EHealth intervention for quality of life in long-term breast cancer survivors: randomized controlled trial

Abstract

Background

Digital health interventions have shown promise for improving quality of life, especially in the short term after treatment. However, evidence regarding long-term breast cancer survivors (LT-BCSs; disease free >5 years) remains limited. This study aimed to evaluate the effectiveness of an eHealth intervention targeting health promotion and late sequelae management to improve LT-BCSs’ quality of life.

Methods

In a randomized controlled clinical trial, 201 LT-BCSs (mean = 11 years posttreatment) were randomly assigned to an intervention group (n = 102) using the CUMACA-M mobile application with specific health advice and recommendations for LT-BCSs, or a control group (n = 99) receiving usual care. Quality of life was measured with the Quality of Life–Cancer Survivors (QOL-CS) scale at baseline and after 3 months. Analyses followed an intention-to-treat approach using t tests, nonparametric tests, and effect sizes.

Results

At 3 months, no statistically or clinically significant differences between the groups in the overall quality of life score (QOL-CS) were found (difference-in-differences = 0.11; 95% confidence interval = −0.10 to 0.32; P = .303). In the intervention group, a small intragroup decrease in spiritual well-being was observed (−0.25; −0.49 to −0.02), of uncertain clinical significance; no significant differences between groups were detected.

Conclusions

This eHealth intervention did not improve the quality of life of LT-BCSs, suggesting that more personalized, interactive, or professionally supported strategies may be needed. Future research should evaluate the long-term outcomes and effectiveness of hybrid or personalized digital strategies in this population.

Trial Registration

ClinicalTrials.gov NCT05322460; https://clinicaltrials.gov/study/NCT05322460.

Introduction

Advances in early detection and treatment have significantly increased breast cancer survival rates,^1-3 which has led to an increase in the population of long-term survivors.^4-7 Globally, the 5-year relative survival of patients with breast cancer is estimated to be approximately 82%.⁶ However, many of these people experience persistent physical, emotional, and psychosocial sequelae that negatively affect their quality of life (QoL), even years after completing treatment.^8-11

Although cancer survivors’ needs have been increasingly recognized, access to personalized and ongoing support remains limited in many health systems.⁶ These systems are not always designed to offer long-term, sustained interventions that comprehensively address the physical, emotional, and social sequelae that persist after the end of active treatment.^9,10,12 As a result, long-term cancer survivors (>5 years disease-free) often receive fragmented and opportunistic care.¹³ The lack of structured posttreatment care models and dedicated resources not only hinders the effective management of survivors’ emerging needs but also increases the healthcare costs stemming from their substantial demand for services.¹⁴

Digital health interventions (eHealth) have become promising tools to address these gaps, offering scalable, flexible, and personalized support to cancer survivors.^15-19 Although previous studies have shown the potential of eHealth to improve QoL outcomes, evidence of its effectiveness in long-term breast cancer survivors (LT-BCSs) remains limited and inconclusive.²⁰

Thus, this study aimed to evaluate the effectiveness of an eHealth intervention targeting health promotion and late sequelae management to improve LT-BCSs’ QoL.

Methods

Study design

A randomized controlled trial (RCT) was conducted. The trial was prospectively registered (ClinicalTrials.gov # NCT05322460) and the trial protocol is published elsewhere.²¹ The methodological framework of the Medical Research Council (MRC) for the development and evaluation of complex interventions was followed.²² This RCT evaluated the effectiveness of a digital health intervention in mobile application format, called CUMACA-M (an acronym in Spanish for Care Beyond Cancer-Breast), to improve the QoL of breast cancer survivors in the long term.²¹ Participants were randomly assigned to either the intervention group (IG), which had access to the CUMACA-M mobile application, or the control group (CG), which received usual care without structured survivorship support. The evaluations were carried out at baseline (T0) and 3 months after the intervention (T1).

Participants

Eligible participants were women older than 18 years with a history of breast cancer, disease-free for >5 years after completing primary treatments (surgery, chemotherapy, immunotherapy, and/or radiotherapy). Ongoing hormonal therapy, common among LT-BCSs, was permitted as it is an adjuvant treatment. Additional criteria included being familiar with the internet and owning a smartphone.

Recruitment

Between November 2023 and January 2025, participants were recruited through a multi-channel strategy. On one hand, the collaboration of the Navarra Breast Cancer Association (Saray) facilitated the dissemination of the project among its members and contacts. Additionally, partnerships were established with various health centers in the region within a 100 km radius to identify and invite potential participants during follow-up consultations. To broaden the scope, outreach efforts included local radio and television interviews, as well as press releases in regional newspapers, aimed at raising awareness and encouraging participation.

The research team presented the study to participants and provided a detailed information booklet. All participants gave written informed consent before receiving a link to the baseline evaluation, which included the collection of sociodemographic and clinical data.

This research complied with the ethical principles established in the Declaration of Helsinki and was approved by the Ethics, Animal Experimentation and Biosafety Committee of the Public University of Navarre (PI-2021/18). The study complied with institutional and international ethical standards, ensuring strict confidentiality. Personal identifiers were removed and replaced with codes, and all data were stored in secure, password-protected systems accessible only to authorized researchers. Information was used solely for research purposes, reported in aggregate, and never disclosed without consent. Participants could withdraw at any time without penalty. Finally, all activities followed applicable data protection regulations.

Intervention: description of the CUMACA-M mobile application

The CUMACA-M mobile application was developed as a supportive resource for LT-BCSs, aiming to improve QoL through self-management, psychoeducation strategies, and personalized health recommendations. The app delivers evidence-based content across 3 modules, namely, LT-BCS-specific health recommendations, physical exercise, and nutrition, along with interactive resources. Personalization is achieved through tailored advice on topics such as healthy lifestyles and emotional well-being, guided by needs identified in baseline self-reported screening questionnaires (https://apps.apple.com/es/app/cumaca-m/id6474200830). The IG participants received full access to the CUMACA-M application and were encouraged to use it freely, at their own pace, for a period of 3 months.

Outcomes

The primary outcome was QoL, which was assessed using the Quality of Life-Cancer Survivors (QOL-CS) scale, which was originally developed by Ferrell et al.²³ For this study, the validated Spanish version of the instrument was employed. The QOL-CS is a multidimensional tool specifically designed for cancer survivors. It consists of 41 items and includes physical, psychological, social, and spiritual well-being domains. Each item is rated on a 10-point Likert scale, with 1 indicating extremely poor QoL and 10 representing excellent QoL. The outcomes were self-reported and collected at baseline (T0) and at 3 months postbaseline (T1).

Sample size

The sample size was estimated based on a standardized mean difference of 0.75 in QoL scores, with a standard deviation of 1.8, a 2-sided type I error of 5%, and a power of 80% (type II error of 20%). Under these parameters, 91 participants per group (intervention and control) were required to detect a statistically significant difference. To account for an anticipated 10% attrition rate, the total target sample was increased to 200 participants (100 per group).

Random assignment and blinding

Participants were randomly assigned (1:1) to the IG or CG using a computer-generated random number list created through the Research Randomizer platform.²⁴ This randomization was performed by an independent researcher who had no contact with participants and no access to personal or clinical information. Randomization was concealed, with group assignments revealed only at the time of allocation. An independent researcher, blinded to participants’ identities and baseline characteristics, communicated group assignments individually.

Statistical analysis

Analyses were conducted using R version 4.5.0²⁵ via RStudio.²⁶ A 2-sided significance level of .05 was considered for all inferential tests. The analyses followed an intention-to-treat (ITT) approach and focused on between-group comparisons (intervention vs control) on the basis of pre-post change scores (T1-T0).

For 14 participants with data only at baseline (T0), missing values were imputed and included in all analyses to preserve sample size and minimize potential bias. Descriptive statistics were used to summarize sociodemographic and clinical data, with continuous variables reported as means and standard deviations and categorical variables as frequencies and percentages. Normality and homogeneity of variance were assessed using the Kolmogorov-Smirnov test with Lilliefors correction and Levene test, respectively. Between-group differences at baseline were evaluated using t tests for continuous variables and χ² tests or Fisher exact tests for categorical variables, as appropriate. The Mann-Whitney U test was used for nonnormally distributed variables.

For outcomes with a normal distribution (QOL-CS and Psychological Well-Being Subscale), within-group changes (T1-T0) were reported as means and 95% confidence intervals (CIs), estimated via paired t tests. Independent-samples t tests (equal variances assumed) were used to compare change scores between groups. Difference-in-differences estimates with 95% confidence intervals were reported, and effect sizes were calculated using Cohen d (<0.20 = very small; 0.20-0.49 = small; 0.50-0.79 = moderate; and ≥0.80 = large).²⁷

For outcomes violating normality assumptions (subscales of physical, social, and spiritual well-being), a nonparametric bootstrap-based approach was applied. Within-group means, standard deviations, and change scores with 95% bootstrap confidence intervals are reported. Between-group comparisons were performed using the Wilcoxon rank-sum test, and effect sizes were estimated using the rank-biserial correlation (r), interpreted as <0.10 (very small), 0.10 to 0.29 (small), 0.30 to 0.49 (moderate), and ≥0.50 (large).²⁷

Results

Sample characteristics

A total of 201 LT-BCSs participated, of which 99 were assigned to the CG and 102 to the IG (Figure 1). The baseline sociodemographic and clinical characteristics were homogeneous between the groups (Table 1). The mean age was 57.90 years (SD = 9.26) in the CG and 57.28 years (SD = 8.70) in the IG. Most of the participants were married or in a relationship, had completed higher education or were professionals, and were employed or retired.

Figure 1.

CONSORT diagram of participant flow. Source: https://clinicaltrials.gov/study/NCT05322460

Table 1.

Characteristics of participants at baseline.

Variable	CG (n = 99)	IG (n = 102)	P
Age, mean (SD)	57.90 (9.26)	57.28 (8.70)	.628
Marital status
Married/partner	86 (86.9)	75 (73.5)	.104
Divorced/separated	3 (3.0)	8 (7.8)
Single	8 (8.1)	13 (12.7)
Widowed	2 (2.0)	6 (5.9)
Education level
Baccalaureate/vocational training	41 (41.4)	34 (33.3)	.318
Primary	16 (16.2)	12 (11.8)
Secondary	4 (4.0)	4 (3.9)
No studies	1 (1.0)	0 (0.0)
University	37 (37.4)	52 (51.0)
Employment status
Active	56 (56.6)	61 (59.8)	.941
Unemployed	5 (5.1)	6 (5.9)
Temporary disability	5 (5.1)	3 (2.9)
Permanent disability	5 (5.1)	5 (4.9)
Retirement	28 (28.3)	27 (26.5)
Years since diagnosis, mean (SD)	12.72 (6.17)	12.10 (5.03)	.436
Years since completion of primary treatment, mean (SD)	11.51 (6.20)	11.02 (5.08)	.544
Stage of cancer
Stage 0	1 (1.0)	0 (0.0)	.305
Stage I	17 (17.2)	21 (20.6)
Stage II	16 (16.2)	21 (20.6)
Stage III	16 (16.2)	9 (8.8)
Stage IV	0 (0.0)	2 (2.0)
I don’t remember	49 (49.5)	49 (48.0)
Hereditary component: Yes	10 (10.1)	12 (11.8)	.879
Surgery type
Lumpectomy tumor removal	39 (39.4)	40 (39.2)	.951
Partial mastectomy	24 (24.2)	27 (26.5)
Total mastectomy	33 (33.3)	33 (32.4)
I don't remember	3 (3.0)	2 (2.0)
Lymph nodes removed from the armpit
1 node	15 (15.2)	29 (28.4)	.053
More than 1 node	62 (62.6)	58 (56.9)
None	22 (22.2)	15 (14.7)
Type of treatment
Brachytherapy	1 (1.0)	1 (1.0)	.606
None	2 (2.0)	3 (2.9)
Chemotherapy	9 (9.1)	12 (11.8)
Chemotherapy + Radiotherapy	52 (52.5)	43 (42.2)
Chemotherapy + Radiotherapy + Immunotherapy	5 (5.1)	11 (10.8)
Chemotherapy + Radiotherapy + Brachytherapy	6 (6.1)	4 (3.9)
Chemotherapy + Radiotherapy + Brachytherapy + Immunotherapy	3 (3.0)	3 (2.9)
Radiotherapy	20 (20.2)	23 (22.5)
Radiotherapy + Immunotherapy	0 (0.0)	2 (2.0)
Radiotherapy + Brachytherapy	1 (1.0)	0 (0.0)
Hormonal therapy: Yes	72 (72.7)	80 (78.4)	.437
Cancer recurrence
No	95 (96.0)	98 (96.1)	.564
1	4 (4.0)	3 (2.9)
2	0 (0.0)	1 (1.0)
No. of pregnancies, mean (SD)	1.62 (1.21)	1.61 (1.16)	.960
No. of births, mean (SD)	1.38 (0.96)	1.43 (0.96)	.725
Early menopause: Yes	61 (61.6)	61 (59.8)	.906
Fertility problems: Yes	21 (21.2)	15 (14.7)	.308
Other treatments: Yes	27 (27.3)	28 (27.5)	1.000
Other pathologies: Yes	33 (33.3)	39 (38.2)	.564
Lymphedema: Yes	19 (19.2)	20 (19.6)	1.000
Tobacco consumption: Yes	11 (11.1)	5 (4.9)	.172
Alcohol consumption: Yes, occasionally	65 (65.7)	62 (60.8)	.569

Data are expressed as No. (%) unless otherwise indicated.

Abbreviations: CG = control group; IG = intervention group.

On average, participants were diagnosed over 12 years prior, with primary treatment completed about 11 years earlier in both groups. Chemotherapy and radiation therapy were the most common treatments, and more than 70% of the participants had received hormonal therapy. No statistically significant differences were found between the groups at the beginning of the study in terms of the sociodemographic, clinical, or treatment-related variables (all Ps > .05).

Primary outcome

The main outcome was QoL, measured by the total score on the QOL-CS scale. No statistically significant intragroup changes from baseline to 3 months were observed in either group (Table 2). In the CG, the mean score of the QOL-CS scale decreased slightly from 5.89 (SD = 1.25) to 5.79 (SD = 1.29), which represents an estimated change of −0.10 (95% CI = −0.24 to 0.04). In the IG, the mean of the QOL-CS scale increased marginally from 5.66 (SD = 1.25) to 5.68 (SD = 1.26), with an estimated change of 0.01 (95% CI = −0.15 to 0.32). No statistically significant between-groups differences were observed (difference-in-differences = 0.11, 95% CI = −0.10 to 0.32, P = .303), and the effect size difference was minimal (−0.15). Changes in the QOL-CS subscales were also analyzed. No statistically significant differences were observed between the groups in the physical, psychological, or social well-being subscales. However, a statistically significant reduction in spiritual well-being was observed in the IG (−0.25; 95% CI = −0.49 to −0.02), although this did not translate into a significant difference between groups. Overall, the intervention did not produce statistically or clinically significant improvements in overall QoL.

Table 2.

Intent-to-treat analysis: change in outcomes between the CG and IG at 0 and 3 months.

SS	Mean (SD)		Estimated post-pre difference (95% CI)	Difference of estimated differences (95% CI)	P	Difference of estimated effect sizes
	Baseline (pre)	3 months (post)
QOL-CSa
CG	5.89 (1.25)	5.79 (1.29)	−0.10 (−0.24 to 0.04)	0.11 (−0.10 to 0.32)	.303	−0.15
IG	5.66 (1.25)	5.68 (1.26)	0.01 (−0.15 to 0.32)
Physical Well-Being Subscaleb
CG	6.78 (1.90)	6.58 (1.92)	−0.19 (−0.45 to 0.06)	0.31 (−0.07 to 0.70)	.176	−0.11
IG	6.53 (1.57)	6.64 (1.54)	0.12 (−0.15 to 0.39)
Psychological Well-Being Subscalea
CG	5.49 (1.52)	5.49 (1.53)	0.00 (−0.18 to 0.17)	0.10 (−0.17 to 0.37)	.465	−0.10
IG	5.31 (1.53)	5.41 (1.57)	0.10 (−0.11 to 0.31)
Social Well-Being Subscaleb
CG	6.57 (1.89)	6.41 (1.89)	−0.16 (−0.42 to 0.11)	0.11 (−0.27 to 0.47)	.133	−0.12
IG	6.32 (1.86)	6.26 (1.92)	−0.05 (−0.30 to 0.20)
Spiritual Well-Being Subscaleb
CG	5.14 (1.47)	4.98 (1.49)	−0.15 (−0.39 to 0.07)	−0.10 (−0.44 to 0.24)	.488	0.06
IG	4.86 (1.45)	4.61 (1.45)	−0.25 (−0.49 to −0.02)

Abbreviations: CG = control group (n = 99); CI = confidence interval; IG = intervention group (n = 102); QOL-CS = Quality of Life–Cancer Survivors scale validated Spanish version

^aIndependent-samples t tests and Cohen d interpreted as: <0.20 (very small), 0.20-0.49 (small), 0.50-0.79 (moderate), and ≥0.80 (large).²⁷

^bWilcoxon rank-sum test and rank-biserial correlation coefficient (r) interpreted as: <0.10 (very small), 0.10-0.29 (small), 0.30-0.49 (moderate), and ≥0.50 (large).²⁷

Discussion

This RCT, the first of its kind, evaluated a digital health intervention to improve QoL in LT-BCSs (>5 years disease-free). Contrary to our initial hypothesis, no statistically or clinically significant differences were observed between groups in the primary outcome (QOL-CS total score) at 3 months.

Our results differ from previous research suggesting that psychosocial and behavioral eHealth interventions can improve QoL in cancer survivors, especially when implemented during or shortly after active treatment.^28-31 Several meta-analyses and systematic reviews have shown that digital or hybrid interventions based on cognitive-behavioral therapy or self-care education can improve psychological well-being and functional outcomes in populations with cancer,^32-34 including breast cancer survivors.^20,32 However, many of these studies involved survivors in the early stages (<5 years) of posttreatment recovery,^20,32 which could partly explain the lower effect size observed here.

In our study, the sample comprised LT-BCSs with an average of more than 11 years posttreatment, a group often underrepresented in intervention trials.³⁵ While including this population contributes to the generalizability of our findings to the growing demographics of survivors, it may also suggest that their unmet needs and responsiveness to health promotion interventions differ substantially from those of women with more recent diagnoses and shorter survival times.³⁴ This difference could influence the magnitude of the impact of the intervention, as women who have been disease-free for more than a decade may have already developed effective coping strategies or perceive a reduced need for additional support, thereby limiting the potential benefits of a digital intervention.

The intervention was specifically designed to address the needs of LT-BCSs identified in the literature and was delivered as a dynamic, engaging, and self-directed web-based intervention, features consistent with current trends in digital health and survivorship care planning.³⁶ However, the absence of significant effects suggests that digital formats alone may not be sufficient to significantly influence complex constructs such as QoL. This is consistent with the findings of previous eHealth trials, which reported modest or even zero effects when interventions lack a high degree of personalization, interactivity, or strategies to maintain sustained user engagement.^20,37-39 In other words, the simple availability of content does not guarantee its effective use or the adoption of behavioral changes, especially in populations that may present motivational and technological barriers or a lower perception of the need for support after long periods free from disease. Therefore, digital interventions supplemented with human support, real-time personalized feedback, or hybrid components that integrate interactions with professionals may be more effective in producing clinically meaningful improvements in complex outcomes such as long-term QoL.

Moreover, self-management programs that combine educational and informational components with digital delivery have been shown to enhance engagement, promote behavior change, and increase the potential impact of interventions.^40-42 Incorporating structured education alongside app-based tools represents a promising approach to optimize outcomes in LT-BCSs.

Notably, a small but statistically significant decrease in spiritual well-being was observed in the IG. However, the intervention was not specifically designed to target spiritual outcomes, and the magnitude of this change is small and of uncertain clinical relevance, as no minimal clinically important difference is established for this subscale. Given the number of QoL domains analyzed, this finding may plausibly reflect chance variation and should therefore be interpreted cautiously. Although the cause of this decline remains unclear, it raises important questions about how survivorship interventions address the existential dimensions of recovery and whether some digital formats may be insufficient to adequately meet these needs.^43,44 A recent systematic review on eHealth interventions to improve the QoL of cancer survivors revealed that while physical, emotional, and social aspects were frequently targeted, none strategies explicitly addressed spiritual well-being.²⁰ Future studies should delve into this area and consider incorporating spiritually oriented components or hybrid models that include human interaction when appropriate.

This study has several limitations that should be noted. First, the 3-month follow-up period may have been insufficient to observe the delayed effects of the intervention. Second, the results were based on self-reported measures, which could be susceptible to recall or desirability biases. Finally, adherence and commitment to the content of the intervention were not measured, and no indicators such as frequency of access to the app were available, limiting our ability to assess participants’ engagement and evaluate whether low exposure may have contributed to the observed lack of significant effects.

Despite these limitations, this RCT provides valuable information on the feasibility and challenges of digital survivorship interventions for LT-BCSs. This highlights the need for more personalized or hybrid approaches that can address the specific and changing needs of this growing population. Future research should investigate personalized digital interventions, explore long-term outcomes, and consider combining self-guided content with professional support to increase their impact.

Although this digital intervention did not significantly improve participants’ QoL, the findings highlight important considerations for designing future digital care strategies for the long-term survivorship stage of cancer. From a clinical perspective, these tools are likely to be more effective if they incorporate higher levels of personalization and mechanisms for professional support. Likewise, longitudinal studies are needed to evaluate the long-term sustainability of these interventions and to confirm or refute our findings in broader clinical settings, ideally through multicenter trials with more diverse samples. These efforts could help optimize digital care strategies and guarantee more effective, needs-tailored support for LT-BCSs.

Author contributions

Gustavo Adolfo Pimentel-Parra (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing—original draft, Writing—review & editing), Nelia Soto-Ruiz (Conceptualization, Funding acquisition, Project administration, Supervision, Visualization, Writing—review & editing), Paula Escalada-Hernández (Writing—review & editing), Leticia San Martín-Rodríguez (Writing—review & editing), and Cristina García-Vivar (Conceptualization, Funding acquisition, Project administration, Supervision, Visualization, Writing—review & editing)

Funding

This study has been funded by Instituto de Salud Carlos III (ISCIII) through the project “PI21/00894” and co-funded by the European Union. The article processing charges for this publication were covered by the Public University of Navarre under the agreement established with the Conference of Rectors of Spanish Universities (CRUE).

Conflicts of interest

The authors have no funding or conflicts of interest to disclose.

Data availability

Data are not publicly available due to applicable Spanish data protection regulations, including the General Data Protection Regulation (GDPR) and the Spanish Organic Law 3/2018 on Personal Data Protection and Digital Rights. Data may be available from the corresponding author upon reasonable request and subject to approval by the relevant ethics committee and data protection authorities.

Ethical approval and consent to participate

This research project was approved by the Clinical Research Ethics Committee of Navarre (PI-2021/18) and complied with the ethical principles stipulated in the Declaration of Helsinki.