Contributing factors to well-being in a sample of long-term survivors of childhood acute lymphoblastic leukemia: the role of social support in emotional regulation

Figures & data

Figure 1. Flowchart of participants. cALL, childhood acute lymphoblastic leukemia.

Note. As all questionnaires were missing for these survivors and the frequency of missing data was low, we decided not to impute missing data.

This flowchart shows that from 100 cALL survivors included in the present study, 8 were removed because of missing data on the variable of interest. Ninety-two were included in the preliminary and main analyses.

Figure 2. Frequencies of health function issues in 92 cALL survivors.

Note. cALL, childhood acute lymphoblastic leukemia. We used a count variable reflecting the number of health domains with at least a mild physical issue. A cut-point on the WHO-5 well-being index of > 50 for high-well being was used (Topp et al., 2015).

This graph shows the number of health domains with at least one mild physical issue in the global sample, the high well-being subgroup and the low well-being subgroup. Breathing, sleeping, discomfort, and vitality are the most affected domains, with higher frequencies reported by survivors.

Table 1. Sociodemographic and clinical characteristics of 92 cALL survivors.

	M (SD) or N (%)
	Total sample (n = 92)	High well-being^a (n = 67)	Low well-being^a (n = 25)	p
Sociodemographic characteristics
Sex				1.000
Female	47 (51.1)	34 (50.7)	13 (52.0)
Male	45 (48.9)	33 (49.3)	12 (48.0)
First language  French  English  Other	86 (93.5) 3 (3.3) 3 (3.3)	62 (92.5) 3 (4.5) 2 (3.0)	24 (96.0) 0 (0.0) 1 (4.0)
Age	24.41 (6.96)	24.76 (7.13)	23.48 (6.55)	.419
Ethnicity  Caucasian  Other	89 (96.7) 3 (3.3)	65 (97.0) 2 (3.0)	24 (96.0) 1 (4.0)	1.000
Marital status  Single  Married/common law  Missing	60 (65.2) 31 (33.7) 1 (1.1)	42 (62.7) 24 (35.8) 1 (1.5)	18 (72.0) 7 (28.0) 0 (0.0)	.621
Employment  Full time  Part time  Unemployed/unpaid	43 (46.7) 15 (16.3) 33 (35.9)	33 (49.3) 10 (14.9) 24 (35.8)	10 (40.0) 5 (20.0) 9 (36.0)	.740
Missing	1 (1.1)	0 (0.0)	1 (4.0)
Clinical characteristics
Age at diagnosis	7.78 (5.16)	8.02 (5.34)	7.16 (4.72)	.461
Age at the end of treatment	9.90 (5.18)	10.15 (5.34)	9.24 (4.75)	.434
ALL risk status				.603
Standard risk	25 (27.2)	17 (25.4)	8 (32.0)
High risk	66 (71.7)	49 (73.1)	17 (68.0)
Missing	1 (1.1)	1 (1.5)	0 (0.0)
Treatment protocol				.569
87–01  91–01	8 (8.7) 17 (18.5)	4 (6.0) 13 (19.4)	4 (16.0) 4 (16.0)
95–01	22 (23.9)	16 (23.9)	6 (24.0)
2000–01  2005–01  Other Radiotherapy  Yes	28 (30.4) 13 (14.1) 4 (4.3) 69 (75.0)	21 (31.3) 9 (13.4) 4 (6.0) 51 (76.1)	7 (28.0) 4 (16.0) 0 (0.0) 18 (72.0)	.788
No	23 (25.0)	16 (23.9)	7 (28.0)

Note. cALL, childhood acute lymphoblastic leukemia.^aCut-point on the WHO-5 well-being index of >50 (Topp et al., 2015).

Table 2. Health status and psychosocial description of the study sample.

M (SD) or N (%)
	Total sample (n = 92)	High well-being^a (n = 67)	Low well-being^a (n = 25)	p
Health status (15D)^b	2.32 (1.94)	1.99 (1.69)	3.20 (2.31)	.022
Health status domains^c
Mobility	6 (6.5)	5 (7.5)	1 (4.0)	1.000
Vision	16 (17.4)	12 (17.9)	4 (16.0)	1.000
Hearing	11 (12.0)	5 (7.5)	6 (24.0)	.064
Breathing	34 (37.0)	25 (37.3)	9 (36.0)	1.000
Sleeping	47 (51.1)	30 (44.8)	17 (68.0)	.062
Eating	0 (0.0)	0 (0.0)	0 (0.0)	-
Speech	15 (16.3)	6 (9.0)	9 (36.0)	.004
Excretion	12 (13.0)	8 (11.9)	4 (16.0)	.729
Discomfort	37 (40.2)	24 (35.8)	13 (52.0)	.232
Vitality	35 (38.0)	18 (26.9)	17 (68.0)	<.001
Well-being (WHO-5)	61.78 (22.70)	73.43 (12.57)	30.56 (11.19)	<.001
Psychosocial characteristics
Social support (N, SSQ-6)	20.03 (8.48)	21.16 (7.91)	17.00 (9.36)	.035
Cognitive reappraisal (ERQ)	4.79 (1.19)	4.97 (1.15)	4.31 (1.17)	.017
Expressive suppression (ERQ)	3.31 (1.33)	3.22 (1.41)	3.54 (1.11)	.309
Emotional processing (EAC)	2.82 (.79)	2.94 (.79)	2.49 (.68)	.015
Emotional expression (EAC)	2.67 (.74)	2.74 (.78)	2.47 (.57)	.116

Note. WHO-5, World Health Organization Well-being Index, SSQ-6, Social Support Questionnaire-Short Form, ERQ, Emotion Regulation Questionnaire, EAC, Emotional Approach Coping Scales.

^a Cut-point on the WHO-5 well-being index of >50 (Topp et al., 2015).

^b Number of domains showing at least a mild severity level.

^c Numbers refer to frequencies of participants with issues on this health domain. Domains are those common to the 15D (n = 63), 16D (n = 28), and 17D (n = 1) excluding for the items of depression and distress (see Methods).

Table 3. Summary of logistic regression moderation analysis predicting the odds of belonging to the high well-being subgroup depending on emotional processing controlling for health status (n = 92).

	Bootstrap Mean Coefficient^a	Odds Ratio	SE^b	OR 95% CI^b
Health status score (15D)^c	-.65	.52	1.36	.27–.91
Emotional processing (EAC)	.75	2.12	1.51	1.09–5.37
Social support (N, SSQ-6)	.53	1.70	1.49	.88–4.26
Interaction Social support x Emotional processing	.34	1.40	1.58	.61–3.82

Note. EAC, Emotional Approach Coping Scales, SSQ-6, Social Support Questionnaire-Short Form.

Odds ratios, SE, and 95% CI were calculated using the formula OR  = e ^{bootstrap mean coefficient}, e^SE, and e^{OR 95% CI} (Hayes, 2020) to produce effects on an odds ratio metric.

R² = .15 (Cox & Snell) .22 (Nagelkerke). Model χ² (4) = 15.47 p = .004.

^a Expressed in log-odds metric.

^b 95% percentile bootstrap confidence intervals and standard errors were based on 5000 resamples.

^c Health status was calculated by adding the number of physical difficulties common to the 15D, 16D, and 17D.

Table 4. Summary of logistic regression moderation analysis predicting the odds of high well-being by expressive suppression, social support, and their interaction (n = 92).

	Bootstrap Mean Coefficient^a	Odds Ratio	SE^b	OR 95% CI^b
Health status score (15D)^c	-.65	.52	1.36	.27–.92
Expressive suppression (ERQ)	-.46	.63	1.43	.28–1.15
Social support (N, SSQ-6)	.74	2.10	1.49	1.09–5.31
Interaction Expressive suppression x Social support	-.91	.40	1.58	.13–.79

Note. ERQ, Emotion Regulation Questionnaire, SSQ-6, Social Support Questionnaire-Short Form.

Odds ratios, SE, and 95% CI were calculated using the formula OR = e ^{bootstrap mean coefficient}, e^SE, and e^{OR 95% CI} (Hayes, 2020).

R² = .17 (Cox & Snell) .25 (Nagelkerke). Model χ² (4) = 17.34 p = .002.

^a Expressed in log-odds metric.

^b 95% percentile bootstrap confidence intervals and standard errors were based on 5000 resamples.

^c Health status was calculated by adding the number of physical difficulties common to the 15D, 16D, and 17D.

Figure 3. Social support moderates the contribution of expressive suppression to well-being in a group of 92 cALL survivors.

Note. ERQ, Emotion Regulation Questionnaire. Models are adjusted for health status; As per the PROCESS output, Low social support = 11 people, Medium social support = 19 people, High social support = 30 people.

This graph shows how social support moderates the contribution of expressive suppression in the sample. When expressive suppression is low, survivors with high social support show considerable higher odds to belong to the high-well being subgroup than survivors with medium or low social support. When expressive suppression is high, however, odds to belong to the high well-being subgroup are the same regardless the level of social support.

Figure 4. Number of health function issues reported by 92 cALL survivors on the 15D.

Note. cALL, childhood acute lymphoblastic leukemia. A cut-point on the WHO-5 well-being index of > 50 for high-well being was used (Topp et al., 2015).

This graph shows the number of health function issues reported by survivors in the high and the low well-being subgroups. In both subgroups, most survivors report 1 or 2 health issues.

Topp, C. W., Østergaard, S. D., Søndergaard, S., & Bech, P. (2015). The WHO-5 Well-Being Index: a systematic review of the literature. Psychotherapy and Psychosomatics, 84(3), 167–176. https://doi.org/10.1159/000376585

 PubMed Web of Science ®Google Scholar

Hayes, A. F. (2020). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford Press.

 Google Scholar

Data availability statement

The full dataset is available in Supplementary material attached to this article.