We conducted a prospective nonrandomized controlled study with a cohort consisting of outpatients who were being treated with systemic therapy (chemotherapy) for early stage breast cancer at the Institute of Oncology in Ljubljana. The study protocol was approved by the Ethical Commission at the Institute of Oncology and Commission of the Republic of Slovenia for Medical Ethics (0120-386/2017/3). The researchers followed the principles of the Declaration of Helsinki.

Patients were recruited after their first appointment with a medical oncologist. All study information was provided by oncologists, medical residents, and medical students working on the project. Patients were eligible if they were scheduled to receive neoadjuvant or adjuvant chemotherapy (and anti–human epidermal growth factor 2 targeted therapy, if indicated) for early stage breast cancer, their smartphone was Android-based, and they were willing to fill in paper and pencil questionnaires reporting their quality of life while receiving treatment. The patients were ineligible if their smartphone was Windows-based or an iPhone, they were unfamiliar with using a smartphone, or they were unable to communicate effectively in Slovenian. All participants provided informed consent prior to being enrolled in the study.

The control group consisted of patients enrolled between December 2017 and March 2018; the intervention group consisted of patients enrolled between April 2018 and September 2018 (since the mobile app was not finished until April). The control group received the standard information about chemotherapy side effects. The intervention group received both the standard information and access to the mPRO Mamma mobile app [17] which was accompanied by a detailed user instruction guide.

Patients in both groups were asked to complete the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ C-30) and Quality of Life Questionnaire Breast Cancer Module (EORTC QLQ BR-23) [18,19] and a questionnaire about health resources use at 4 time points: before systemic therapy (baseline), after the first week, after the first cycle, and at the end of systemic therapy. Upon enrollment, each patient received questionnaires (4 sets) along with instructions to return completed questionnaires by mail in a timely manner. The patients who returned the baseline questionnaires and questionnaires from at least one other time point were included in the analysis.

For the purpose of this study, we developed an Android mobile app called mPRO Mamma [17]. It conforms to standard Google Material Design guidelines [20], but additionally features customized slider controls to visually indicate symptom severity levels (Figure 1). It allows for quick daily recording of symptoms and symptom severity (with reminder notifications) and can send encrypted reports to the patient’s oncologist, if selected. The app was designed and revised several times in cooperation with experienced medical oncologists, medical students, and computer scientists. It has been implemented according to the best principles of Material Design multilingual support and currently supports Slovenian and English.

We selected 50 symptoms that were relevant to the expected adverse effects of either cytotoxic regimens or targeted drugs administered to patients during the neoadjuvant or adjuvant breast cancer treatments from the 78 symptoms of PRO-CTCAE. Each symptom had a description, a severity scale, and advice for its management. Patients labeled and reported a symptom only if it occurred. To make our app clear and useful, symptom severity was graded on a 3-point scale as mild, moderate, or severe.

For each grade of symptom severity, a health professional’s advice was presented. For mild or moderate ratings by the patient, suggestions that would allow them to alleviate the symptoms on their own were displayed. If symptoms were rated as severe, the app advised the patient to visit either their general practitioner or an emergency department. Patients reported their symptoms in the app on a daily basis and sent the symptom report to their treating oncologist as an encrypted email on the day before their next therapy cycle (every 2 or 3 weeks, depending on their chemotherapy schedule). Screenshots of the symptom entry page and daily report are presented in Multimedia Appendix 1, Multimedia Appendix 2, and Multimedia Appendix 3.

The EORTC QLQ C-30 and BR-23 consist of a global quality of life score, and several functional and symptom scales, which can be combined in a summary score. Our primary outcomes were the global quality of life score and summary score at three time points (after the first week, after the first cycle, and at the end of treatment). Other scales were used in our exploratory analysis. The EORTC QLQ scoring manual was followed [21]. All scales had values from 0 to 100, where 100 represented the best global quality of life, the best functioning, or the worst symptoms. The summary score, which ranged from 0 (worst) to 100 (best), was calculated from 13 out of 15 EORTC QLQ-C30 scales (the global quality of life and the financial impact scale were excluded) in accordance with Giesinger et al [22] and instructions from the EORTC [21-23]. We used previously established thresholds [24] for important differences in quality of life—clinically meaningful difference was set to 10 points.

The secondary outcome was the use of health resources which was measured by self-reported number of visits to the doctor and number of hospitalizations in the first week and in the first cycle of therapy (questionnaires at the end of treatment were not included because they may not have been reliable due to the long time span between the start and the end of treatment).

The study flow chart is presented in Figure 2. We performed a CONSORT-EHEALTH checklist (Multimedia Appendix 4) according to pilot testing [28].

The demographic, clinical, and treatment data for the control (n=45) and the intervention (n=46) groups are presented in Table 1. Patients in both groups were similar with respect to age (t₈₉=0.86, P=.39), tumor size (Fisher exact test, P=.48), nodal involvement (Fisher exact test, P=.13), histology (Fisher exact test, P=.42), tumor grade (Fisher exact test, P=.38), hormone receptors (progesterone: χ²₁=1.23, P=.27; estrogen: χ²₁=2.33, P=.13), human epidermal growth factor 2 status (χ²₁=1.76, P=.19), tumor subtype (Fisher exact test, P=.30), type of chemotherapy (Fisher exact test, P=.41), and number of chemotherapy cycles (Mann-Whitney U test, P=.73); however, there was a significant difference between the intervention and control groups with respect to type of surgery (χ²₃=12.60, P=.006). Patients in the control group had undergone more mastectomies, and consequently, had also undergone reconstruction more often (Fisher exact test, P=.01). At baseline, patients in the control group had lower quality of life and summary scores than those of the intervention group (quality of life: Mann-Whitney U test, P=.06; summary: Mann-Whitney U test, P=.02). Baseline differences in quality of life and summary scores and for type of surgery (mastectomy or not) between the groups were, therefore, taken into account and adjusted for, within the model.

In Table 2 and in Table 3, we have presented only effects that are of interest (ie, adjusted mean difference between the groups for each of the 3 time periods with 95% confidence intervals and P values; for all mean differences, a positive difference indicates a higher value in the intervention group than the corresponding value in the control group). The adjusted mean differences between the intervention and the control group after the first week of chemotherapy were statistically significant for both the global quality of life score (10.1, 95% CI 1.8 to 18.5, P=.02) and the summary score (8.9, 95% CI 3.1 to 14.7, P=.003). The adjusted mean difference for summary score at the end of treatment (10.6, 95% CI 3.9 to 17.3, P=.002) was also significant. Complete model results are presented in Multimedia Appendix 5 and Multimedia Appendix 6. Clinically meaningful differences in the adjusted means between groups were found for global quality of life score after the first week and for summary score at the end of treatment (ie, an adjusted mean difference greater than 10).

In the first week, 44% (20/46) of patients in the intervention group visited the doctor, and 60% (27/45) of patients in the control group visited the doctor. In the first cycle, 37% (16/43) of patients in the intervention group visited the doctor at least twice, and 54% (21/39) of patients in the control group visited the doctor at least twice. The differences between groups were not statistically significant (first week: χ²₁=2.49, P=.12; first cycle: χ²₁=2.29, P=.13; Table 4). The number of hospitalizations was low—3% (3/91) in the first week and 7% (6/84) in the first cycle—without substantial or statistically significant differences between the groups (first week: Fisher exact test, P=.62; first cycle: Fisher exact test, P>.999).

Scales with clinically important differences between the groups (baseline adjusted), and with P<.10 (Mann-Whitney U tests) were social functioning (first week: P=.001, adjusted P=.04; first cycle: P=.005, adjusted P=.29; end of treatment: P=.003, adjusted P=.14), pain (first week: P=.03, adjusted P>.999; end of treatment: P=.005, adjusted P=.29), physical functioning (first week: P=.051, adjusted P>.999; end of treatment: P=.005, adjusted P=.27), role functioning (end of treatment: P=.01, adjusted P=.49), cognitive functioning (end of treatment: P=.09, adjusted P>.999), appetite loss (first week: P=.04, adjusted P>.999), and systemic therapy side effects (first cycle: P=.08, adjusted P>.999). These scales are presented in Figures 3 and 4. Complete exploratory analysis results are available in the Multimedia Appendix 7. The only difference between the groups that could be generalized to the population after adjusting for multiple testing was in social functioning after the first week (adjusted P=.04).

In this study, we investigated the impact of using of a mobile app for symptom reporting and self-management on patient-reported outcomes in early stage breast cancer patients who were receiving chemotherapy. Chemotherapy can cause numerous adverse effects which have a negative impact on health-related quality of life. Our hypothesis was that the intervention group (those using the mobile app) would cope better with adverse effects, which would result in the improved global quality of life and summary scores, especially at the beginning of chemotherapy, and would result in reduced use of health resources.

Our first hypothesis about the improved global quality of life and summary score in the intervention group using the mobile app was confirmed. After the first week of treatment, global quality of life reported by the intervention group was better than that reported by the control group, with clinically important mean differences (Table 2). The global quality of life score represented patient perception of overall health and overall quality of life in the previous week, but was not sensitive enough to detect group differences over time [22]. This was confirmed in our analysis as both groups had similar global quality of life scores after the first cycle and at the end of treatment. Summary score, however, was a more robust alternative to the global quality of life score, as it took into account 13 scales that were calculated from 27 out of 30 EORTC QLQ C-30 items. We found statistically significant differences between groups in summary scores after the first week and at the end of treatment; the latter demonstrated a clinically important difference (Table 3). Our mobile app assisted patients as an interventional tool for the self-management of symptoms. As the patient reported a symptom, the mobile app automatically provided the most suitable information for self-management according to the reported level of severity (Multimedia Appendix 1 and Multimedia Appendix 2). This improved patient operative capability and patient self-efficacy, which resulted in an improved quality of life in accordance with Bandura’s theory of self-efficacy theory [29].

To investigate which EORTC QLQ C-30 scales most significantly contributed to the difference in summary score, we performed an exploratory analysis. QLQ BR-23 scales were also included in this analysis to shed light on symptoms and functioning that were specific to breast cancer. We found clinically important improvements in social, physical, role, and cognitive functioning in the intervention group versus the control group, as well as less severe pain, less appetite loss, and fewer systemic therapy side effects in the intervention group (Multimedia Appendix 7, Figure 3, and Figure 4). In a recently published paper on lung cancer [30], patients with access to a web-based health education program reported better global quality of life and emotional functioning, as well as a significant decrease in ten of the most important symptoms compared to those in the control group; however, no differences in physical, role, cognitive, or social functioning were found [30]. Patients with lung cancer, however, cannot be directly compared to patients with breast cancer. To our knowledge, comparable studies for breast cancer patients do not exist. We believe that the intervention group had better social, physical, role, and cognitive functioning due to the use of the mobile app (better self-management). For cancer patients, solving the problem (ie, self-managing of symptoms) correlates significantly with improvements in their levels of psychological distress, overall quality of life, role, and physical functioning [31,32].

We found the strongest effect was for social functioning after the first week, as the difference between groups remained statistically significant even after adjusting for multiple comparisons. Patients who receive chemotherapy treatments experience imposed limitations in social functioning, which is also influenced by family support, outlook on life and opportunities for social exchange [33]. The mobile app had a positive impact on social functioning, although, a bias as a result of nonrandom patient allocation may have been present.

Less severe systemic therapy side effects, pain, and appetite loss in the intervention group were probably due to self-management techniques employed by these patients. On the other hand, the surgery-related group differences may have contributed.

Our second hypothesis was not confirmed. We did not find any statistically significant difference between groups for use of health resources. Self-management strategies employed by patients themselves probably resolved mild and moderate side effects. For severe symptoms, we presumed that an option of immediate triggering alert to medical staff at its onset would reduce health resource use. Basch et al [5] reported significantly lower health resource usage in the group allocated to symptom reporting, probably because a severe symptom grade triggered an email alert to nurses.

Ginossar et al [34] performed a systematic review of literature and found over 100 breast cancer-related mobile apps; however, many of them did not lead to behavior change and many were not evidence-based. Recent mobile apps for women with breast cancer have a positive effect by promoting weight loss, decreasing stress, and improving the quality of life [35,36]. Apps for reporting symptoms and promoting self-care during cancer chemotherapy treatment are still rare [6-11,37,38]. Electronic capture and monitoring of patient-reported outcome for symptoms during systemic treatment of cancer has been found to be feasible, even though patients were not provided with feedback about their management [5]; however, in studies [6,7,9,10] where automated alerts for self-management of reported symptoms were available, a rapid benefit in the form of decreased symptom severity has been identified. It has been reported [6] that automated alerts were more effective than nurse-administrated symptom management. In our study, we also rapidly found a benefit from the use of our mobile app, which was demonstrated by better global quality of life and summary scores in the first week after receiving chemotherapy. In our opinion, the first week after chemotherapy is a very important time to alleviate distress in patients. A recently published randomized trial from Japan [37] tested a mobile app similar to ours, to see whether the app affected anxiety and depression, but no improvement in anxiety, depression, or health literacy were found at the end of treatment [37].

Using a mobile app in collaboration with the treating physician improved patient well-being and their awareness of chemotherapy adverse effects [38]; however, the impact of the mobile app on quality of life outcomes was not assessed [37,38]. Zhu et al [9] reported findings similar to ours, namely that the e-support group had better quality of life, self-efficacy, and symptom control than that of the group receiving standard care. They found beneficial effects at 3 months, but which disappeared at 6 months [9]. In our study, even at the end of treatment (approximately 6 months), the beneficial effect in summary score (but not in global quality of life) remained; however, Zhu et al [9] assessed quality of life using a different questionnaire, ie, the Functional Assessment of Cancer Treatment-B. Recently, a mobile phone–based system for the remote monitoring and management of chemotherapy-related side effects was evaluated in Canadian patients with cancer; however, its impact on health-related quality of life has not yet been tested [11]. Several randomized controlled trials evaluating remote electronic monitoring and symptom management are still ongoing [14-16].

The advantage of our study was the detailed patient-reported outcome analysis of the effect of mobile app usage, with data collected using validated questionnaires (EORTC QLQ C-30 and BR-23). In addition to using a standard tool (global quality of life score), we used summary score, a new tool suggested by EORTC. To our knowledge, our study is the first that uses both quality of life and summary score as primary outcomes when analyzing the impact of a mobile app on the care of patients with breast cancer who are receiving chemotherapy. Moreover, we appropriately planned the first assessment for after the first week, indicated by the statistically significant difference in the global quality of life and summary score at this time.

Our study has some limitations. First, it was a nonrandomized controlled cohort study. In order to avoid potential bias, it would have been better to conduct a randomized controlled trial. In fact, the groups in our study differed with respect to type of breast surgery, quality of life score, and summary score at baseline, so the results were adjusted for these differences using appropriate statistical methods. There may also have been seasonal differences affecting patient well-being since patients were enrolled at different times of the year. In addition, a possible bias may have existed as a result of different recruitment techniques used by the medical oncologists. At the beginning of the clinical study, we were still in the process of programming our mobile app which prevented us from randomly allocating patients.

Second, not all symptoms that can arise during the systemic therapy can be assessed with the QLQ C-30 and BR-23 questionnaires. Our app included approximately 50 of the 78 recommended PRO-CTCAE symptoms, so instructions for patient self-management for the missing symptoms were not included in the app; however, we aimed to include the symptoms that occur most commonly over the course of breast cancer treatments. Our collection was broader than that used by Zhang et al [39] who developed and evaluated a patient-reported outcome scale for breast cancer containing 38 items [39].

Our app could also be improved by utilizing wearable sensors (smart bracelets, watches, rings, etc) and incorporating additional questionnaires capable of measuring psychological distress, mental distress, and anxiety; it could be also equipped with alerts reminding patients to provide data at predetermined dates.

Based on the positive effects on health-related quality of life measured in the systemic therapy part of breast cancer treatment, we plan to include other treatment modalities for breast cancer patients, ie, surgical interventions and radiotherapy. Another interesting upgrade would be to include a follow-up period. Our aim is to provide regular updates to ensure software compatibility as well as to avoid potential security issues. Since our app is likely most useful for mild and moderate symptoms where patients are empowered to manage symptoms themselves, we plan to also make the app useful for patients with severe symptoms. We could establish a system where alerts about patient symptoms are sent to a dedicated hospital server. If the alert contained any severe symptoms, it would be labelled as urgent and an appointed research nurse would be trained to react in accordance with the prespecified medical algorithms. Such algorithms are also incorporated in recently reported proposals of studies [14-16] to assess symptom burden, quality of life, supportive care needs, anxiety, self-care, self-efficacy, work limitations, and cost effectiveness.

Our findings regarding the impact of mobile app use on the quality of life and summary scores in combination with those regarding other scales that were only a part of our exploratory analysis, should be confirmed in a larger randomized controlled trial. Further research is also needed in order to extend these findings to a broader population of patients, including those suffering from other ailments, and to different organizations in the medical field.

Our mobile app has interventional value for the self-management of symptoms for patients with breast cancer who are receiving systemic therapy which was shown by a better global quality of life scores (in the first week of therapy) and a better summary scores (in the first week and at end of treatment) for the intervention group than for the control group. Based on the exploratory analysis, the app contributed to clinically important improvements in social, physical, role, and cognitive functioning while diminishing pain, appetite loss, and systemic therapy side effects.