Finding a balance: resilience in older adults after depression in later life

Abstract

Older adults who have had a major depressive disorder (MDD) have a high risk of relapse. Although risk factors for depression have been researched extensively, less is known about protective factors, and what experiences might strengthen subsequent resilience and help to prevent relapse. Therefore, this qualitative study explored factors of resilience in older adults who recovered from MDD and did not relapse across at least six years. Twenty-five semi-structured interviews were held with older adults aged 73–85 years who participated in the Netherlands Study of Depression in Older Persons from 2008 to 2014 and were re-interviewed for the present study in 2020–2021. Participants were defined as resilient based on having an MDD diagnosis at baseline but not on two- and six-year follow-up. We used grounded theory coding techniques and thematic analysis to identify factors contributing to resilience. Factors contributing to resilience included: taking agency; receiving social support and engaging in social activities; doing activities individually; and managing thought processes. Resilience after late-life depression appeared to be a dynamic process involving internal and external factors, including finding a balance between rest and activity, between taking initiative and receiving support by others, and between accepting negative emotions and ignoring negative thoughts. Additionally, the ability to learn from depression shows that resilience is not only about avoiding psychopathology, but also about the recovery process and preventing relapse. These findings highlight the need for research and interventions to focus on understanding and influencing the dynamics underlying resilience.

Background and objectives

Recent research has demonstrated that the course trajectory of major depressive disorder (MDD) in older adults is much poorer than in younger adults (Schaakxs et al., 2018). Late-life depression is typically more chronic, once in remission older people tend to relapse more often (Haigh et al., 2018), and mortality rates tend to be higher in older people with depression because of less self-care and higher risk of suicide (Gottfries, 2001). In addition, older adults more often experience somatic symptoms of depression (Hegeman et al., 2012), making depression difficult to diagnose. Symptoms of depression are often regarded as a normal part of the ageing process, hampering treatment of and recovery from MDD (Stoppe and de Mendonça Lima, 2019).

However, despite the relatively poor course of depression in older adults, some older adults successfully recover from MDD and manage to prevent relapse. These older adults can be considered ‘resilient’. Research into the factors contributing to their resilience can help to improve prevention and treatment of depression in the older population in general. Although much research has been done to identify risk factors for depression (Hinrichsen and Emery, 2005; Vink et al., 2009), less is known about protective factors contributing to recovery and sustainable remission of depression (Tanaka, 2018; Ungar and Theron, 2019). By focusing on resilience after late-life depression, the current study contributes to a more general shift away from disease models towards resilience models (Almedom, 2008; Cassidy and Cassidy, 2019), and does so from the perspective of older adults themselves.

Approaching resilience as a dynamic and multi-level concept

During the past decades, research on resilience has been mushrooming (Windle, 2011; Huisman et al., 2017). Originally, resilience was approached from a psychopathological perspective, focusing on individual characteristics that predicted absence of psychopathology despite traumatic experiences, mainly in children and adolescents (Rutter, 1985; Almedom, 2008). Later, resilience research expanded to include ageing (Hayslip and Smith, 2012), and adopted a broader and more dynamic perspective. Recent literature emphasises that resilience may vary across the lifecourse (Windle, 2011) and can itself be shaped by adverse experiences (Rutter, 1981; Windle, 2011). For example, managing to recover successfully from depression may teach older persons how to prevent relapse. Therefore, it is relevant to move beyond the original conception of resilience as avoiding psychopathology despite adversity to a more dynamic view of resilience that includes successful recovery from depression and preventing new depressive episodes.

To date, studies adopting this dynamic view on late-life depression are scant; most studies have focused on identifying factors that contribute to preventing depressive symptoms rather than those contributing to recovery and relapse prevention of depression. For example, the cross-sectional study of Boman et al. (2015), in a population-based sample of older women, found that inner strength was negatively associated with depression, and Burnette et al. (2017) found that social support was negatively associated with depressive symptoms in American Indian and Caucasian older adults. One exception is a qualitative study of Polacsek et al. (2020), which focused on the factors influencing self-management of depression in older adults. They found that self-stigma almost withheld these older adults from seeking help in the first place. The older adults felt treated differently because of their age, emphasising the importance of external factors such as stigmatisation in health care and the patient–care-giver relationship in fostering or hampering resilience with regard to depression. These results suggest that individual and social factors are involved in resilience after depression. Given that the factors contributing to resilience after late-life depression are largely unknown, explorative qualitative studies on this topic can help to illuminate them (Ungar, 2003; Angevaare et al., 2020).

In the present study we adopt an a priori approach to resilience (Huisman et al., 2017), which means that we applied specific researcher-defined criteria to identify a ‘resilient’ group that is most likely to possess relevant experience with resilience after late-life depression. In line with this approach, we conducted semi-structured interviews with older adults who demonstrably recovered from depression in later life and managed to prevent relapse for at least six years. We aimed to answer the following research question:

• • What factors contribute to successful recovery of depression and longer-term relapse prevention from the perspective of older adults themselves?

To answer this question, we conducted semi-structured interviews with selected older adults participating in the Netherlands Study of Depression in Older Persons (NESDO) and applied a thematic analysis to identify factors that contributed to their resilience.

Research design and methods

Research design

In line with our aim to explore the experiences of older adults themselves to identify factors contributing to resilience after depression, we used qualitative research methods. We chose semi-structured interviews to be able to keep the interview focused on the last episode of depression and on potential protective factors contributing to relapse prevention, yet also to provide sufficient flexibility and time to explore with the participant exactly how these factors of resilience worked from their perspective.

Participant selection

Participants were recruited from the sample originally included in NESDO, a multi-site naturalistic cohort study that started in 2008 and lasted until 2014. At baseline, depressed (N = 378) and non-depressed (N = 132) older adults were recruited from in- and out-patient clinics from mental health-care facilities and general practitioners in five regions across the Netherlands (for details, see Comijs et al., 2011). Depression diagnosis, including MDD and/or dysthymic disorder were obtained using the Composite International Diagnostic Interview (CIDI, version 2.1).

For the current study, using purposive sampling methods, we re-approached all NESDO participants through a newsletter sent out at the beginning of 2020, containing information on the current study and an informed consent form to approve being approached for future studies. Although these respondents had taken part in the NESDO study until about six years ago, they had not been in contact with current researchers before. Therefore, all contact with participants was established via the first author LG, in order to create some extent of familiarity with the participants as well as the data gained from them. Based on several criteria, we approached respondents who, based on available quantitative data, were expected to have the most relevant information about resilience after depression. Those who fulfilled our criteria for resilience received a phone call from LG and were asked to participate in an interview. Inclusion criteria were: (1) MDD at NESDO baseline, but not at the two-year and six-year follow-up measurements; (2) not being presently in treatment for psychological problems nor during the past two years; and (3) being cognitively able to participate in this study. Criterion 1 was assessed through analysing available quantitative data, and criteria 2 and 3 were assessed during the approach by phone (see Figure 1). N = 69 participants fulfilled criterion 1. Of these, N = 7 had passed away and N = 12 could not be contacted; N = 9 had not provided consent to be re-approached for research; N = 4 had current emotional problems; N = 1 had dementia; and N = 11 refused to participate because they regarded their depression to be ‘a closed chapter’ that they did not want to talk about.

Figure 1. Criteria for participant selection.

Notes: NESDO: Netherlands Study of Depression in Older Persons. MDD: major depressive disorder.

We assumed that around 20 interviews would be sufficient to answer the research question because the aim was to find patterns in the data based on thematic analysis. Moreover, we expected this number to be achievable, given that N = 41 could be contacted for inclusion in the study and assuming a response of 50 per cent (Figure 1). In total, N = 25 participants were interviewed by LG (see Table 1). For N = 4 participants, it appeared they did relapse into depression during the current study, and therefore did not, in hindsight, meet criteria for inclusion. We decided to include them, still, because data resulting from these participants were found to be useful to gain insight into the boundaries of resilience and negative reinforcement of resilience factors. Two of these participants relapsed in early 2020, one of whom was recovered at the time of interview and one was not. Another participant had a high Inventory of Depressive Symptomatology (IDS) score and said she was not doing well, but she did not want to seek treatment, and one participant relapsed after the interview due to a combination of factors (of which the interview itself was not one according to the participant).

Table 1. Characteristics of participants

Notes: 1. Participant relapsed during the course of the study. IDS: Inventory of Depressive Symptomatology.

Within two months of the interview, LG called the participants to verify whether they received the summary with key points discussed during the interview and a voucher, and to ask the participants how they were doing, as a form of aftercare after talking about a difficult time in their life.

Data collection

We used semi-structured interviews to ensure that several topics would be covered in all interviews, while leaving room for tailoring each interview to the specific story of the participant (see the online supplementary material). Fixed topics were: present daily functioning, experiences and events around the depressive episode which was the reason for their original inclusion in the NESDO study; their subsequent recovery and emotional functioning; and how they manage their mental health in daily life. Furthermore, we asked for their own definition of resilience and the advice they would give to other older adults with depression and their care-givers. The interviews lasted between 45 minutes and two hours, with an average of 90 minutes. Interviews were held between LG and the participant only, except for two participants where the partner was present during the interview, only to complement information the participant had given. To measure current depressive symptomatology in a standardised way, the IDS (Trivedi et al., 2004) was administered at the end of the interview, which took about 10 minutes to complete. During data collection, LG and AK continuously discussed the interview questions, interviews, transcripts, research questions and findings. No significant alterations were made, the emphasis remained on finding out about resilience factors and we wanted to specify how these work in our participants.

Due to the outbreak of COVID-19 in 2020, in the Netherlands and worldwide at the start of this research project, most interviews were held by phone (N = 15) or through video calls (N = 3), based on the preference of the participant. The subsequent (N = 7) interviews were held face-to-face at the participants' homes, adhering to the national hygiene and social distancing guidelines. The summary participants received after the interview to verify key information and provide feedback did not lead to any changes to the results.

The interviews were audio recorded and transcribed verbatim. We adopted a social-constructivist perspective on the data, meaning that the way in which knowledge is brought about is regarded as constructed by joint efforts from the participant and the researcher.

Data analysis

Data collection and data analysis were conducted iteratively. The goal was to derive a set of themes and subthemes that reflect the most important factors contributing to resilience based on the participants' experiences. We applied a thematic analysis to the data, while making use of grounded theory coding techniques, attempting to stay as close as possible to the wordings of the participants – with codes, subthemes and themes thus remaining ‘grounded’ in the data (e.g. ‘being around people’) (Charmaz, 2006: 4). We also coded for actions rather than concepts in order to stay close to how resilience works in the participants' daily lives (e.g. ‘taking part in physical exercise’). MAXQDA was used to document and support the analysis.

The transcripts were first open coded by LG, who was trained in anthropology, based on sections of text mentioning any factors that could have contributed to resilience. To examine coding reliability, the first two transcripts were coded by a second coder who was trained in psychiatry and was not involved in the project. Of the 85 sections of text that were highlighted as being important to be coded, 76 per cent was highlighted by both coders. Codes were also continuously discussed with the author team, to remain reflexive of what would be regarded as important codes.

A preliminary coding tree was made based on open coding of the first seven interview transcripts. Next, axial coding was applied to reach higher levels of abstraction, grouping multiple quotations that received similar codes (e.g. ‘going on with daily life’ and ‘putting things into perspective’ were grouped under ‘managing thought processes’). Then, a set of preliminary themes was derived by LG and AK, and further discussed with the author team, including an old-age psychiatrist, a sociologist, a social epidemiologist and a medical ethicist.

Furthermore, codes were not determined beforehand. Resilience was a sensitising concept in the gathering and analysis of the data, but the data were leading in arriving at codes. Through alternating rounds of interviews and data analysis, including open and axial coding, and discussion among all authors, a definitive set of themes and subthemes was arrived at. After the 16th interview, no new themes emerged, and saturation was initially reached. The following nine interviews were used to enrich these themes, among which were the four participants who appeared to have relapsed. During the writing stage of the final manuscript, LG worked together with a native English editor skilled in Dutch to translate the participants' quotations into English.

Results

Of the 25 participants, nine were male and 16 were female. The age at interview ranged between 73 and 85 years (mean = 77.9, standard deviation = 3.5).

We identified four themes and nine subthemes that reflect factors indicated by participants to contribute to resilience after depression (Table 2). All themes appeared to be important for immediate recovery and for subsequent relapse prevention, except for managing thought processes, which was difficult during the immediate recovery phase, but highly relevant to prevent relapse.

Table 2. Themes and subthemes

Taking agency

Almost all participants indicated that taking agency over the processes of recovery and relapse prevention was crucial for resilience after depression. The following three elements contributed to the transition to a more active role in these processes.

Realising you have to do it yourself

Almost all participants indicated that at some point during their depression they realised that ‘you have to do it yourself’. This was often mentioned as a precondition for recovery. This realisation sometimes occurred prior to treatment, e.g. when one participant decided himself that he wanted to be hospitalised. For others, this happened during treatment, through realising that despite having support, the situation will not change as long as you do not change yourself:

You have to be able to pull yourself together, like, ‘But if I don't want this anymore, so depressed, then I have to do something about it.’ (Ms de G)

For several participants, the realisation that ‘you are not your depression’ was key to taking on a more active role in their recovery and developing motivation to get better. For some participants it was helpful to think of depression as ‘a disease of the mind’, which for them meant that they could play an active role in curing it:

By saying ‘I feel depressed’ – not ‘I am depressed’ – ‘I feel depressed’ makes you aware that you can control that depression … That I am able to do something about it myself. (Mr van L)

Finding a balance between rest, safety and activity

Agency also became salient in the ability to actively seek a balance between rest, safety and activity. In order to profit from rest, it was deemed important to sometimes accept that one was ‘not feeling like doing anything’ and allowing oneself to feel less good for a while. During these periods of rest, it could help to ‘pamper yourself’, e.g. by ‘cooking a nice meal for yourself’ or ‘watching a movie that you like’.

In the context of clinical admittance, feeling safe seemed to be a precondition for being able to really rest and to share personal issues with the therapist. Despite sometimes experiencing unpleasant situations with other patients, the fact that there was always care personnel around made participants feel at ease:

When you're so depressed, you also think about people who ended it. You don't want to, but you do think about it. Then I think ‘As bad as I feel, that must never happen’, so, er, ‘I want to go to the clinic’. And the clinic, yes, I feel safe there, I think ‘Nothing can happen here.’ (Mr G)

Several participants indicated that resting could be crucial to recovery, and this was sometimes also advised by their therapist:

In the morning and in the afternoon, I go to bed. I have to rest for a while again, and will do just that – I don't mind at all … Just last week I went to the doctor, who said ‘you'll have to take it easy’, I said ‘well, I do that already’. And I've learned that too, that I need to slow down. (Mr O)

Nevertheless, participants also mentioned that periods of rest should be followed by periods of activity. As one participant put it, you should not ‘just sit in a chair and hope it will pass’. Finding a balance between rest and activities thus seems crucial.

Understanding which factors make you emotionally vulnerable

Treatment helped participants to recover from depression by providing them with insights into the factors that made them emotionally vulnerable. Some participants were not aware of the impact that (previous) life events had on their emotional functioning, and treatment helped them to feel ‘lighter’, as if they could handle more: ‘It doesn't affect me as deeply. I can distance myself a little further from it’, thus making them more resilient to new challenges.

More generally, some participants stated that successful treatment depends on improving factors that ‘caused the depression’. For instance, for one participant resilience literally meant ‘knowing how to ward off whatever it is that makes you feel depressed’, and treatment could contribute to this. As such, an understanding of vulnerability factors contributed to taking agency over one's recovery, but it was also helpful in order to prevent later depression relapse.

Receiving social-emotional support and engaging in social activities

Social contacts of the participants contributed in two distinct ways to their resilience: while social-emotional support specifically helped to alleviate depressive moods through talking with others, social activity primarily provided distraction, a sense of usefulness and a way to alleviate loneliness that helped to prevent depressive moods.

Receiving social-emotional support

Social-emotional support could come from family, friends or the therapist, and in two forms. First, in the possibility to share your story with someone who is there for you in tough times and listens to you. However, this was only experienced as supportive if these people were genuinely interested and understanding. Concerning the relationship with therapists, it was important for participants to be regarded as ‘a person’ instead of ‘a case’. With this they meant that during treatment, rather than simply making a diagnosis and starting treatment of the respondent, the therapist invited patients to actively express themselves, making them feel understood and providing them with choices regarding their treatment. For other contacts, acceptance, attention, and small acts of care and encouragement made participants feel better, during depression as well as after recovery:

Contact with people around you, who cheer you up, who approach you positively. Because some said, like, ‘This guy is crazy’, but most friends – real friends, those who stayed, yeah … I've had times when I couldn't [play tennis], but he [friend] still called me in. (Mr de W)

The second type of emotional support was more reciprocal, emerging through shared suffering. For example, through exchanging emotional support with someone who had similar experiences:

The week after the day of death of F [son], I also cried here. Yeah. Then Z came in, and he says ‘What's up girl?’ and I say ‘Goddamn it, it's the anniversary of F's death.’ Yeah. Then he strokes my hair anyway, ‘Girl, I know all about it.’ And so, we still support each other, eh? (Ms W)

Still, despite these positive experiences with social-emotional support, some participants felt that depression is a taboo for many people, which made some of them reluctant to talk about it.

Being around people and building relationships

Participants emphasised that staying in touch with people in general contributed to resilience after depression. This ‘being among people’ was enabled through social activities, which according to some participants required taking an active role in ‘gathering people around you’:

Someone who was actually in the same position as me asked, like, ‘How did you do that [dealing with the death of a loved one]?’ and I said ‘You just have to make sure you gather a lot of people around you.’ I say ‘You shouldn't just sit idly at home; you should entertain yourself and find people you can talk to.’ (Ms de H)

Examples of establishing social contacts and meeting people were joining a (sports) club, a course, a workshop or doing voluntary work. These social activities primarily acted to prevent loneliness and as such contributed to resilience. Volunteer work could be a powerful form of social activity, as it provided distraction while also connecting participants to other people. Through this connectedness, participants realised that they were not the only ones with struggles:

For me, it wasn't about the work, it was about the group. Just being together – one has this issue and the other has that issue. That offered me a bit of distraction. (Ms S)

Another respondent who had relapsed reinforced this notion of the importance of social contact, by stating her lack thereof:

I don't really have contact with anyone anymore, that also dulls you, doesn't it? (Ms L)

This example shows in another way that social contact plays a significant role in resilience after depression, and that lacking this contact is not beneficial to the process of staying recovered from depression.

Doing activities individually

Participants also indicated a range of more individually oriented activities that positively influenced their moods and as such contributed to resilience. We grouped them in two subthemes.

Taking part in physical exercise

Participants told that during clinical admission, physical activity was highly recommended by care-givers, and that it indeed worked well to curb depression. Some participants walked often to make sure that they were getting some exercise, and one participant explained that she still went to her yoga classes when she was ‘in the depths of her depression’. Exercise was also often mentioned by participants as a major recommendation to other depressed older adults:

I'd advise people who are temporarily depressed to ‘Please do something – do something you like to do, mostly, go and exercise.’ (Ms B)

Outside the clinical setting, physical activity could offer distraction from sitting at home and produce a feeling of fulfilment. It was deemed important to identify a way of exercising that fits you. However, the extent to which participants could benefit from exercise depended on their physical health; where some participants considered themselves to still be very mobile, others were not able to walk properly anymore. The importance of physical health for maintaining resilience was confirmed by negative examples of two participants who recently relapsed into depression. They indicated that ‘not being able to move around’ like they used to before could or did make it difficult to stay optimistic.

Engaging in activities that feel meaningful

Almost all participants explained that activities they really enjoyed doing individually often provided distraction from difficult events or simply from sitting at home with possible lurking negative thoughts.

These activities were often small and recurrent ones that provided structure to daily life, such as getting dressed every day and putting on make-up. Other examples of satisfactory activities were reading and working on a personal project or hobby. Maintaining discipline and structure through such activities was sometimes actively encouraged by the participants' partner. In the acute phase of depression, ‘doing small things’ could also contribute to resilience, because it produced recurrent success experiences. Being able to complete daily tasks, such as doing the dishes or cleaning up, helped participants to gain back self-confidence, and celebrating small successes was therefore helpful on the way to recovery:

That you can do things, even if they are only very small daily things, or just one thing a day, just to keep on going further. And learn to look back, not like ‘It's not over yet’, but ‘I did it’, that kind of small steps to build yourself up again, so to say. (Mr van L)

Furthermore, for some participants, helping and supporting others provided a sense of purpose that bolstered motivation to tackle one's own negative thoughts and to engage in activities individually. For example, Mr P stated:

I had now set some sort of goal for myself. I think to myself ‘Dammit, I'm going to take care of her: we're having a great time, I don't have to work anymore, we don't have money problems, we have a house.’ (Mr P)

This example shows how setting goals and finding meaningful activities can motivate and stimulate recovery from depression.

Managing thought processes

Although having been recovered from clinical depression for several years, participants indicated that negative thoughts can lie in wait. They described several ways to manage such thoughts, which we grouped in two subthemes.

Leaving negative thoughts behind and going on with daily life

Several participants knew that if they would give in to negative thoughts and worries, it ‘would spoil their day’. Participants suggested that they were able to closely monitor their own moods and thoughts, which helped them to act timely, before negative thoughts would overwhelm them. This enabled them to ‘continue with daily life’. For example, one participant used the mood state in which she woke up as a warning sign for the rest of the day:

There have also been times when I always woke up feeling shitty, if I may put it that way, and that's always an indicator for me. But I also have periods when I wake up happy again, you know. You have to keep a close eye on such things yourself. (Ms T)

Some participants acknowledged that they had a tendency for pessimism, but found ways to actively and consciously shape their own thoughts that otherwise could put them at risk of relapse:

If I sometimes have the tendency to think that [I'm always unlucky], then I think ‘No, no, you shouldn't say that – you shouldn't think that either – because there are very many good other things, which are really good, you are not always unlucky.’ (Ms V)

On some days it could take substantial effort to really get rid of negative thoughts. Various participants said that they were raised with a mentality of ‘don't complain, just get on with it’. But although some participants deemed it necessary to sometimes ‘put your shoulders to the wheel’ on such bad days, they also said that this could be counterproductive. Sometimes they deemed it better to adopt a more flexible attitude to negative thoughts and ‘not be too hard on oneself’. They acknowledged that adversity is a normal part of life: ‘Life is a gift for everyone; sometimes it is a challenge.’ Curbing negative thoughts thus seemed to require a delicate balance between ‘getting yourself up to do something’ and forgiveness to oneself.

Taking lessons forward to life after depression

Some participants indicated that they ‘learned a lot’ from their depression, even though they described it as an unpleasant experience. For example, one participant learned to indicate boundaries:

When people say ‘Oh, you are such a go-getter’, yes, you take that as a compliment, but do you always have to be a go-getter? Sometimes not. Sometimes you have to say ‘That's enough and I won't go any further; that is not good for me.’ I learned that, things like that. (Ms T)

One recently relapsed participant explained that having difficulty with saying ‘no’ was part of the reason for her relapse:

Those were all things at once, and I can't handle that very well, yet it happens, because I can't really say ‘no’. (Ms R)

Other participants appreciated the importance of patience in recovery, and obtained self-confidence from the experience of having recovered from depression before:

And then, that's the most important thing, that you show patience, and when I had it [depression] the second time with that cancer, then I thought ‘I think I will get better, because the first time it also worked out.’ (Mr G)

Such lessons provided participants with more understanding of their own thoughts and behaviours and of how to steer those in order to stay emotionally healthy. As one participant put it: ‘The older, the wiser.’

Discussion and implications

The aim of this study was to provide insight into resilience factors in older adults who have demonstrably recovered from depression in later life persistently or for a longer period of time, while this group tends to have high chances of relapse (Haigh et al., 2018). Based on semi-structured interviews with 25 older adults, we identified four themes reflecting these factors: taking agency, receiving social support and engaging in social activities, doing activities individually and managing thought processes. All themes, except managing thought processes, appeared to be important for the initial recovery from depression as well as for subsequent relapse prevention. Managing thought processes was difficult in the acute phase of depression, but highly relevant for later relapse prevention.

The four themes demonstrate the importance of a multi-level and multi-disciplinary approach to resilience, as suggested by Stainton et al. (2019), Ungar and Theron (2019) and Windle (2011). Resilience after depression resides in individual dispositions and resources, in social contacts, and in the institutional context. Firstly, this research contributes to a better understanding of how internal factors found in previous studies, such as self-management, optimism and an active coping style (Lavretsky, 2012; Boman et al., 2015; Polacsek et al., 2020), contribute to resilience in the face of depression. We found that in treatment as well as in sustaining social contacts after recovery, it is important to take initiative yourself, and not to wait for others to do so. Managing thought processes played an important role in being conscious of your thoughts and allowing negative thoughts to exist and to pass by only to a certain extent, next to putting things into perspective and learning from previous experiences in life.

Secondly, this research emphasises the importance of external factors, such as social support (Singh and Okereke, 2015), often neglected in resilience studies focused on mental health (Ungar and Theron, 2019). Being among others, keeping in touch with people and feeling regarded as a person are related to ‘a feeling of connectedness’ (Aléx, 2010). This boosted participants' self-confidence, provided distraction from lurking loneliness, and sometimes allowed them to care for others and to feel purposeful in doing so.

Thirdly, our research shows that besides social activities, individual activities either as physical exercise or small recurrent daily activities help to sustain resilience. Such activities distracted from lurking negative thoughts when sitting at home alone, provided structure to daily life and helped the participants feel better when they succeeded in carrying out such tasks or activities. In addition, previous research shows that physical activity works in two ways: people with depression are more likely to engage less in physical exercise, and people who engage less in physical activity have a higher risk of depression (Schuch et al., 2017). Therefore, physical exercise seems to be an essential factor of resilience with regard to mental health (Schuch et al., 2016), and our results give insight into which ways this works: by providing distraction, structure and a better feeling about oneself.

Participants' accounts also suggested that resilience can be a precarious process; for some, relapse was always a possibility, and for a few, relapse became a reality. At the same time, participants were confident that they would recover again because they had done so before. This seems in line with the steeling hypothesis by Rutter (1999: 125), who writes that earlier experience(s) of dealing with stress and life challenges can influence ‘our confidence in our ability to deal effectively with life challenges’. As such, earlier experiences may partly mitigate the impact of new adversities on wellbeing. Similar processes have been referred to by others as ‘a new, more evolved centre of self’ (Rosowsky, 2011: 46), a ‘discovery’ of insights (Tanaka, 2018) or ‘growth after adversity’ (Joseph and Linley, 2006). This outcome emphasises a dynamic aspect of resilience, as described by Ungar and Theron (2019), in which resilience is not solely focused on the avoidance of psychopathology despite predisposing factors, as in earlier conceptions of resilience. Rather, extending the concept of resilience to include recovery from psychopathology and relapse prevention can help understand the dynamic nature of resilience processes and highlight potential positive gains of going through depression for maintaining future resilience against stressors (i.e. growth through adversity).

Originally, much research on psychological resilience has focused solely on the question why some people manage to avoid psychopathology despite adversity and trauma (Rutter, 1985). We extend this literature by adopting a dynamic view of resilience, assuming that resilience does not stop at psychopathology. We focused on how older adults bounce back from a depressive episode and may use these experiences to strengthen their resilience to future stressors and prevent relapse. Moreover, while many (quantitative) studies indicated general resources that may contribute to resilience (e.g. Johnston et al., 2015), there is a need to better understand how these resilience factors work in the daily lives of people (Stainton et al., 2019). Our results indicate the importance of finding a balance when recovering from MDD; a balance between rest and activity, between taking initiative and being supported by others, and between sometimes accepting negative emotions and ignoring negative thoughts. The ability to learn from earlier experiences seemed crucial for monitoring and preserving this balance.

Finally, older adults are underrepresented in research on resilience and mental health (Netuveli et al., 2008). Yet one contribution of our study in older adults is that we found that physical ageing might compromise resilience after depression. Some participants related their physical health to the possibility to give shape to relapse prevention, such as being able to engage in physical exercise or to take part in activities, which might be hampered by limited mobility. This is in line with Domajnko and Pahor (2015: 188), who state that ‘health can affect the very ability to be resilient’. The accounts of two participants from our study who relapsed into depression appeared to confirm this. However, some participants were resilient despite physical dysfunction and could draw sufficiently from other resources. The extent to which ageing experiences interfere with resilience could be an interesting avenue for further research.

Implications and recommendations

Participants emphasised the importance of social-emotional support in dealing with depression, e.g. from experts-by-experience. Making use of a buddy system and a peer-to-peer-group, possibly during treatment of depression, may help patients to find such support and it can provide reassurance that they are not alone in this situation.

Based on the current study, personal resources such as having a sense of agency and managing thought processes can be drawn from in treatment and recommendations for relapse prevention. However, this does not mean that individuals should be held personally responsible for a resilient outcome, as social support turned out to be important too (Atkinson et al., 2009). For example, those who have trouble networking might benefit from policy interventions to support social contact.

It thus seems that multiple resilience factors can be combined to support an individual facing adversity in the best possible way. Participants described it as being helpful when therapists make them feel safe and regarded as a person, while they acknowledged that it was also crucial to be encouraged to start feeling responsible for their own recovery at some point. Given the intricate balancing act that we highlighted above, it is important to create space for the personal needs of someone, such as creating rest versus stimulating activity and offering positive affirmation and support versus stimulating own initiatives. Such personal needs and preferences could also concern the choice of medication or type of interventions.

Finally, given the precariousness of resilience in some of the participants, future research could study the sustainability of the resilience factors that we have identified across other difficult situations and new stressors. This could be done in other groups exposed to other adversities, or in a similar group that is followed over time.

Strengths and limitations

A strength of this research is that we interviewed a heterogeneous group of older adults in terms of region, gender and treatment (e.g. whether they were institutionalised for treatment or not). This heterogeneity also extended to four participants who recently had emotional problems despite their previous persistent recovery from MDD, which allowed us to explore the edges of resilience and to look for negative reinforcement of themes. In addition, we used quantitative clinical data obtained earlier over the course of six years to identify a ‘resilient’ group of individuals that was most likely to possess the specific experiences in which we were interested.

In terms of transferability, it might be considered a limitation that among participants, there was no diversity in origin; all participants were from Dutch descent, and we studied resilience processes within the Dutch cultural context. Yet our results can also be relevant for an international audience, because they are largely in line with results from studies in other countries, and add to them. For example, the findings of Polacsek et al. (2020) about self-management after depression among Australian older adults are similar to our findings, focusing on the importance of taking agency and managing thought processes for resilience; and Boman et al. (2015), writing about the importance of inner strength for managing depressive symptoms among older women in Finland, mirror the relevance of extending the focus of resilience research beyond the avoidance of pathology to how people manage psychopathology and relapse prevention. However, we should be careful with interpretation because perspectives on depression and resilience processes elsewhere may be shaped differently (Karasz, 2005; Ward et al., 2014). For example, Ward et al. (2014: 46), writing about a study in older African American women with depression, note that their findings ‘show these older women think of depression as a normal reaction to difficult life situations, rather than an illness’. Still, even in that context, the resilience factors we found may be useful, as they can encourage professionals to be aware of different ways of viewing depression. We believe these resilience factors provide relevant insights into resilience processes within older adults more generally, although it would be valuable to have this confirmed in another study. Also, a relatively high number of participants had finished higher education (eight out of 25 participants). However, there was variation within achieved education. Next to that, it is worth asking the question whether these findings are transferable to people who might be considered not resilient. While the protective factors that we identified may not be directly transferable to people who are not resilient, they may contain useful information and insights that may apply to vulnerable populations too, such as future older adults with depression. On another note, recall bias might have played a role because the diagnosis of MDD occurred around the year 2008. However, for most participants this period was highly imposing and they appeared to remember it very well.

Furthermore, the fact that we spoke to most participants during the initial phase of the COVID-19 pandemic is possibly an important context to be aware of. During interviews, respondents expressed differences in how the pandemic affected them. Some respondents were not so much bothered by the situation, thereby probably not affecting their views on the previous depression. Others emphasised that they found it difficult to maintain good mental health due to the restrictions and uncertainties of the pandemic, potentially leading to more negative accounts of their experiences with and after depression. Although we cannot verify this, we considered it worthwhile mentioning. Additionally, the difference in interview modes between participants (face-to-face versus by (video) call) might have had an influence on the data. However, codes resulting from all interviews did not show any discrepancy between interview modes, and no notable differences in rapport with participants were experienced between interview modes.

Conclusion

This study identified factors that contributed to resilience in older adults after depression from the perspective of older adults themselves. We found that resilience resides in internal and external factors; taking agency after realising your own responsibility in recovering from depression was important, but could only be fruitful in interaction with and supported by others. Furthermore, finding a balance between rest and activity and between acceptance or ignoring negative thoughts or emotions is key to resilience. Finally, learning from earlier experience(s) with depression and self-monitoring is part of relapse prevention. Taking note of these factors in treatment and in providing general mental health information for older adults, their families and their care-givers can support resilience in older adults after depression.

Ethical aspects of the study

Even though the depressive episode took place years ago, and the interview focused on the successful recovery from depression rather than on the conditions that might have given rise to that depression, we were aware that the interview could evoke negative emotion and memories in the participants. Therefore, a number of precautions was made to protect the wellbeing of the participants. First, to provide sufficient time for the participant to change their minds, we planned at least a week between acceptance of the invitation for the interview and the interview itself. Second, we also emphasised that participants were allowed to withdraw from the study at any time without explanation. Third, a psychiatrist was on call during the interviews in case a participant showed signs of requiring acute psychiatric support. Fourth, after the interview we provided a summary of the interview to the participants so they could reflect on it if necessary, and called them by phone to make sure no emotional problems emerged shortly after the interview. All participants were given an explanation of the study before the interview started and were assured that data were pseudonymised. All participants provided written informed consent.