An explorative analysis of the differences in levels of happiness between cancer patients, informal caregivers and the general population

A pooled analysis of two cross-sectional samples: [1] healthy social media users who completed surveys using electronic tools and [2] cancer patients and informal caregivers of cancer patients who answered face-to-face questionnaires.

Individuals who have a Facebook® account and/or used the WhatsApp® application and caregivers of cancer patients and cancer patients treated at a cancer hospital.

Participants from the five regions of Brazil were recruited online through the social network site Facebook® and the WhatsApp® application. In brief, three different strategies were used for the Facebook® survey: [1] From a total of 5570 Brazilian cities, 300 were randomly selected. One of the researchers (MCG) searched for Facebook® pages that could be representative of the city (preferentially official prefecture pages) and requested permission from the owners of those pages to post the link to the survey [2]; three university professors from the northeastern, northern and southern regions of Brazil were invited to post the research link on their personal Facebook® pages [3]; the authors posted the link on their personal pages. Additionally, people from the authors’ contact list networks were also contacted via the Whatsapp® application to disseminate and replicate the research so that the invitation to participate would reach as many Brazilian individuals as possible. In this case, explanatory text about the study was sent with the research link. In addition to being asked to respond to the questionnaire online (using either Facebook® or WhatsApp®), the participants were encouraged to share the study text/link with their own Facebook® and WhatsApp® contacts. The survey was completed using the SurveyMonkey® platform. Unfortunately, regardless of the participation route (Facebook or WhatsApp), all responses completed online were sent to the SurveyMonkey program data collector (the program through which the responses were collected and stored). Because the collector was the same for all methodologies, it was not possible to accurately identify the most effective search strategy.

A convenience sample of cancer patients and informal caregivers was interviewed in person using the same evaluation forms that were used for the online survey. Three trained interviewers (two research nurses and one medical student) collected all the data. Patients were recruited from oncology outpatient clinics, and caregivers were recruited from two different institutional support houses where cancer patients and informal caregivers from other locations stay while being treated in the city of Barretos.

Individuals who met all the inclusion criteria and none of the exclusion criteria were included.

The sample size was calculated using the Power Analysis and Sample Size application (PASS v. 2002) via multiple linear regression analysis with seven predictors and a priori coefficients of determination (effect size) of R² = 0.01, α = 0.05 and β = 0.05, thus resulting in an a priori statistical power of 95%. The estimated sample size was 2191 healthy participants.

The participants in this study (general population, cancer patients and informal caregivers) were grouped so that the sample size calculation for each group was performed separately. However, one yet-unpublished study separately examined patients and caregivers. The sample size calculated in that study aimed to compare happiness levels according to different subgroups of patients: [1] survivors (no evidence of disease and no antineoplastic treatment) [2]; patients receiving adjuvant therapy (those with no evidence of disease but receiving antineoplastic treatment); and [3] patients receiving palliative care only (those with evidence of disease and no antineoplastic treatment). Thus, assuming a rate of 70 and 50% of happy individuals in groups 1 and 3, α = 0.05, β = 0.10, at least 100 patients for each group should be included. Considering that caregivers would mainly be caring for patients whose profiles were similar to those of group 3, initial planning expected that caregivers and patients from group 3 would be compared. Thus, the minimum number of caregivers was also defined as 100.

The final sample comprised a minimum of 2591 participants.

Data were collected from October 2015 to October 2016. A total of 134 caregivers of cancer patients were invited to participate in the study. Eight refused to participate, and 126 were included. SurveyMonkey® (via Facebook and WhatsApp) collected data from 2688 respondents, but 90 were under 18 years old, and 13 declined, remaining 2585. Of these, 433 did not complete the Pemberton Happiness Index (PHI) and Satisfaction with Life Scale (SWLS), and 40 reported having cancer at the time of survey completion; thus, a total of 2112 responses were analyzed using the online tool. A total of 328 cancer patients were invited to participate in the study; 26 refused to participate, and the other 302 were included. Forty online respondents who reported having cancer at the time of survey completion were added to these, for a total of 342 cancer patients. The final sample included 2580 participants (2112 online respondents from the general population, 342 cancer patients and 126 caregivers).

The study followed Resolution CNS 466/12, which regulates Brazilian ethical research issues. In Brazil, there are several accredited local Research Ethics Committees and a National Research Ethics Commission (CONEP). There are well-established criteria that define the need for additional approval by CONEP; the present study was approved by the Research Ethics Committee of the Barretos Cancer Hospital (opinion no. 1.098.789 and 1.114.730) and did not need to be sent to CONEP for further approval. All participants read and signed the informed consent form (ICF) before voluntarily agreeing to participate via either the online or the face-to-face formats.

Additional analyses were conducted to compare individual item scores on the PHI between groups of participants by means of the Kruskal-Wallis test. Scores for the items that evaluate “personal growth” and “meaning in life”, as well as an item that measures “the ability to enjoy small things in daily life”, were chosen. Additionally, an item measuring “bad moments in daily life” was chosen to validate the hypothesis of greater perceived happiness among those facing cancer (informal caregivers and cancer patients) despite reporting more suffering/negative affect in daily life. The main contents of and item descriptions for the selected PHI items were as follows: eudaimonic well-being/life meaning (“I think my life is useful and worthwhile”); eudaimonic well-being/personal growth (“My life is full of learning experiences and challenges that make me grow”); hedonic well-being/positive affect (I enjoy many little things every day); and hedonic well-being/negative affect (“I have many bad moments in my daily life”).

In the multivariate model, the informal caregivers (β = 0.8, p ≤ 0.001) and cancer patients (β = 0.5, p ≤ 0.001) presented higher rates of happiness in relation to the general population. A positive self-assessment of health (β = 0.6, p ≤ 0.001) and reports of optimism (β = 1.4, p = < 0.001) were associated with increased happiness scores. Regarding aspects of the participants’ daily life, a higher frequency of family gatherings (β = 0.4, p ≤ 0.001), contact with nature (β = 0.2, p = 0.044), and leisure time (β = 0.3, p ≤ 0.001), as well as physical activity ≥3 times per week (β = 0.3, p ≤ 0.001), were associated with higher PHI scores in the multivariate model. In contrast, previous diagnoses of depression (β = − 0.7, p ≤ 0.001), anxiety (β = − 0.3, p ≤ 0.001) and other psychiatric/psychological problems (β = − 0.7, p = 0.002) were negatively associated with happiness scores (Table 2).

Regarding life satisfaction, the informal caregivers (β = 2.9, p ≤ 0.001) and cancer patients (β = 0.9, p = 0.051) had higher scores than the general population. In the multivariate model, greater life satisfaction was associated with positive self-assessment of health (β = 2.3, p ≤ 0.001), optimism (β = 3.4, p ≤ 0.001), happiness with work (β = 2.4, p ≤ 0.001) and satisfaction with financial issues (β = 3.4, p ≤ 0.001). Being separated or divorced (β = − 2.0, p ≤ 0.001) and having a previous diagnosis of depression (β = − 2.5, p ≤ 0.001) or anxiety (β = − 0.9, p = 0.002) were negatively associated with life satisfaction (Table 3).

Higher positive affect scores were associated reports of optimism (β = 5.2, p ≤ 0.001), positive self-assessments of health (β = 1.2, p = 0.001), happiness with work (β = 1.5, p ≤ 0.001), satisfaction with financial issues (β = 1.9, p ≤ 0.001) and leisure time (β = 1.7, p < 0.001). Lower positive affect levels were noted in cancer patients (β = − 3.8, p ≤ 0.001) and informal caregivers (β = − 2.0, p ≤ 0.001) than in the general population. Previous diagnoses of depression (β = − 2.1, p ≤ 0.001), anxiety (β = − 0.9, p ≤ 0.001), and some other psychiatric/psychological problems (β = − 1.4, p = 0.027) and not having a job (β = − 1.0, p = 0.007) were also associated with lower positive affect levels (Table 4).

Subsequently, higher levels of negative affect were observed in the informal caregivers (β = 3.6, p ≤ 0.001) and cancer patients (β = 2.3, p ≤ 0.001) than in the general population. Individuals with a previous diagnosis of depression (β = 2.2, p ≤ 0.001) and anxiety (β = 1.5, p ≤ 0.001) also exhibited higher negative affect levels. In contrast, a positive self-assessment of health (β = − 0.7, p = 0.005), optimism (β = − 3.5, p ≤ 0.001), happiness with work (β = − 1.3, p ≤ 0.001), satisfaction with financial issues (β = − 0.9, p ≤ 0.001) and not having a sick loved one (β = − 0.5, p = 0.005) were associated with lower negative affect levels (Table 5).

The median and interquartile range (IQR) values of the PHI items were compared among the cancer patients, informal caregivers and healthy people: the “life meaning” item (cancer patients = 10(1); caregivers = 10(1); general population = 9(3); p < 0.001); the “personal growth” item (cancer patients = 10(2); caregivers = 10(2); general population = 9(4); p < 0.001); the “enjoy little things” item (cancer patients = 10(2); caregivers = 9(3); general population = 7(4); p < 0.001); and the “bad moments” item (cancer patients = 5(6); caregivers = 6(5); general population = 3(5); p < 0.001).

The feeling of happiness is subject to a wide range of external influences; however, hereditary components are present that can be explained by the genetic architecture of the personality and are responsible for 30 to 50% of the score variance [4, 28, 29]. The “genetics of happiness” and personality traits maintain relatively constant happiness scores over time. However, engaging in activities that promote increased happiness may also potentially improve positive feeling levels for significant periods of time [4]. In this study, the final linear regression models explained between 54 and 65% of the variance in the participants’ scores for happiness, life satisfaction and positive and negative affect, suggesting that the remaining percentage of variance may be explained by genetic or personality-specific aspects that were not investigated.

In the present study, the characteristics associated with higher happiness scores enhance the multidimensional nature of the “happiness” construct. Some categories of factors that tend to positively influence the general population’s sense of happiness have been described in prior studies; these factors include optimism, the pursuit of goals and purposes, acts of kindness, expressions of gratitude, financial satisfaction, physical exercise, social relationships, rest, and altruism [30‐32]. The present study was the first to detail happiness-related conditions in a large sample of the Brazilian population using properly validated tools.

For an adequate understanding of the results of this study, it should be emphasized that the scales used tend to measure slightly different constructs. The PNES uses more direct questions and investigates more transient life aspects. The SWLS and PHI-r include more reflective questions addressing existential aspects. Examples of such items are as follows: from the SWLS, “As much as possible, I have achieved the important things I want out of life”; from the PHI, “My life is full of learning experiences and challenges that make me grow,” and “I enjoy many little things every day”. These differences explain why patients and informal caregivers have lower positive and higher negative affect and simultaneously have better happiness and life satisfaction scores, which at first glance would appear incoherent. To validate this finding, a different analysis was performed; it determined that the scores for items that evaluate “personal growth” and “meaning in life” and for the item that measures “ability to enjoy small things in daily life” were higher among the patients and caregivers than among the healthy people. Again, both the patients and the caregivers reported higher scores on the “bad moments in daily life” item than the healthy people did.

Patients with cancer and cancer survivors who are happy tend to have a better quality of life and fewer symptoms than unhappy patients do. In addition, happier cancer patients are more likely to report positive expectations about the future; have more life goals; have greater optimism and hope; experience and positive changes within their relationships; are better at coping with problems; and have higher spirituality levels [18, 21, 33]. Acceptance of the cancer diagnosis and the “path” through treatment are often seen as psychosocial and spiritual transitions that can lead to a restructuring of values and can reflect how patients assess life and their sense of happiness [20, 21]. Thus, it is important to evaluate happiness in cancer patients and to seek methods to promote it, i.e., through providing support during treatment or encouraging the development of positive characteristics unrelated to the disease or its treatment. The role of the health professional should not be limited exclusively to issues related to the disease itself.

The psycho-oncological literature shows characteristics of posttraumatic growth in patients during and after treatment for cancer. This process refers to the positive psychological changes experienced as a result of struggling with highly challenging life circumstances [18]. The changes occur mainly in interpersonal relationships, perceptions of personal strength, appreciation of life, spirituality and values ​​priorities [18]. In the present study, it was possible to identify some aspects that reflect the posttraumatic growth of the participants who had cancer and the informal caregivers of patients with cancer. These groups even had higher happiness scores than the healthy people in the general population did.

The well-being of informal caregivers of cancer patients is also affected by the cancer diagnosis [10, 11]. Caring for suffering patients can cause a significant negative emotional load for their caregivers in addition to causing fatigue, physical exhaustion, etc. Despite these negative parameters, caregivers of cancer patients may present high well-being levels that are influenced by characteristics and events inherent to their own life and the lives of those for whom they care [17].

According to a Buddhist aphorism, “we all want to be happy and not suffer”. However, the results of this study suggest that suffering (either experiencing it oneself and/or observing it in someone close), although undesirable, has a role in individual growth in healthy individuals and individuals with chronic diseases (such as cancer) [15, 17, 20, 21]. Cancer patients exclusively receiving palliative care had reduced perceptions of happiness and life satisfaction compared with the other groups of cancer patients, as expected. In these patients, physical (pain, fatigue, nausea, dyspnea, among others) and psychosocial suffering are high, especially at the end of life. After adequate control of physical and psychological symptoms, strategies that focus on problem solving, personal issues, forgiveness and dignity therapy may play a fundamental role in the existential domain. If managed properly, these patients will probably report higher happiness and life satisfaction scores. However, this hypothesis needs to be tested in future studies.

A comparative study [34] of students from Brazil and Norway evaluated the use and efficiency of 14 strategies for regulating emotion with the aim of stopping anger, anxiety and sadness. Norway and Brazil were chosen because they were thought to represent individualistic and collectivistic cultures, respectively. In general, the Norwegians used a greater number of strategies than the Brazilians. Cultural differences were suggested by comparatively high user ratings of “working”, “acting out”, and “going for a walk” by Norwegians and “relaxation” and “entertainment” by Brazilians. The “pray to God” strategy was also more frequently used by Brazilians than by Norwegians. Thus, it is suggested that people from collectivist countries (such as Brazil) seek strategies for regulating emotions that foster individualism and self-knowledge. In the present study, at the time of emotional crisis (cancer diagnosis), the Brazilian participants presented higher scores for happiness and satisfaction with life, which may suggest that the pursuit of self-knowledge and individualization could have led to personal growth. However, subsequent studies need to be designed to explain these findings.

This study has some limitations. For example, the participating patients’ caregivers were not evaluated/included, and there was no information about the conditions of the patients whom the participating caregivers were accompanying; however, it was assumed that they were patients with reduced functionality given that many patients do not need caregivers when they are in good physical condition. Although the healthy participants were selected from the general population via social media, they probably do not characterize and represent the entire Brazilian population.

In contrast, the study is original because no other study has been conducted in Brazil that evaluates the constructs studied in the different groups addressed and because happiness is rarely measured in cancer patients. In addition, the study allowed the comparison of a large population of the same nationality and from different regions.

The pooled analysis of the general population and people affected by cancer, whether as patients or caregivers, has its limitations. Although statistical adjustment techniques were used, the combined populations have different profiles, which can be considered a potential selection bias. Our sample is consistent with the most recent Brazilian population census, which showed that the most populous region is the Southeast and that the largest number of white Brazilians live in urban areas and are Catholics; on the other hand, our study included a higher percentage of women and had a higher income and educational level than the general population described in the census. The modes of administration (on-line and face-to-face) also represent a study limitation. However, the authors were careful to use the same research instruments for data collection in both modes and to properly train the interviewers to standardize data collection. In addition, previous comparative studies [35, 36] have shown small or no magnitude of influence on the responses to the questionnaires.