This systematic review and meta-analysis was registered in PROSPERO (CRD42024611768), and the study was reported following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines (Multimedia Appendix 1) [34].

We conducted electronic literature searches in PubMed, Scopus, Web of Science, Cochrane Library, CINAHL, and EMBASE from the inception of each database to October 15, 2024. The search terms consisted of subject headings and free terms. The full search strategy and other relevant information are provided in Table S1 in Multimedia Appendix 2. Meanwhile, manual searches were carried out according to reference lists of relevant articles.

Literature searches were limited to published English-language articles. Eligible studies had to meet the inclusion criteria provided in Textbox 1.

Studies meeting the inclusion criteria provided in Textbox 1 were included in this review.

Search results from all databases were imported into EndNote X9 (Clarivate) software to remove duplicate studies and manage the research. Two independent reviewers (LT and YW) screened the titles and abstracts of the studies according to the inclusion criteria. After identifying potentially eligible studies, the full texts were screened again to determine the final included studies. Then, the 2 reviewers (LT and YW) discussed and reached a consensus. A third reviewer (JF) was available to assist and mediate throughout the process.

Data extraction followed the PRISMA statement to ensure a systematic approach. Two reviewers (LT and YW) independently extracted data using a pretested data extraction form, and a third author (JF) cross-checked the accuracy of the results. The data extracted included publication information (author, research year, and country); study design; characteristics of the study participants (sample size, age, sex, and type of cancer); duration of intervention; intervention methods in the intervention group and the control group, respectively; primary outcomes and assessment scales (methods); and main results. The primary outcome data were extracted as mean and SDs.

Two authors (LT and YW) independently evaluated the risk of bias, methodological quality, and certainty of evidence of the included studies. In case of inconsistencies, discussions were held with a third reviewer (JF) until a consensus was reached. The Revised Cochrane Risk-of-Bias Tool was used to assess the risk of bias from the following aspects: randomization process, deviation from the intended interventions, missing outcome data, outcome measurement, and selection of reported results [35]. The overall quality of evidence for each outcome was rated according to the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) guidelines [36]. Considering that only randomized controlled trials (RCTs) were included, each outcome was initially assigned a high quality. Two reviewers (LT and YW) used the GRADE system to evaluate the quality of evidence, and potential disagreements were resolved through discussions with a third reviewer (JF). Table S2 in Multimedia Appendix 2 presents the GRADE downgrading details.

Due to the expected heterogeneity among the studies, we used a random-effects model to conduct a pooled estimate of the postintervention effects. Using the inverse-variance method, we calculated the standardized mean differences (SMDs) and 95% CIs for continuous variables. The effect sizes were interpreted as small (0.2), medium (0.5), or large (0.8). Statistical heterogeneity was evaluated using I², which was classified as low (I²<25%), moderate (I²=25%-50%), high (I²=50%-75%), and considerably high (I²>75%). Publication bias was assessed by calculating Begg and Egger test values, and P<.05 was considered publication bias [37,38]. A leave-one-out sensitivity analysis was performed to explore the robustness of the pooled effect [39]. We performed all statistical analyses using Review Manager (version 5.4; Cochrane) and Stata (version 17; StataCorp LLC).

In the initial stage of the research design, our focus was on the impact of remote web-based interventions on physiological states (pain and fatigue) and the quality of life of patients with cancer. When we extracted the relevant data from the included studies, we found that there were controversies regarding the efficacy of remote web-based interventions on the anxiety and depression of patients with cancer. Although anxiety and depression were not listed as outcome indicators in the initial registration protocol, they are closely related to the research topic. Considering that anxiety and depression are common indicators of the psychological state of patients with cancer and have a significant impact on their lives, anxiety and depression cannot be ignored when evaluating the effect of remote web-based interventions. Therefore, to enrich the research content and enhance the clinical significance of the results, compared with the registered protocol, we have added outcome indicators of anxiety and depression. There were no other inconsistencies with the preregistration protocol.

Figure 1 illustrates the flow of this study and the reasons for exclusion. A total of 15,209 records were identified from the databases. After removing duplicate literature, 9309 (61.2%) potentially eligible studies were determined. After screening the titles and abstracts, we retrieved 101 studies for a more in-depth review. Eventually, 35 (34.6%) RCTs met the inclusion criteria. Among the 66 excluded studies, 19 (29%) had intervention methods that did not meet the inclusion criteria, 19 (29%) lacked complete data, and 28 (42%) did not contain relevant outcome indicators. In addition, we tracked and supplemented relevant citations. Finally, 36 studies met the inclusion criteria for our meta-analysis [15,19-33,40-59].

A total of 36 RCTs were included in this study [15,19-33,40-59]. The study sites involved countries such as China, Iran, the United States, Spain, the Netherlands, Japan, the United Kingdom, Australia, and Germany. The sample size of the studies varied from 29 to 417, with 4521 patients with cancer included. The age range of patients with cancer was wide (>18 years). The patients with cancer covered various types of cancers, including non–small cell lung cancer, breast cancer, urinary system cancers, digestive system cancers, acute leukemia, esophageal cancer, ovarian cancer, cervical cancer, uterine cancer, vaginal cancer, and mixed cancer types. The apps, miniprograms, or other web-based platforms used in the intervention group included WeChat, WhatsApp, ePAL, Microsoft Teams, Zoom, KLIK Pijnmonitor, Welby My Carte ONC, ASyMS, MyBreastHealth, Kanker Nazorg Wijzer, Breast Cancer Support Zone program, the miniprogram “Electronic Intestinal Safety,” e-CUIDATE system, Breast Cancer Self-Management Support app, ENCOURAGE project, Steps Toward Improving Diet and Exercise website, the website named “Together,” Shafayar, Mika, and Untire mobile app. The intervention duration ranged from 1 week to 18 months, and most of the studies focused on 3 to 12 months. The outcome assessment indicators included anxiety, depression, pain, fatigue, and quality of life. More information about the main results of each study is provided in Table 1.

As shown in Figure S1 in Multimedia Appendix 2 [15,19-33,40-59], a total of 25 (69%) studies were assessed as having an overall low risk [15,20-23,25-27,32,33,40-45, 47,48,50,52,54-56,58,59], 7 (19%) studies were assessed as having some concerns [19,28-31,49,51], and 4 (11%) studies were assessed as having a high risk [24,46,53,57]. Except for 1 (3%) study [46], which was assessed as having a high risk, the remaining 35 (97%) studies reported an adequate randomization process and were evaluated as having a low risk of bias. Regarding deviation from the intended interventions, 35 (97%) studies were considered to have a low risk of bias, and 1 (3%) study had some concerns [49]. Regarding missing outcome data, 28 (78%) studies were rated as having a low risk of bias, 6 (17%) had some concerns [19,28-31,51], and another 2 (5%) were assessed as having a high risk of bias [24,57]. The outcome measurement risks of 2 (5%) studies were rated as having a high risk of bias [53,57], and the remaining 34 (94%) studies were rated as having a low risk of bias. Regarding the selection of reported results, 36 (100%) studies had a low risk of bias. More detailed information on the risk-of-bias assessment is provided in Table S3 in Multimedia Appendix 2 [15,19-33,40-59].

The sensitivity analysis for the meta-analysis results of pain (Table 3), anxiety (Table S8 in Multimedia Appendix 2 [26-33,40,43,45,47,53,55]), depression (Table S9 in Multimedia Appendix 2 [26-28,30,32,33,40,43,47]), and quality of life (Table S10 in Multimedia Appendix 2 [15,19,20,22,23,25, 27,28,30,32,33,40-42,44-48,50,52,54-56,58,59]) showed that when any single study was excluded, the pooled estimates of the main analysis were robust. However, for the meta-analysis result of fatigue, the sensitivity analysis indicated that after excluding the studies by Xia [45] and Wang et al [28], respectively, the results of the overall pooled analysis changed from statistically significant (SMD –0.52, 95% CI –0.95 to –0.09; I²=95%) to nonstatistically significant (SMD –0.28, 95% CI –0.57 to –0.01; I²=91% and SMD –0.39, 95% CI –0.81 to –0.03; I²=91%), suggesting that the main analysis was not robust in the pooled estimates (Table S11 in Multimedia Appendix 2 [20,22,26,28,30,32,42,45,48,51,57]).

As shown in Table 4, the results of Begg and Egger tests (P＞.05) indicate no publication bias in the pooled analyses of pain, fatigue, anxiety, depression, and quality of life.

In recent years, the health management of patients with cancer has shifted from single-disease treatment to a holistic focus on both physiological and psychological states. Patients with cancer not only have to cope with physiological symptoms (pain and fatigue) brought about by chemotherapy, radiotherapy, and other treatments, but they are also often burdened with psychosocial issues, including anxiety, depression, posttraumatic stress disorder, and fear of recurrence [60-63]. Although traditional face-to-face interventions can partially alleviate these problems, their accessibility and sustainability are limited due to the uneven distribution of medical resources, patients’ limited mobility, or remote areas [64,65]. Remote web-based interventions, with their advantages of high accessibility, great flexibility, and low cost, have gradually become an important means for patients with cancer to manage their symptoms by themselves [14]. However, although previous studies have suggested that remote web-based interventions may improve physiological and psychological states and quality of life in patients with cancer, the effectiveness remains controversial [19-25,27-33]. To this end, this systematic review and meta-analysis aimed to summarize the relevant evidence regarding the physiological and psychological impacts of remote web-based interventions on patients with cancer. A total of 36 studies were included to determine the clinical efficacy of remote web-based interventions on physiological pain, fatigue, anxiety, depression, and quality of life in patients with cancer.

This study, through meta-analysis, found that remote web-based interventions significantly reduced the pain intensity in patients with cancer (n=14, 39% studies; SMD –0.39, 95% CI –0.64 to –0.14; I²=82%; GRADE rating=low). This result is consistent with the results of some previous studies. For example, in the review by Villinger et al [66], it was mentioned that some apps showed certain positive effects in cancer pain management. Through the real-time chat consultation function, patients can report pain and receive suggestions, which is helpful for pain management. The review by Wu et al [67] also pointed out that mHealth apps have certain potential in monitoring and managing cancer pain. After some patients used pain-related apps, their physiological pain intensity improved. Through remote monitoring and digital platforms, patients with cancer can report symptoms (pain) in real time, enabling health care providers to adjust treatment plans promptly [68]. Meanwhile, the health education module enhances the understanding of relevant symptoms in patients with cancer and improves their self-management [69]. In this study, the pooled effect value of pain showed high heterogeneity, with an overall heterogeneity of 82%. Further subgroup analyses revealed the impacts of different regions, app types, and intervention durations on heterogeneity: the heterogeneity in the Asian subgroup was 65%, 85% in the American subgroup, and 79% in the European subgroup. Regarding the type of intervention app, the heterogeneity in the WeChat subgroup was 46%, while it was 86% in the non-WeChat subgroup. In addition, the heterogeneity in the subgroup with an intervention duration of less than 3 months was 89%, and it decreased to 68% in the subgroup with an intervention duration of more than 3 months. The above results indicate that differences in intervention apps may be one of the main sources of heterogeneity. In contrast, factors such as region and intervention duration are related to heterogeneity but are not the main sources. The sensitivity analysis showed that when any single study was excluded, the main analysis was robust in the pooled estimates, indicating that the study’s results regarding pain have certain reliability.

Studies show that, in addition to symptom management and educational empowerment [68,69], remote web-based interventions can also customize exercise programs (Tai Chi and aerobic training) for patients with cancer to improve their cardiopulmonary function and muscle strength and alleviate fatigue [70]. Our results indicated that remote web-based interventions significantly affected the fatigue state of patients with cancer (n=11, 31% studies, SMD –0.52, 95% CI –0.95 to –0.09; I²=95%; GRADE rating=low). However, the sensitivity analysis showed that after excluding the studies by Xia [45] and Wang et al [28], respectively, the statistical significance of the pooled analysis results changed. This result suggests that the main analysis was not robust in the pooled estimates. These 2 studies might have been unique regarding intervention measures, sample characteristics, or fatigue assessment methods, significantly influencing the overall results. The results of the sensitivity analysis indicate that there is still a high degree of uncertainty in current research on the impact of remote web-based interventions on fatigue state of patients with cancer, and this result should be interpreted with caution.

This study found that remote web-based interventions significantly improved the anxiety state of patients with cancer. This result is consistent with some of the research results mentioned in the review by Villinger et al [66]. Some apps may help relieve patients’ anxiety by providing information support and social interaction functions. Remote web-based intervention can continuously provide psychological support for patients with cancer [71] and comprehensive, personalized support through multidisciplinary collaboration [64]. In addition, the web-based community established by remote web-based interventions can reduce loneliness among patients with cancer. Sharing experiences among patients can enhance their sense of hope [72], thereby improving their psychological states. The pooled effect value of anxiety showed a very high degree of heterogeneity, with an overall heterogeneity as high as 91%. Subgroup analyses further revealed the potential sources of heterogeneity: the heterogeneities of the Asian subgroup, the American subgroup, and the European subgroup were 88%, 54%, and 0%, respectively, indicating significant differences among regions. Regarding the app type, the heterogeneity of the WeChat subgroup was 90%, and that of the non-WeChat subgroup was 74%. Regarding the intervention duration, the heterogeneity of the subgroup with a duration of less than 3 months was 87%, while that of the subgroup with a duration of more than 3 months was 95%. These findings suggest that the high heterogeneity may stem from regional differences and other unknown factors, and the intervention duration and the type of intervention app are not the main sources of high heterogeneity. The sensitivity analysis showed that when any single study was excluded, the main analysis was robust in the pooled estimates, indicating that the results of this study have a certain degree of reliability.

This study demonstrated that remote web-based interventions could significantly improve the depressive state of patients with cancer. In the review by Samadbeik et al [73], it was pointed out that mHealth interventions can help improve patients’ psychological states. For example, providing psychological support, self-management education, and other content may positively impact depressive symptoms, which is consistent with the results of our study. The overall heterogeneity of the pooled effect value for depression was 81%. In the subgroup analysis, the heterogeneity of the European subgroup was 0%, and the heterogeneities of the other subgroups were all greater than 50%. This result indicates that the region may be one of the main sources of high heterogeneity. At the same time, the intervention duration and the type of intervention app are not the main sources of high heterogeneity. The results of the sensitivity analysis suggest that the results of this study have a certain degree of reliability.

Several studies have shown that digital tools can establish a closed-loop intervention model based on the biopsychosocial medical model through mechanisms such as psychosocial support, social connection and peer support, symptom management and educational empowerment, and multidisciplinary collaborative comprehensive interventions [64,68,69,71]. This intervention model can synergistically improve patients’ physiological and psychological states, enhancing their quality of life. For example, some app interventions mentioned by Villinger et al [66] showed positive effects in aspects such as cancer-related symptom management, thereby improving patients’ quality of life. Samadbeik et al [73] also pointed out that mHealth interventions could improve patients’ quality of life by enhancing self-management and promoting healthy lifestyles. The results of our study also showed that remote web-based interventions had a significant positive impact on the quality of life of patients with cancer. The overall heterogeneity of the pooled effect value for quality of life and the heterogeneities of all subgroups exceeded 50%, which meant that region, the type of intervention app, and intervention duration were not the main sources of high heterogeneity. The results of Begg and Egger tests indicate no publication bias in our study. In addition, the sensitivity analysis results showed that our study was robust.

Previously, 4 systematic reviews [66,67,73,74] and 1 meta-analysis [75] have investigated the effects of apps, miniprograms, or other web-based platforms on symptom management in patients with cancer. We compared our study with the aforementioned articles. We found that Zou et al [74] conducted electronic database searches for studies published until December 2021 and included 20 studies (RCTs=14; quasi-experimental nonrandomized trials=4; and pre-post study=2). Villinger et al [66] carried out a comprehensive and systematic literature search in 5 electronic databases in January 2022 and finally included 17 studies (RCTs=5; nonrandomized trials=3; randomized trial study protocols=3; quasi-experimental design study protocols=1; and pre-post studies=5). Li et al [75] conducted electronic database searches for studies published until April 14, 2021, and included 28 studies (all were RCTs). Wu et al [67] conducted electronic database searches for studies published from 2013 to 2023 and included 20 studies (RCTs=4; quasi-experiments=3; prospective observational cohort studies=4; cross-sectional design=1; qualitative design=3; and mixed methods design=5). Samadbeik et al [73] conducted electronic database searches for studies published between January 2010 and February 2021 and included 28 studies (interventional studies=19, including RCTs; feasibility pilot studies=5; descriptive studies=1; qualitative studies=1; developmental studies=1; nonrandomized controlled studies=1; and mixed methods studies=1). In contrast, we conducted electronic database searches for studies published until October 15, 2024, and included 36 studies (RCTs=36). Compared with the studies by Zou et al [74], Villinger et al [66], Li et al [75], Wu et al [67], and Samadbeik et al [73], our study included newer, more voluminous studies, and all of them were RCTs. Regarding statistical methods, Zou et al [74], Villinger et al [66], Wu et al [67], and Samadbeik et al [73] reported results through comprehensive narrative descriptions. In contrast, we conducted meta-analyses on outcome indicators and reported the results based on them. Furthermore, our study conducted subgroup analysis, sensitivity analysis, and publication bias detection. Although the study by Li et al [75] also reported results using meta-analysis, compared with ours, the outcome indicators in their study were rather single and lacked comprehensiveness. Regarding the conclusions, the studies by Zou et al [74], Villinger et al [66], and Samadbeik et al [73] showed that it was impossible to determine the efficacy of web-based interventions based on apps and miniprograms on pain, fatigue, anxiety, depression, and quality of life in patients with cancer. The study by Li et al [75] indicated that web-based interventions could promote the quality of life of patients with cancer. Wu et al [67] demonstrated that mHealth apps were effective and acceptable in supporting the self-management of cancer pain and could provide patients with multimodal interventions to monitor, track, and manage pain. However, we conducted meta-analyses and quantitatively reported our conclusion: For patients with cancer, remote web-based interventions could alleviate pain, relieve fatigue, improve psychological states, such as anxiety and depression, and enhance their quality of life. In addition, we explored the potential mechanisms of remote web-based interventions to improve the physiological and psychological states of patients with cancer, which enriches the content of our research and theoretical basis. Finally, our conclusions are drawn based on subgroup analysis, sensitivity analysis, and publication bias detection. Therefore, our conclusions are more robust and persuasive.

The results of this study provided valuable references for clinical practice, indicating that remote web-based interventions can serve as a beneficial supplement to the routine care of patients with cancer. Clinical medical staff should fully recognize the potential of remote web-based interventions in improving the symptoms and quality of life of patients with cancer. On the basis of patients’ specific needs and conditions, they should reasonably select and recommend appropriate remote web-based intervention platforms or tools to provide patients with personalized physiological and mental support.

Remote web-based interventions have significant effects in reducing pain, relieving fatigue, improving anxiety, alleviating depression, and improving the quality of life of patients with cancer. However, we could not draw firm conclusions about these findings due to the evidence’s low overall certainty. In the future, more high-quality, large-sample RCTs are needed to validate the effectiveness of remote web-based interventions further.