A randomised controlled trial (RCT) exploring the impact of a photography intervention on wellbeing and posttraumatic growth during the COVID-19 pandemic

ABSTRACT

Background

Emerging evidence points to rising levels of psychological distress resulting from the COVID-19 pandemic. There is a need for self-administered, low-cost, and accessible interventions that facilitate wellbeing and growth.

Methods

This study used a randomised controlled trial (RCT) design to investigate the effects of a two-week positivity-oriented photography intervention on wellbeing and posttraumatic growth in comparison to a control group. Participants were adults between the ages of 21 and 80 living in the UK recruited between May and August 2020 (n = 109).

Results

After adjusting for baseline wellbeing, both wellbeing and PTG were significantly higher in the intervention group compared to the control group following intervention completion, with this effect remaining similar at one-month follow-up.

Conclusions

The study offers preliminary evidence that a brief self-administered photography intervention could hold therapeutic value.

KEYWORDS:

• Visual arts
• digital media
• mental health
• psychological wellbeing
• randomised controlled trials

Introduction

Therapeutic impact of photography

Photography can enable the capturing of experiences that may be difficult to summarise with language alone (Frith & Harcourt, 2007), and due to the proliferation of smartphones, may act as an accessible, portable, and flexible self-expression tool. Buchan (2020) reviewed the therapeutic benefits of “participatory photography”, defined as “participant-taken photographs followed by oral or written communication of its meaning” (p. 658), for adults with mental health problems. Participants engaged in different forms of mental processing through their photography and communication, such as reflecting on their lives and experiences, making sense and meaning of their mental health difficulties, and re-authoring self-narratives. Photography was frequently reported to facilitate creative expression, with the representation of experiences using “metaphor, composition, and staging” being therapeutic (Reid & Alonso, 2018, p. 35). Other reported benefits include enabling a sense of achievement and enjoyment, increasing social support through group interactions, and empowering participants (Buchan, 2020). Participatory photography projects that are therapeutically beneficial for individuals may arguably fall under the category of “therapeutic photography” (Saita & Tramontano, 2018, p. 2), which is distinct from “photo therapy” (p.2.), as it does not require a trained professional and is not conducted as part of a course of therapy. A recent pilot RCT found that an intervention involving college students representing the theme “living with my mental illness” through photography led to a significantly greater reduction in anxiety compared to a group counselling intervention (Werremeyer et al., 2020). However, in general there is a dearth of quantitative data in this field, and very few randomised controlled trials (Buchan, 2020), which has been identified as an essential future direction for advancing understanding of photography as a therapeutic tool (Loewenthal, 2020).

Benefit-finding, post-traumatic growth, and positive writing

It has been proposed that photography may enable people to find hope in difficult or painful circumstances and experience personal development (Sackett & Jenkins, 2015), a process which may be likened to post-traumatic growth (PTG; Tedeschi & Calhoun, 1995). PTG refers to the reconstruction of perceptions of self, others, and meanings of traumatic events in the aftermath of trauma (Tedeschi & Calhoun, 1995). PTG has been linked to a variety of positive mental and physical wellbeing outcomes (Affleck et al., 1987; McMillen et al., 1997). Prag and Vogel (2013) delivered photography workshops to Shan migrant youths in northern Thailand, which included photographing things that they loved and hated about living in Thailand. Qualitative analysis suggested that photography had fostered several areas of PTG, including a greater appreciation of life, a recognition of the importance of intimate relationships, and identification of personal strengths.

COVID-19

The COVID-19 pandemic has profoundly changed the everyday lives of people across the globe. On 23 March 2020, the UK government announced a national lockdown, with strict measures introduced to curb the spread of the virus (Diver et al., 2020), including people being allowed to leave their home for only essential purposes, a ban on inter-household mixing, and the closure of restaurants, bars, shops, and offices. Longitudinal cohort studies have identified that anxiety and mental distress have increased, and mental wellbeing has reduced in the UK since the pandemic began (Pierce et al., 2020). Hood et al. (2020) have estimated that there will be an 11% increase in new referrals to mental health services in the UK for the next three years, with associated costs amounting to an extra £1 billion a year, around 8% of the annual NHS expenditure on mental health services. Many have lost loved ones to the disease, are recovering from the trauma of suffering with the illness, or are coping with the fear of catching a potentially life-limiting and life-threatening disease (Maddrell, 2020; Tingey et al., 2020). These are just a few of the many pathways by which COVID-19 may be having a detrimental effect on psychological wellbeing.

However, research has also identified factors which can protect against deteriorating wellbeing under these restrictions. Self-efficacy and optimism have been found to be predictive of resilience (Robles-Bello et al., 2020), psychological flexibility and optimism have been associated with lower COVID-19-related stress (Arslan et al., 2020), and feelings of hope, higher levels of meaning in life, and life satisfaction have been found to correlate with lower state anxiety during the pandemic (Trzebiński et al., 2020). Shamblaw et al. (2021) found that out of a variety of approach and avoidance coping strategies employed by a large Canadian sample of participants during the pandemic, looking for positive aspects of the situation and recognising how they could learn from the circumstances, was the most protective of mental health and quality of life. Amongst the plethora of concerning findings on worsening mental health, these findings highlight the importance of exploring themes of hope, positivity, and growth in the face of COVID-19.

Rationale

Drawing upon research exploring the therapeutic aspects of photography and the mental health benefits of hope and optimism during the COVID-19 pandemic, the following study aimed to explore the effects of a photography intervention delivered to general members of the UK public between May and August 2020. Participants took photographs of, and wrote captions about things which made them feel positive and hopeful. In line with other expressive interventions, such as written emotional disclosure (Riddle et al., 2016), the intervention was self-administered over a relatively brief two-week timespan, which enabled us to capture data specifically during the first national lockdown, at a time when restrictions had begun to lift. Differences in wellbeing and Posttraumatic Growth (PTG) between an intervention and control group were explored.

Hypotheses

Our primary hypotheses were that after completing a photography intervention, participants would show higher levels of wellbeing and PTG compared to the control group. We also predicted that these results would be evident at one-month follow-up.

Method

Design

The study used a two-armed unblinded randomised controlled trial design, with repeated measures administered to participants at baseline, immediately following the two-week intervention, and one-month post-intervention completion. The trial protocol was not pre-registered due to the time limitations of seeking approval to collect data during the early stages of the pandemic. Participants were randomly allocated to the intervention (photography) or control arm. Control group participants were not sent the study instructions, but were told they could opt in to receive these after study completion.

Measures

Demographic information

All online questionnaires and consent forms were designed and administered through the online survey platform Qualtrics© (www.qualtrics.com/uk). Demographic details were gathered using a questionnaire which included gender, age, ethnicity, employment status, education, whether the participant or a family member/loved one had been ill or was currently ill with coronavirus, and their level of photography engagement on a four-point Likert scale, ranging from 0 = never to 3 = regularly. They were also asked whether they were providing care to a person or people, due to the documented impact of the pandemic on caregiving responsibilities (Gallagher & Wetherell, 2020).

Wellbeing

Wellbeing was measured using the Warwick-Edinburgh Mental Well-being Scale (WEMWBS; Tennant et al., 2007). This is a 14-item scale which measures both the feeling and functioning aspects of mental wellbeing over the past two weeks. Items are answered on a 1 to 5 Likert scale, and the scale is scored by summing responses, with a minimum score of 14 and a maximum of 70. It has strong internal consistency (Cronbach’s α = .91) and high test-retest reliability (r = 0.83; Tennant et al., 2007).

Posttraumatic growth

Posttraumatic growth was measured using an adapted version of the Posttraumatic Growth Inventory (PTGI; Tedeschi & Calhoun, 1996). This is a 21-item scale measuring perception of positive changes following a traumatic experience, with five subscales: new possibilities, relating to others, personal strength, appreciation of life, and spiritual change. Items are answered on a 0 to 5 Likert scale, and the inventory is scored by summing responses, with a minimum score of 0 and a maximum score of 105. It has strong internal consistency (Cronbach’s α = .90) and acceptable test-retest reliability (r = .71; Tedeschi & Calhoun, 1996). The questionnaire was reworded to ask participants to rate the degree to which they have experienced such changes since the COVID-19 crisis began, rather than following a traumatic event. Reliability and internal consistency were calculated for the PTGI due to this change in wording, which found excellent internal consistency (Cronbach’s α = .95) and good test-retest reliability (r = .87).

Participants

Ethical approval was obtained from a university ethics committee (1920–010). Individuals were eligible if they self-identified as living in the UK, were over the age of 18, in possession of a smartphone and able to use it to take photographs, had sufficient command of English to participate, and were able to provide informed consent for participation in the study. They were excluded if they self-identified as currently receiving formal psychological or family therapy, or were due to begin this within the course of the study. This was to reduce the potential of interventions received outside of the study confounding outcomes. One hundred and nine participants were recruited between the ages of 21 and 80. There were no statistically significant differences between the groups on any of the demographic variables measured (Table 1).

Table 1. Demographic data for the two study groups

	Whole sample (N=109)	Control (N=53)	Photovoice intervention (N=56)	p-value for group differences
Gender, n(%)				.80
Female	96 (88.1)	46 (86.8)	50 (89.3)
Male	11 (10.1)	6 (11.3)	5 (8.9)
Gender variant/non-conforming	2 (1.8)	1 (1.9)	1 (1.8)
Age, Median(Range)	51.00 (21.00-80.00)	47.00 (24.00-80.00)	52.00 (21.00-72.00)	.56
Ethnicity, n(%)				.54
White British	94 (86.2)	47 (88.7)	47 (83.9)
White Irish	4 (3.7)	1 (1.9)	3 (5.4)
Other White	5 (4.6)	1 (1.9)	4 (7.1)
Indian	3 (2.8)	2 (3.8)	1 (1.8)
Other Asian	2 (1.8)	1 (1.9)	1 (1.8)
Mixed Black and White British	1 (0.9)	1 (1.9)	0 (0.0)
Employment, n(%)				.85
Employed full-time	27 (24.8)	12 (22.6)	15 (26.8)
Employed part-time	16 (14.7)	8 (15.1)	8 (14.3)
Self-employed	11 (10.1)	7 (13.2)	4 (7.1)
Recently unemployed due to COVID-19 situation	4 (3.7)	1 (1.9)	3 (5.4)
Unemployed	4 (3.7)	2 (3.8)	2 (3.6)
Retired	26 (23.9)	13 (24.5)	13 (23.2)
Student	5 (4.6)	1 (1.9)	4 (7.1)
Unable to work due to caring responsibilities, ill health, or disability	8 (7.3)	4 (7.5)	4 (7.1)
Furloughed due to COVID-19 situation	8 (7.3)	5 (9.4)	3 (5.4)
Highest education level, n(%)				.79
Postgraduate qualification	48 (44.0)	22 (41.5)	26 (46.4)
Diploma or degree	40 (36.7)	21 (39.6)	19 (33.9)
A-Levels or equivalent	12 (11.0)	7 (13.2)	5 (8.9)
GCSEs, CSEs, or equivalent	7 (6.4)	2 (3.8)	5 (8.9)
Left school at 16, did not sit/pass exams	1 (0.9)	1 (1.9)	0 (0.0)
Missing	1 (0.9)	0 (0.0)	1 (1.8)
Experienced COVID-19, n(%)				.88
No	76 (69.7)	36 (67.9)	40 (71.4)
Yes	6 (5.5)	3 (5.7)	3 (5.4)
Had symptoms but not tested	12 (11.0)	5 (9.4)	7 (12.5)
Unsure	13 (11.9)	8 (15.1)	5 (8.9)
Prefer not to say	2 (1.8)	1 (1.9)	1 (1.8)
Family members or loved ones experienced COVID-19				.72
No	74 (67.9)	34 (64.2)	40 (71.4)
Yes	8 (7.3)	4 (7.5)	4 (7.1)
Had symptoms but not tested	17 (15.6)	10 (18.9)	7 (12.5)
Unsure	9 (8.3)	5 (9.4)	4 (7.1)
Prefer not to say	1 (.9)	0 (0.0)	1 (1.8)
Acting as a caregiver, n(%)				.54
Yes	28 (25.7)	15 (28.3)	13 (23.2)
No	81 (74.3)	38 (71.7)	43 (76.8)
Engagement with photography, Median(IQR)	3.00 (1.00)	3.00 (1.00)	3.00 (1.00)	.97

Table 2. Descriptive statistics and tests of difference on outcome measures between the control and intervention group at baseline

Measure	Whole sample (N=109)	Control (N=53)	Intervention (N=56)	p-value for group differences
Wellbeing, M(SD)	45.68 (7.74)	47.26 (6.86)	44.19 (8.27)	.04*
PTGI, Median(IQR)	57.00 (35.00)	56.00 (33.50)	57.00 (36.50)	.87
Positive emotions and coping, M(SD)	55.41 (18.87)	57.42 (19.60)	53.51 (18.12)	.28

* indicates a significant result at the α=.05 level

Figure 1 shows the assignment and attrition at various points during the study.

Figure 1. Flow of participants through the study.

Figure 2. Graph showing wellbeing estimated marginal means for the control and intervention group. Error bars represent 95% confidence intervals.

Figure 3. Graph showing posttraumatic growth estimated marginal means for the control and intervention group. Error bars represent 95% confidence intervals.

Procedure

Recruitment took place from May 20th to August 31st in 2020. Participants were recruited using volunteer sampling. A study advertisement was shared widely on Twitter and posted in community Facebook groups, which indicated that the research was exploring the effects of a photography intervention on wellbeing during the COVID-19 crisis, emphasising that no prior interest or skill in photography was required. Whilst an effort was made to share the advertisement to community Facebook groups across the UK, stratified sampling was not employed. The study advertisement instructed those who were interested in finding out more about the research to email the lead researcher. They were then sent an information sheet, and if they remained interested, were emailed a link to complete an online consent form. Once this was completed, participants were randomly allocated to the control or intervention arm according to a simple computer-generated block randomisation list produced using Sealed Envelope™ (https://www.sealedenvelope.com/). They were not stratified on any baseline characteristics. Participants were emailed a link to complete the baseline questionnaires, which included the demographic questionnaire, the WEMWBS, and the PTGI. Once the questionnaires had been completed, they were informed of their group allocation by email.

Intervention group

Participants were informed of their allocation and received study instructions to take photographs over a two-week period of things that made them feel positive and hopeful (see supplemental online material for instructions). There were no restrictions as to the quantity, content, or quality of the photographs, and it was explained that the photographs could be as abstract or as literal as they liked. Participants were asked to write captions for the photographs which could be of any length. It was suggested that these could include descriptions of the photographs or reflections or insights about them. At the end of each week, participants were instructed to choose five photographs to send to the researcher, and it was suggested that these could be the photographs that they felt were particularly meaningful or representative of their experiences, or simply their favourites. Guidance was provided on how to securely share the photographs with the researcher. Participants were emailed reminders about sending their photographs on the appropriate dates, but flexibility was offered if they needed extra time to complete the intervention or send their photographs. Examples of photographs and captions submitted by participants can be found in the supplemental online material.

Control group

Participants in the control group were informed that they would not receive the study instructions at that time, but could opt in to receive them once the one-month follow-up period had been reached (see supplemental material for instructions). They were not instructed to stop any current photography practice during the study period.

Questionnaires

After completing the intervention, or after the two-week study period for the control group, all participants were emailed the WEMWBS and the PTGI to complete a second time. The intervention group participants were also asked to state which photographs, if any, they would be happy to be included in the write up of the research. Approximately one month after these questionnaires were sent, the third and final questionnaires were emailed. This final questionnaire also asked participants to rate the extent to which they had been engaging with photography over the past month. Continuing the photography was not encouraged as part of the research but the question was included to determine if practices had continued for analysis. Participants were sent a debriefing statement, and instructions were sent to six control group participants who opted to receive them. These instructions were the same but suggested that instead of sharing the photographs with the researcher, they could either keep the photographs for themselves, or share them with friends or family members.

Analysis plan

An a priori analysis plan was made to compare intervention and control group participants, as well as participants who dropped out of the intervention and those completed it, on baseline characteristics. It was planned that two mixed ANOVAs would be conducted, with “group” being the between-subjects factor, “timepoint” being the within-subjects factor, and wellbeing and PTG being separate dependent variables. This test is fairly robust to violations of normality so would have been conducted regardless of whether the data suggested a parametric or non-parametric distribution (Blanca Mena et al., 2017).

Results

Testing for baseline outcome measure differences

The control and intervention groups were compared on their baseline scores on the three outcome measures using two-tailed independent samples t-tests. Table 2 shows the descriptive statistics and difference tests. Wellbeing was significantly higher in the control group than the intervention group at baseline; M = 3.08, 95% CI [0.18, 5.97], t(107) = 2.19, p = .04. Therefore, randomisation had failed to control for this outcome measure.

Baseline differences between completers and non-completers

Tests of difference were performed on baseline characteristics between those who completed the intervention (N = 42) and those who had dropped out of the study (N = 14). A Fisher’s exact test showed there was a significantly higher proportion of caregivers in the group of participants who did not complete the intervention (50% caregivers) compared to those who completed it (14.3% caregivers), p = .01. There were no other statistically significant differences on baseline characteristics between the groups.

Missing data analysis

Due to drop-out at the follow-up point reducing statistical power and the risk of bias associated with only analysing participants with complete data, a missing data analysis was conducted post-hoc. Missing data were judged to be at random, assessed by Little’s Missing Completely at Random (MCAR) test; χ²(8) = 12.55, p = .13. The Last Observation Carried Forward (LOCF) method was used due to this being more likely to lead to conservative estimates of the effect (Salkind, 2012).

Wellbeing

Due to a baseline imbalance between the intervention and control group on the primary outcome measure of wellbeing, the a priori analysis plan was deviated from, and it was decided post-hoc to include baseline wellbeing as a covariate in both analyses. Two mixed ANCOVAs were conducted to determine whether there were differences between the control and intervention groups on wellbeing and PTG at the post-intervention and one-month follow-up timepoints when controlling for baseline wellbeing. Assumptions of linearity, homogeneity of regression slopes, normality of residuals, homoscedasticity, homogeneity of variances, and residual outliers were largely met, indicating that ANCOVAs could be conducted. The output for the main analysis can be found in the supplemental online material.

The mixed ANCOVA identified a significant main effect of group, with wellbeing being significantly higher in the intervention group (adjusted M = 47.52, SE = 0.71) compared to the control group (adjusted M = 44.89, SE = 0.73) after controlling for baseline wellbeing; F(1, 106) = 6.59, p = .01. The effect size was medium (partial η² = .06; Cohen, 1988). Figure 2 shows the estimated marginal means for wellbeing for the two groups at the post-intervention and follow-up timepoints. There was no significant interaction effect between time and group, indicating that both the intervention group (post-intervention adjusted M = 47.62, SE = 0.70; follow-up adjusted M = 47.42, SE = 0.81) and control group (post-intervention adjusted M = 44.66, SE = 0.72; follow-up adjusted M = 45.11, SE = 0 .83) showed a comparable lack of change from post-intervention to follow-up; F(1, 106) = 0.71, p = .40. The effect size was small (partial η² = .01).

There was a significant main effect of time; F(1, 106) = 7.04, p = .009, with a medium effect size (partial η² = .06). However, the pairwise comparison showed no significant difference between wellbeing at post-intervention (adjusted M = 46.14, SE = 0.50) and follow-up (adjusted M = 46.27, SE = 0.57) across groups when baseline wellbeing was controlled for; p = .74. There was also a significant interaction effect between time and baseline wellbeing; F(1, 106) = 6.95, p = .01, with a medium effect size (partial η² = .06). To explore this relationship further, a post-hoc Pearson’s correlation analysis was conducted, which showed a significant weak negative correlation between baseline wellbeing and post-intervention to follow-up change scores; r(109) = −.24, p = .01. This suggests the lower a person’s baseline wellbeing, the more likely they were to experience some positive change from post-intervention to follow-up and vice versa. A significant main effect of baseline wellbeing was found, suggesting that the covariate was significantly related to wellbeing at the post-intervention and follow-up timepoints; F(1, 106) = 146.39, p < .001. The effect size was large (partial η² = .58). Post-hoc exploratory Pearson’s correlation analyses showed strong significant positive correlations between baseline wellbeing and post-intervention wellbeing; r(109) = .77, p < .001, and follow-up wellbeing; r(109) = .67, p < .001.

Posttraumatic growth

The mixed ANCOVA identified a significant main effect of group, with PTG being significantly higher in the intervention group (adjusted M = 67.49, SE = 2.73) compared to the control group (adjusted M = 57.85, SE = 2.81) after controlling for baseline wellbeing; F(1, 106) = 5.93, p = .02. The effect size was small (partial η² = .05). Figure 3 shows the estimated marginal means for PTG for the two groups at the post-intervention and follow-up timepoints. There was no significant interaction effect between time and group, indicating that both the control group (post-intervention adjusted M = 57.96, SE = 2.88; follow-up adjusted M = 57.74, SE = 2.85) and intervention group (post-intervention adjusted M = 66.83, SE = 2.80; follow-up adjusted M = 68.15, SE = 2.77) showed a comparable lack of change from post-intervention to follow-up; F(1, 106) = 0.97, p = .33. The effect size was small (partial η² = .01). There was no significant main effect of time (post-intervention adjusted M = 62.39, SE = 1.99; follow-up adjusted M = 62.95, SE = 1.97); F(1, 106) = 0.01, p = .94, and no interaction effect between time and baseline wellbeing; F(1, 106) = 0.001, p = .97. The effect sizes were small (partial η² < .001).

A significant main effect of baseline wellbeing was found, suggesting that the covariate was significantly related to posttraumatic growth at the post-intervention and follow-up timepoints; F(1, 106) = 10.71, p = .001. The effect size was medium (partial η² = .09). Post-hoc exploratory Pearson’s correlation analyses showed weak significant positive correlations between baseline wellbeing and post-intervention PTG; r(109) = .26, p = .01, and follow-up PTG; r (109) = .25, p = .01.

Association between photography engagement at one-month follow-up and outcome measure change scores

It was decided post-hoc to conduct an exploratory correlation analysis between one-month follow-up photography engagement and outcome measure scores to determine whether continued engagement with photography following the intervention was associated with improvement. Spearman’s rank-order correlations were run on the data to assess the relationship between photography engagement and the outcome measures at the one-month follow-up point. This test was selected as photography engagement was measured on a Likert scale and therefore yielded ordinal data. There was a statistically significant moderately strong positive correlation between photography engagement and wellbeing and photography engagement and PTG for the intervention group (r (56) = .33, p = .01 and r(56) = .32, p = .02 respectively), but not for the control group (r(53) = .05, p = .73, and r(53) = −.14, p = .33 respectively).

Discussion

This study aimed to explore the impact of a photography intervention on wellbeing and Posttraumatic Growth (PTG) for participants from the UK, recruited whilst national restrictions relating to the COVID-19 pandemic were beginning to ease. Participants who took photographs of things that made them feel positive and hopeful and wrote accompanying captions over a two-week period experienced significantly greater PTG and wellbeing following intervention completion compared to a control group when baseline wellbeing was controlled for, with small and medium effect sizes observed respectively. Outcome measure scores appeared to remain similar at one-month follow-up, although those with lower baseline levels of wellbeing were more likely to experience some further positive change in wellbeing from post-intervention to follow-up, whilst those with higher levels baseline wellbeing were more likely to experience some deterioration between these timepoints. Taken together, the results suggest that the photography intervention may hold therapeutic value, although those with lower wellbeing are somewhat more likely to continue experiencing improvements past the completion of the intervention. This is in line with other research suggesting that self-expressive interventions are more effective for those with poorer outcomes at baseline (Smith et al., 2015).

It must be noted that the current study is unable to provide a logic model for explaining why these positive changes occurred. Many factors unique to this period of the pandemic were not measured and could have contributed to improvements in wellbeing, such as increased leisure time, feelings of safety in lockdown, and the loss of daily commutes (Kowalski et al., 2022). However, given that a large proportion of the participants already regularly engaged with photography at baseline, the fact that outcomes changed following participation, and that positive associations between photography engagement and the outcome measures at one-month follow-up only emerged for the intervention group, does suggest that the intervention offered something over and above their usual photography practice. Written disclosure can facilitate cathartic inhibition of thoughts and feelings (Lepore & Smyth, 2002), sense-making around an experience (Pennebaker, 1993), and a stronger sense of being able to regulate one’s own emotions (Frattaroli, 2006; King, 2002). It is possible that the captioning aspect of the intervention afforded a therapeutic opportunity to reflect on photographs and disclose experiences of the pandemic. Whilst this study was a small-scale proof of concept study, future studies with larger sample sizes could explore potential mediators of therapeutic change, for example, investigating whether wellbeing mediated changes in PTG or vice versa, or stratifying participants based on the length of their photograph captions and examining the effect of writing quantity on outcomes.

When considering other self-administered writing interventions that have emerged during the pandemic, it has been demonstrated that writing about gratitude and imagining one’s best possible self following the pandemic have been associated with improved positive affect, social connectedness, and lower pandemic-related distress (Dennis et al., 2020; Jiang, 2020). In contrast, Vukčević Marković et al. (2020) found that participants experienced heightened levels of stress after engaging in an expressive writing exercise where they documented their experiences and thoughts about their lives during the pandemic. Whilst research has found that writing about a difficult or traumatic experience can be beneficial for psychological wellbeing (Barry & Singer, 2001), the authors acknowledged the possibility that COVID-19 news was so widely prevalent and difficult to avoid, that further space to reflect on one’s innermost feelings about the situation may have been distressing. Therefore, photographing and writing specifically about topics of positivity and hope may have contributed to improved outcomes in the current study.

In terms of limitations, the sample of participants in this study were predominantly women, from educated backgrounds, who already regularly engaged with photography. Given that this was a voluntary and reasonably time-intensive research project with no financial incentive, it is likely that participants were highly motivated to engage with the intervention. Therefore, inferences cannot be made about whether the photography intervention would be beneficial for the wider population. Whilst it was emphasised in recruitment posters that no prior skill with photography was required, it may have been beneficial to place the focus of advertisements more on wellbeing during COVID-19 rather than the photography intervention, so that the sample was less skewed to those interested in photography. The requirement for participants to be in possession of a smartphone and the use of social media for recruitment will have limited access for those who are digitally excluded, although it should be noted that around 90% of adults in the UK own a smartphone (Deloitte, 2019), making smartphones a widely accessible tool for engaging in photography. It is important to note that the participants who dropped out of the intervention were significantly more likely to be caregivers, which seems to indicate that the intervention was less acceptable to those with caring responsibilities. Jones et al. (2016) found that lack of time was reported as a barrier to caregivers of people with psychosis engaging in a written disclosure intervention, which may have applied in this research. It may be of value to explore how the time demands of photography can be reduced to improve accessibility, and trial this intervention using purposive sampling with more specified populations, such as those with mental health difficulties or caregivers. Across the wider sample, whilst the dropout rate of 25% was reasonably high (Furlan et al., 2009), the study took place during the pandemic which may well have contributed to more frequent difficulties in personal circumstances. It is therefore not clear if it was the intervention that was unacceptable or not, and more research exploring its acceptability is needed.

Furthermore, there were significant levels of non-completion of questionnaires at one-month follow-up. Whilst the Last Observation Carried Forward (LOCF) method was used to address missing data, no method of substitution or imputation can compare to actual data collection, and thus may lead to biased estimates. It would have been beneficial to recruit a larger sample at baseline so that the study was still sufficiently powered at follow-up even with high levels of dropout. Findings for the datasets with and without LOCF data could then have been compared. It would have also been of value to collect data over a longer follow-up period. Whilst this study was a pragmatic pilot study, the findings point to the importance of exploring the longevity of the therapeutic effects of photography in future research.

Finally, for the purpose of this pilot study, participants were not allocated to groups using stratified randomisation. Future definitive trials may benefit from stratifying participants based on certain characteristics, for example, by allocating participants to groups based on their baseline wellbeing (e.g. low, medium, and high). This would prevent baseline imbalances needing to be corrected for retrospectively. In this research, control group participants were not instructed to stop their usual photography practice, as the aim of the intervention was not to replace, but act as an adjunct to usual photography practices. Although participants were asked to rate their level of photography engagement over the past month, it would have been useful to collect more detailed information on the types of photography and frequency of practice, as well as engagement with other creative pursuits, to determine whether the groups were balanced on such variables, and if engagement affected outcome measure scores.

This study offers preliminary evidence that a positivity and hope-oriented photography intervention, conducted in the midst of the COVID-19 pandemic in the UK, can have beneficial effects on wellbeing and posttraumatic growth. This contributes to an emerging body of research highlighting the benefits of tools which aid self-expression and reflection on sources of hope found during the pandemic, and builds upon evidence of the therapeutic potential of photography.

Acknowledgments

The authors would like to thank the University of Surrey Department of Clinical Psychology Service User and Carers Advisory Group for their contribution to selecting outcome measures for their study, and providing feedback on the intervention and study materials.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that supports the findings of this study are available from the corresponding author, RR, upon reasonable request.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/17533015.2022.2107033

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.