Social relations and loneliness among older patients consulting their general practitioner

Besides affecting quality of life, limited social relation and loneliness among older persons increase the risk of functional decline and the risk of both physical and mental morbidity (particularly cardiovascular disease and depression) and premature mortality [1-4]. The influence on mortality has been found to be comparable to established risk factors such as physical activity, obesity and smoking [5]. In Denmark, it is assumed that poor social relationships cause 1,000-1,500 annual deaths, equivalent to about 2% of all deaths [6]. The prevalence of loneliness among older persons varies in international studies [7].

The general practitioner (GP) may play a significant role in identifying lonely older persons and in helping prevent that a feeling of loneliness leads to illness and social isolation [8]. The aim of this study was to describe and analyse social relations and loneliness among older patients consulting their GP.

METHODS

Material

This study was a survey counting 12 general practices with a total of 20 GPs located in the Capital Region of Denmark. During a three-week period, each practice consecutively invited their patients aged 65 years and above, regardless of the reason for their visit, to fill out a questionnaire regarding health, social relations and loneliness. Patients gave informed written consent for their participation. Excluded were patients who were unable to speak or read Danish, unable to answer the questionnaire, unable to sign an informed consent form, and patients with severe acute or terminal illness (Figure 1).

Data were collected from February to September 2014. The first author instructed all participating practices in data collection. The GPs received an honorarium of 18 euro for each recruited patient.

Questionnaire

The questionnaire consisted of three parts:

1. Socio-demographics, use of homecare and patient affiliation to the practice.

2. Information about health, smoking and alcohol consumption:

a. Self-rated health measured by a single item

from the Short Form 36 (SF-36) health questionnaire: “In general, would you say your

health is” with the following five response

categories: excellent, good, fair, poor and

very poor [9].

b. Subjective memory complaints measured by

a single item used in primary care studies:

“How would you assess your memory?” with

the following five response categories: excel-lent, good, less good, poor, and miserable [10].

c. Quality of life: The patients completed the

Danish Validated Version of the EQ-5D.

The EQ-5D measures five dimensions –

mobility, self-care, usual activities, pain/-

discomfort, and anxiety/depression –

each by three levels of severity [11].

d. Information on mobility and ability to

see/read a newspaper text and hear a normal

conversation with minimum three people

using items from the Danish national health

interview surveys [12].

e. Information about smoking and drinking habits

using items from the Danish national health

interview surveys [12]. The question on drink-

ing habits was simplified to use per week

instead of each day during the past week.

3. Information about social participation and feelings of loneliness:

a. Social participation: Social participation within

the last months was measured by three

questions: “How often did you” a) have

visitors at home? b) visit others? and

c) participate in social activities outside your

home? With the response categories

“At least once a week”, “Less than once a

week” and “Never” [1].

b. Loneliness was measured by the following

item: “Do you ever feel lonely?” With the

response categories “Yes – often”, “Yes –

occasionally”, “Yes - but rarely” and “No”.

We also used questions from the Danish

national health interview surveys [12]:

“Does it happen that you are alone even

though you want to be with others?” and

“Do you have someone to talk to if you

have problems or need support?”

c. An item for those who were lonely: Asking

them to state if they had talked with their

GP about their loneliness.

We computed two scores based on the three social participation questions. One was the score introduced by Avlund et al [1]. Here the answer “weekly” is assigned one point and the other answers zero points. A total of three points is considered high social participation; while a total score of 0-2 is considered lower social participation. Furthermore, we constructed our own three-level score, since we assumed a likely profound difference

between the categories “less than once a week” and “never” and therefore considered that a dichotomised scale might be too crude. The score we constructed was divided into “high”, “medium” and “low” social participation. We assigned one point to the answer “weekly”, two points to “less than once a week” and three points to the answer “never”. In the cumulated score, three points was considered high social participation equivalent to the scale by Avlund et al, 4-5 points medium and six points and above were considered low social participation.

The loneliness question was dichotomised. The responses “Yes – often” and “Yes – occasionally” were labelled “lonely” and the responses “Yes – but rarely” and “No” were labelled “not lonely”.

Statistics

Differences in variables stratified according to loneliness were analysed using chi-square tests. We used Monte Carlo simulated p-values in cases of a table cell count of less than six observations. For variables with p < 0.05, the associations between the variables and loneliness were analysed using univariate logistic regression. Age and gender were unadjusted; the remaining variables were adjusted for age and gender, and odds ratios were computed. The comparative impact of the variables on the probability of being lonely was assessed by relative importance [13], i.e. the mean increase in model fit

attributable to the addition of a variable to the model (variables with p < 0.05 were included in the relative

importance algorithm). Ethics
The Scientific Ethical Committee for Copenhagen has been informed about the study and assessed it unnecessary to notify (R. no. H-C-FSP-2011-04). The Danish Data Protection Agency (R. no. 2013-41-2393) and the DSAM Multipractice Study Committee (R. no. MPU 24-2013) approved the project.

Trial registration: not relevant.

RESULTS

Of the 762 eligible patients, 476 were included in the study. Of the 291 patients not included in the study,

148 declined participation, 39 had filled out the questionnaire at a previous consultation and 47 were excluded based on the exclusion criteria. A total of 459

patients filled out at least one item about social participation or loneliness (Figure 1). In comparison with the included patients, the non-included participants were significantly older (40% versus 57% above 75 years) (p = 0.0002), but there was no difference in gender distribution (58% and 59% women) (p = 0.7751).

Table 1 presents the distribution of the variables in total and divided according to loneliness.

In the three

individual social participation questions, 59-68% responded “at least once a week”. A cumulated social participation score could be calculated for 437 patients (95%). Based on the scale by Avlund et al, 63.8% had low social participation. Based on our scale, these were further divided into 45.5% with medium and 18.3% with low social participation. Additionally, 17.9% of the patients reported feeling lonely either often or occasionally.

Several items were associated with loneliness (p < 0.05) (Table 1). As seen in Table 2, the odds of feeling lonely were 3.5 times higher for those living alone and four times higher for those with the lowest social participation compared with those with the highest social participation in our scale.

Further, the odds were 39 times higher for those stating often or occasionally being alone when wanting to be with others, and ten times higher for those who only sometimes had people to talk to when having problems or needing support compared with those who often had someone. Additionally, the odds of feeling lonely were 2.5 times higher for people receiving home care, 1.8 times higher for women, 3.3 times higher for those with the lowest compared with the highest self-rated health, 2.9 times higher for those with difficulties hearing a normal conversation and lastly eight times higher for people feeling anxious or depressed.

Additionally, there was a trend towards an association between age and loneliness (p = 0.0747) (Table 1). In an additional analysis with age divided into three

levels (65-74, 75-84, ≥ 85 years), the p-value decreased to 0.0518, and there was a significant odds ratio of 2.6 between the youngest and oldest group (p = 0.0211).

Based on relative importance, the three most predictive variables for feelings of loneliness were whether patients were anxious or depressed (39%), were living alone (27%) and their level of social participation (21%), whereas the remaining variables each explained 0.5-4% of the variance (Table 3).

Despite a clear association between loneliness and the three variables social participation, being alone when wanting to be with others and not having someone to talk to in case of problems or need for support, several patients answered in a manner not fitting the expected association (Table 1). For instance, 12.7% with high social participation reported feeling lonely and, conversely, 63.3% with low social participation reported not feeling lonely. Among the patients who responded that they often or occasionally felt lonely, only 15.2% had discussed their loneliness with their GP (Table 1).

DISCUSSION

In this study, 17.9% of the older patients felt lonely either often or occasionally. The prevalence of low social participation and loneliness in our general practice setting is similar to that found in population-based studies [8, 14]. We found a significant association between loneliness and social participation, being a woman, living alone, receiving home care, being unable to perform usual activities, anxiety/depression, ability to hear a normal conversation and self-rated health. These variables have also been identified in other studies [2, 7, 8, 14]. We also found a trend towards the oldest patients being lonelier. Three of the associated variables accounted for almost all of the variation in reported loneliness; anxiety or depression, living alone and social participation. As expected, loneliness increased with lower social participation. However, several patients were lonely despite having a high social participation or were not lonely despite having a low social participation. A review by Courtin & Knapp [2] also reported a mixed result for the association between social isolation and loneliness.

The measures used are debated in the research about social relations and loneliness [2, 15-17]. Some state that asking directly about loneliness might be stigmatising and might result in incorrect answers [15, 18]. Often, scales based on multiple questions like the UCLA or the Jong Gierveld Loneliness Scale are used, but the contents and the differences between the scales are also debated [2, 16]. We chose to ask directly about loneliness, but also about social participation, being unwantedly alone and having someone to talk to. We found that thought these issues were significantly associated with loneliness; several respondents’ answers fell outside of the expected associations. The study also indicates that a dichotomised version of social participation like the one by Avlund el al might be too crude. Here all questions need to be answered with “weekly” to obtain a score of high social participation and we found big differences in the feelings of loneliness among people with medium and low social participation on our scale. Lastly, it should be noted that to each of the three social participation questions, around two thirds answered “at least once a week”, but when combined in a social participation score only about one third had a high social participation. Hence, this study underlines the need for discussion of the assessment method both in research and practice.

Given the influence of loneliness and low social participation on health and wellbeing, these dimensions of social life are important public health issues [7]. GPs have been proposed as the professional group that is most likely to come into contact with these people, and they are therefore in a unique position to identify them [8]. Municipal nurses conducting preventive home visits with older people will likely have similar opportunities. Based on this study, GPs and nurses should be attentive to loneliness, especially among those who are anxious or depressed, who have low social participation and those living alone, but also among those receiving home care, those who have difficulties performing usual activities or hearing a normal conversation, and finally those with a low self-rated health and likely also the oldest patients. We found that lonely patients rarely discuss these issues with their GP, and a qualitative study by van Ravesteijn et al [19], reported that GPs rarely asked patients directly about loneliness, but either asked indirectly or not

at all. In this context practitioners should be aware of our finding that information about social participation is not always transferable to people’s feelings of loneliness.

As a general-practice-based study, it is a strength that the sampling of participants reflects daily clinical practice and a population in which GPs have an opportunity to consider problems with low social participation and loneliness. However, it is a limitation that we only included patients who are able to visit the practice and to fill out the questionnaire. By not including those receiving home visits, very old and frail patients are probably underrepresented, and they are most likely more lonely than our respondents.

CONCLUSIONS

Among older patients consulting their GP, 17.9% reported being lonely. Only 15.2% of the lonely patients had discussed their loneliness with their GP. Loneliness was associated with low social participation (visits to and by others and leisure activities), being a woman, living alone, receiving home care, not being able to perform usual activities, being unable to hear a normal conversation, self-rated health and feelings of anxiety or depression. GPs should be aware of potential loneliness among patients with these characteristics, especially those who are anxious or depressed, those with low social participation or living alone since these characteristics are highly predictive of feelings of loneliness. They should also be aware that information about social participation is not always transferable to patients’ feelings of loneliness since several patients were lonely despite having high social participation or were not lonely despite low social participation.

Correspondence: Tina Drud Due. E-mail: tina.due@sund.ku.dk

Accepted: 6 January 2017

Conflicts of interest: Disclosure forms provided by the authors are available with the full text of this article at www.danmedj.dk

Acknowledgements: We would like to express our gratitude to Willy Karlslund, Research Unit for General Practice and Section of General Practice, Department of Public Health, University of Copenhagen, for data management. Furthermore, we extend our thanks to Volkert Dirk Siersma, Research Unit for General Practice and Section of General Practice, Department of Public Health, University of Copenhagen, for statistical advice.