4): 193–201.

37.

Braun

. Obstacles to social networking website use among older adults. Comput Hum Behav 2013; 29(3): 673–680.

38.

Gatto

Tak

. Computer, Internet, and e-mail use among older adults: benefits and barriers. Educ Gerontol 2008; 34(9): 800–811.

39.

Lagergren

Fratiglioni

Hallberg

. A longitudinal study integrating population, care and social services data. The Swedish National Study on Aging and Care (SNAC). Aging Clin Exp Res 2004; 16(2): 158–168.

40.

Städer och stadsstruktur. En förstudie om hur städers betydelse för tillväxt kan beskrivas och analyseras (Internet). Report no. Dnr 2010/014. Östersund: Myndigheten för tillväxtpolitiska utvärderingar och analyser, 2010, http://www.tillvaxtanalys.se

41.

Statistics Sweden (SCB). Immigrant’s migration patterns (Internet). SCB, 2008, http://www.scb.se/en_/Finding-statistics/Publishing-calendar/Show-detailed-information/?publobjid=9033++

42.

Statistics Sweden (SCB). Undersökning av levnadsförhållanden 2007 [Investigation of living standards 2007] (Internet). SCB, 2007, http://www.scb.se/Statistik/LE/LE0101/_dokument/LE0101_BS_2007.pdf

43.

Folstein

McHugh

. ‘Mini-mental state’: a practical method for grading the cognitive state of patients for the clinician. J Psychiat Res 1975; 12: 189–198.

44.

Statistics Sweden (SCB). Undersökning av levnadsförhållanden 2004–2005 [Investigation of living standards 2004–2005] (Internet). SCB, 2005, http://www.scb.se/Statistik/LE/LE0101/_dokument/LE0101_BS_2004_2005.pdf

45.

Field

. Discovering statistics using SPSS. 2nd ed. London: SAGE, 2005.

46.

Hosmer

Lemeshow

Sturvivant

. Applied logistic regression. Hoboken, NJ: Wiley-Blackwell, 2013.

47.

Austin

Steyerberg

. Interpreting the concordance statistic of a logistic regression model: relation to the variance and odds ratio of a continuous explanatory variable. BMC Med Res Methodol 2012; 12(1): 82.

48.

Kiel

. The digital divide: Internet and e-mail use by the elderly. Med Inform Internet 2005; 30(1): 19–23.

49.

Lustria

MLA

Smith

Hinnant

. Exploring digital divides: an examination of eHealth technology use in health information seeking, communication and personal health information management in the USA. Health Informat J 2011; 17(3): 224–243.

50.

Dwivedi

Irani

. Understanding the adopters and non-adopters of broadband. Commun ACM 2009; 52(1): 122–125.

51.

Jung

M-L

Loria

. Acceptance of Swedish e-health services. J Multidiscip Healthc 2010; 3: 55–63.

52.

Zimmer

Wen

Kaneda

. A multi-level analysis of urban/rural and socioeconomic differences in functional health status transition among older Chinese. Soc Sci Med 2010; 71: 559–567.

53.

Biggs

Carstensen

Hogan

. Social capital, lifelong learning and social innovation. In: Beard

Biggs

Bloom

. (eds) Global population ageing: peril or promise? Geneva: World Economic Forum, Global Agenda Council on Ageing Society, 2012, pp. 39–41.

54.

Mante-Meijer

Loos

. Innovation and the role of push and pull. In: Pierson

Mante-Meijer

Loos

(eds) New media technologies and user empowerment. Frankfurt: Peter Lang, 2011, pp. 27–43.

55.

Cotten

Ford

. Internet use and depression among older adults. Comput Hum Behav 2012; 28(2): 496–499.