The effect of tele-interventions on family caregiver burden: a systematic review and meta-analysis Authors: Samane Najafi, Zahra Sadat Manzari, Abbas Heydari Published: 2026 ABSTRACT BackgroundFamily caregiving is often associated with a significant physical and psychological burden. Tele‑interventions have increasingly been used to support family caregivers; however, existing studies report inconsistent findings. This systematic review and meta‑analysis aimed to synthesize evidence on the effect of tele‑interventions on family caregiver burden across different patient populations.MethodsThis systematic review and meta‑analysis was conducted in accordance with PRISMA guidelines and included randomized controlled trials published from 2015 to 2025. Two independent reviewers performed study selection and data extraction. Twenty‑three studies were included in the qualitative synthesis, of which 17 were eligible for meta‑analysis. Pooled, subgroup, and sensitivity analyses were conducted utilizing Stata software, with results reported as standardized mean differences and 95% confidence intervals.ResultsMost tele‑interventions were conducted among caregivers of individuals with neurological conditions, with web‑based formats and educational components being the most common intervention features. Seventeen randomized controlled trials were included in the meta‑analysis. The pooled analysis indicated that tele‑interventions were associated with a greater reduction in family caregiver burden compared with usual care; however, substantial heterogeneity was observed across studies (SMD = -0.39; 95% CI: -0.76 to − 0.01; p < 0.001; I² = 94.50%). Subgroup analyses based on disorder and intervention modality and content were conducted to explore potential sources of heterogeneity and should be interpreted as exploratory. Larger standardized mean differences were observed in studies involving caregivers of individuals with chronic conditions compared with other disease groups, and in interventions delivered via mobile applications.ConclusionThis systematic review and meta‑analysis suggests that tele‑interventions may reduce family caregiver burden, although the findings are limited by substantial heterogeneity across studies. Therefore, further well‑designed randomized controlled trials with standardized outcome measures and diverse caregiver populations are needed to clarify the effectiveness of tele‑interventions and to identify the contexts in which they may be most beneficial.Clinical trial numberNot applicable.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12912-026-04863-x. Background The global increase in chronic, progressive, and life‑limiting diseases has led to a shift in healthcare delivery from hospital‑based models toward home‑ and community‑based care [1]. Consequently, family caregivers, unpaid individuals who provide physical, emotional, and practical support to people with acute or chronic conditions [2], have become integral to the provision of nursing care outside formal healthcare settings [3, 4]. Worldwide, an estimated 349 million individuals serve as family caregivers, including more than 53 million in the United States alone [5, 6]. Although caregiving is a cornerstone of contemporary health systems, it is frequently associated with substantial burden [7]. Caregiver burden refers to the perceived strain experienced by family caregivers across physical, psychological, emotional, social, and financial dimensions during long-term care of patients. This burden typically arises when the demands of caregiving exceed the available resources and coping strategies of the caregiver [8]. The nature and intensity of this burden vary across disease contexts due to differences in disease trajectories, symptom patterns, and self-management demands. For example, neurodegenerative diseases such as dementia often require prolonged supervision as cognitive and functional decline progresses [5, 7, 9, 10], whereas in cancer care the burden often intensifies during active treatment or advanced stages [11–13]. Chronic conditions such as cardiovascular disease and diabetes involve ongoing responsibilities related to medication management and lifestyle monitoring [14–16], while illnesses with episodic exacerbations, including chronic obstructive pulmonary disease (COPD) and severe mental illness, expose caregivers to unpredictable care demands [17–19]. Across these conditions, approximately one-third to one-half of family caregivers experience significant burden [20, 21]. Caregiving responsibilities can also impose considerable economic and psychosocial strain. Many caregivers experience income loss, reduced workforce participation, and substantial out-of-pocket expenses, sometimes accounting for 15–20% of household income [22, 23]. These pressures negatively affect caregivers’ mental health and quality of life and may also influence patient outcomes, as caregiver psychological distress has been associated with poorer patient adherence to treatment [2]. Such challenges highlight the need for accessible education, resources, and professional support for family caregivers [24]. Traditionally, face-to-face interventions to reduce caregiver burden include psychoeducational programs, skills training, CBT, counselling, and supportive services such as peer groups and adult day care [25–28]. The core components, such as education, skills training, and psychological support, aim to empower caregivers and reduce burden [26, 27], but access is often limited by logistical, time, and geographic barriers. To overcome these challenges, tele-interventions, defined as healthcare services delivered remotely via digital technologies such as videoconferencing, mobile applications, and web-based platforms [29], have emerged as a promising alternative [24]. Tele-interventions provide an appropriate substitute by translating traditional elements into digital formats without altering their essential content [3]. Within these models, nurses frequently play a central role in assessing caregiver needs, providing tailored education, coordinating care, and delivering follow-up support [30–33]. Many tele-interventions incorporate psychoeducational or cognitive-behavioural components aimed at enhancing caregiving skills and self-efficacy, which may help reduce caregiver burden and improve well-being [12, 34]. Additionally, tele-interventions have been associated with improved quality of life and more favourable healthcare utilization outcomes, including shorter hospital stays and smoother transitions from hospital to home [3, 12, 35]. Despite these potential benefits, evidence regarding the effectiveness and acceptance of tele-interventions among caregivers remains inconsistent. Variability across disease groups, intervention modalities, and levels of professional involvement complicates comparisons across studies. Moreover, factors such as digital literacy and unequal access to technology may influence engagement and effectiveness, potentially increasing cognitive workload or widening disparities among caregivers [36–39]. Evidence suggests that nurse‑led tele‑interventions can reduce technology‑related stress, enhance caregiver confidence, and improve engagement, thereby strengthening the overall effectiveness of digital care models [31]. Nonetheless, variations in intervention design, intensity, professional involvement, and disease context continue to complicate comparisons across studies. Several systematic reviews have examined tele-interventions for caregivers in specific disease contexts, most commonly dementia [9] or cancer [40, 41]. Although these reviews provide valuable disease-specific insights, their focus on single conditions limits generalizability across caregiving contexts and intervention modalities. In addition, many earlier reviews included heterogeneous study designs or did not quantitatively synthesize outcomes related specifically to caregiver burden. Given the rapid expansion of tele-interventions technologies, particularly following the COVID-19 pandemic, an updated and broader synthesis of evidence is needed. Despite, the expanding body of primary research, there remains no comprehensive synthesis evaluating the effectiveness of tele-interventions on family caregiver burden across diverse disease populations and delivery modalities. Accordingly, this systematic review and meta-analysis aimed to evaluate the effect of tele-interventions on family caregiver burden across a broad range of disorders and intervention types. Methods The present study is a systematic review and meta-analysis conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA, 2020) guidelines. Protocol registration The protocol of this study was registered in PROSPERO with registration number CRD42025648217. Search strategy An initial scoping search without date restrictions was conducted to explore the breadth and evolution of the literature. This preliminary assessment indicated that many studies published before 2015 focused on early-stage or low-technology tele-interventions (e.g., telephone-based support, simple web portals, or non-interactive platforms). These approaches differ substantially from contemporary digital health interventions in terms of functionality, interactivity, accessibility, and user engagement. Given the widespread adoption of smartphones, mobile health applications, broadband internet, and more advanced tele‑interventions infrastructures after approximately 2015 [42], a priori date limits were applied to enhance the methodological and technological comparability of included studies and to ensure relevance to current digital health practice. Accordingly, the final systematic search was restricted to studies published between January 1, 2015 and January 1, 2025. The systematic search was conducted on February 25, 2025, using Web of Science, Scopus, and PubMed as principal bibliographic databases. Google Scholar was also searched as a supplementary source to enhance coverage. The search strategy used combinations of MeSH terms and keywords related to “tele-interventions” and “caregiver burden,” such as: (caregiver OR caregivers OR carer OR carers OR family OR families OR spouse OR spouses OR “care giver” OR “care givers” OR relative OR relatives OR partner OR partners carer OR caregiver OR family member OR relatives OR informal caregivers) AND (burden OR burnout) AND (telemedicine OR telenursing OR telecare OR telecommunicat* OR telepalliative OR telehospice OR “remote consult*” OR telemonitor* OR teleconsult* OR telehealth* OR app OR apps OR application OR “virtual medicine” OR e-health OR m-health OR mobile OR cellphone OR online OR internet OR website OR smartphone OR web-based OR Video OR technolog* OR digital* OR electronic* OR “web-based intervention” OR ‘technology intervention’ OR “internet intervention” OR “tele-health intervention”). Filters were used according to databases’ options: MeSH terms/keywords included in title and abstract, RCTs, free full text, English language, and limited to the 2015–2025 period. The complete search strategies for each databases are available in Supplementary Material 1. Eligibility criteria The eligibility criteria were defined according to the PICO (Population, Intervention, Comparison, Outcome) framework [43], which is commonly used to structure research questions and eligibility criteria in systematic reviews. Eligible studies included clinical trials that investigated the effects of tele‑interventions (Intervention) on caregiver burden (Outcome) among family caregivers of patients (Population). The comparison group (Comparison) comprised family caregivers receiving usual care or standard support. Only studies published in English with full‑text availability were considered. Studies were excluded if they were non‑interventional, qualitative in design, unavailable in full text, in languages other than English, or assessed outcomes unrelated to caregiver burden. In line with our focus on estimating the effects of tele‑interventions using methodologically robust evidence, studies judged to be at high risk of bias were excluded a priori. This decision aimed to enhance internal validity and reduce the likelihood of spurious effect estimates driven by serious methodological flaws, particularly in the presence of substantial clinical and methodological heterogeneity. The implications of this decision were further considered in the GRADE (Grading of Recommendations Assessment, Development, and Evaluation) assessment and sensitivity analyses. Study selection and data extraction All records obtained from the searches were imported into EndNote version 8.0, and duplicate studies were systematically eliminated. Two researchers independently screened the remaining records for relevance based on titles and abstracts, followed by full-text assessments of potentially eligible studies. To further identify relevant literature, the reference lists of all included studies were reviewed manually. In addition, to access gray literature, the reference lists of relevant systematic reviews and meta‑analyses were manually examined to identify potentially eligible randomized controlled trials. However, no additional eligible studies were identified, as the retrieved articles either did not report caregiver burden as an outcome or did not involve tele-interventions. The extracted data included authors, publication year, country, study design, participants, sample size in both intervention and control groups, type and duration of interventions, type of control group, follow‑up period, outcome measurement tools, and main findings. Disagreements at any stage were resolved by consensus, with the involvement of a third researcher. Quality assessment The methodological quality of the randomized controlled trials included in the review was appraised through the assessment of risk of bias using the Cochrane Risk of Bias version 2.0 (RoB-2) tool. The RoB-2 tool is a widely recommended instrument for appraising the internal validity of randomized trials and is frequently employed as a surrogate measure for methodological quality in systematic reviews. The RoB-2 tool evaluates bias across the following domains: random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and other bias. Each of these domains is evaluated as “low risk”, “high risk”, or “unclear risk” of bias [44]. The overall risk of bias was determined based on domain‑level judgments under Cochrane guidance. Studies were classified as high risk of bias if at least one key domain was judged as high risk. Two reviewers independently assessed the risk of bias for each included study. Any disagreements were resolved through discussion and, when necessary, consultation with a third reviewer. Risk of bias assessments were conducted using Review Manager (RevMan; Cochrane Community, Oxford, UK), version 5.0. Studies judged to be at high risk of bias were excluded from the review. Publication bias was assessed through a variety of complementary methodologies. A visual examination of funnel plots was conducted to investigate potential asymmetry. The application of Egger’s regression test facilitated the statistical assessment of small-study effects, adhering to a significance level of 0.05. In addition, Rosenberg’s Fail‑Safe N (N-fs) method was employed, which appraises the robustness of statistically significant outcomes by estimating the number of unpublished null studies required to render the pooled effect non‑significant. Rosenberg’s Fail‑Safe N (N-fs), a more conservative alternative to traditional fail-safe approaches [45, 46], was applied. The tolerance level was calculated as 5k + 10. To further examine the potential impact of publication bias, the trim‑and‑fill method was applied to estimate the number of potentially missing studies and to evaluate whether adjustment for such studies would alter the magnitude or direction of the pooled effect size. To examine publication bias, StataMP software version 17 was used. All statistical analyses were performed with a significance threshold established at 0.05. Assessment of the certainty of evidence The certainty of the evidence was appraised utilizing the GRADE approach [47]. The quality of evidence was evaluated for the primary outcome (Caregiver Burden) across five domains: risk of bias, inconsistency, indirectness, imprecision, and publication bias. The certainty was classified into four hierarchical levels: “High”, “Moderate”, “Low”, or “Very Low”. The GRADEpro GDT software (Guideline Development Tool) was employed to generate the Summary of Findings (SoF) table. Data analysis A descriptive synthesis was conducted to summarize study characteristics and intervention features. Quantitative data synthesis was then performed through meta-analysis using a random-effects model to estimate pooled effect sizes. A leave‑one‑out sensitivity analysis was performed by sequentially excluding each study and recalculating the pooled standardized mean difference along with 95% confidence interval to assess the impact of individual studies on the overall estimate. A standardized mean difference (SMD) with a 95% confidence interval was also applied to compare the effect sizes of different interventions, ensuring accurate comparisons between studies with diverse scales. The heterogeneity of studies was assessed using the I-squared (I²) index to determine the degree of variation among studies. Subgroup analyses were conducted according to disorder type, as well as the type and content of the interventions. Forest plots were computed for the visual display of results. The meta-analysis was conducted using Stata 17. Results Study selection The database search of Web of Science, Scopus, and PubMed identified 741 records. An additional 373 records were identified through supplementary sources, including Google Scholar. These supplementary records were screened at the title and abstract level; however, no additional studies meeting the predefined inclusion criteria were identified. After removing duplicates (n = 72), 669 records remained for title and abstract screening. Following this screening process, 510 records were excluded due to irrelevance to the review topic. The full texts of the remaining 159 articles were assessed for eligibility. In this step, 127 articles were excluded. The reasons included unavailability of full text (29 articles), a different language (1 Spanish article), inappropriate study type (97 articles), and high risk of bias (9 articles). Ultimately, 23 studies met the inclusion criteria and were included in the final analysis. The study identification, screening, eligibility assessment, and inclusion process are illustrated in the PRISMA 2020 flow diagram (Fig. 1). Fig. 1Flowchart of study selection based on PRISMA 2020 Flowchart of study selection based on PRISMA 2020 Assessment of risk of bias The risk of bias in the 23 randomized controlled trials included in this systematic review and meta-analysis is illustrated in Figs. 2 and 3. Figure 2 indicates that all studies had a low risk of bias in the randomization process (100%). Most studies showed a low risk of bias for allocation concealment (47.82%). The review of 17 studies (73.91%) revealed a low risk of bias regarding blinding of participants and personnel. However, the risk of bias was unclear in most studies for several domains. These included blinding of outcome assessment (56.52%), incomplete outcome data (52.17%), and selective reporting (56.52%). Regarding other biases, more than half of the studies (56.52%) showed a low risk (Fig. 3). Fig. 2Risk of bias graph of the included studies Risk of bias graph of the included studies Fig. 3Risk of bias summary in the included studies Risk of bias summary in the included studies Visual inspection of the funnel plot suggested asymmetry, particularly among studies with larger standard errors. This pattern raised the possibility of small-study effects rather than definitive publication bias (Supplementary Material 2). Consistent with this observation, Egger’s regression test indicated statistically significant asymmetry (p < 0.05). However, substantial between-study heterogeneity was observed (I² = 94.50%). Therefore, the interpretation of funnel plot asymmetry and Egger’s test should be approached with caution. Both methods are sensitive to heterogeneity and may yield false-positive indications of asymmetry. To further explore the potential influence of small-study effects on the pooled estimate, Rosenberg’s fail-safe N (N-fs) and the trim-and-fill method were applied. The fail-safe N was estimated at approximately 114. This value exceeds the tolerance threshold of 95 for the 17 included studies. This finding suggests that a large number of null-effect studies would be required to overturn the statistical significance of the pooled result. Application of the trim-and-fill procedure did not identify any missing studies, and the adjusted pooled effect size remained unchanged (SMD = -0.020, 95% CI -0.108 to 0.069). This finding does not necessarily indicate the absence of small-study effects. Instead, it may reflect the limited sensitivity of the trim-and-fill method under conditions of extreme heterogeneity and imprecision. Taken together, these analyses suggest that although small-study effects cannot be fully excluded, the overall pooled estimate was not materially altered by statistical adjustment for funnel plot asymmetry. Study characteristics Table 1 presents key characteristics and findings of included studies. Study designs included one exploratory RCT [48], one field trial [49], 16 single-centre RCTs [50–65], four multi-centre RCTs [66–69], and one cluster RCT [70]. Sample sizes ranged from 58 to 377 participants, with a total sample of 3,678 participants. The setting of the study was the United States in 9 of the studies (39.13%). 17 studies (73.91%) were published in 2022 or later, indicating the contemporary relevance of tele-interventions for family caregiver burden. Regarding control conditions, usual care emerged as the most prevalent comparison, utilized in 20 studies (86.96%), reflecting standard clinical practice without structured tele-interventions. Attention control conditions were implemented in two studies (8.70%). In these studies, participants received non-therapeutic contact to control for time and attention effects. One study (4.35%) used a waitlist control, allowing participants to access the intervention after the follow-up period. The duration of the intervention and the subsequent follow-up period varied significantly. Tele‑interventions ranged from 30 days [62] to 12 months [60]. One study lacked a follow-up period [49], whereas the longest follow-up lasted 18 months after the tele-interventions’ completion [53]. There is substantial variability in the duration of intervention and follow-up period. All included studies were RCTs, and the outcome variable of caregiver burden was measured in all of them (100%). Table 1Key characteristics and findings of included studiesNo.Authors (year)CountryDesignParticipants(Family caregivers of patient with)Sample sizeInterventionIntervention durationIntervention content/typeControl conditionFollow-upOutcome measurement (Tools)Caregiver burdena(Control group)Caregiver burdena(Intervention group)Main resultControl groupIntervention groupBaselinePost testBaselinePost test1Bahadori et al. (2023)IranRandomized Controlled Field TrialCOPD3535A smartphone-based’ pulmonary rehabilitation education through an android application6 weeksRehabilitative/Mobile appUsual careImmediately after the study interventionZBI50.77 (10.64)51.57 (7.32)49.77 (7.65)37.31 (6.95)Significantly decreasing caregiver burden (p < 0.05)2Barnes et al. (2024)U.S.RCTCognitive impairment4354An online, livestream, mind-body group movement program called Moving Together1-hour sessions held 2 days per week for a total of 12 weeksSupportive/Web-basedWaitlist control24 weeks after the baselineZBI8.9(5.2)9.2(5.3)8.5(5.1)10.4(5.4)No differences in caregivers burden (P > 0.05)3Chae & Lee (2024)South KoreaRCTDementia3030An online comprehensive cognitive training program called Smart Brain5 sessions per week, delivered through a tablet application for cognitive training, emotional support, and healthcare activities for 8 weeksEducational/Mobile appUsual careweek 4 and week 8 after the interventionZBI38.50 (20.56)49.96 (23.20)32.38 (16.62)31.71 (14.16)Effective in improving caregiver burden (P < 0.05)4Dröes et al. (2019)NetherlandsRCTDementia189iMCSP group: 39MCSP group: 54iMCSP consisted of three components:1) DemenTalent, 2) Dementelcoach (telephone coaching), and 3) STAR e-Learning (an online course)6 monthsSupportive/Telephone-basedUsual care6 months after baselineNPI burden subscale4.61 (2.17)4.66 (2.19)iMCSP group: 11.57 (5.93)MCSP group: 9.29 (8.28)iMCSP group: 7.67 (5.77)MCSP group: 10.03 (8.13)No significant differences in caregivers burden (p > 0.05)5Fitzgeraldson et al. (2024)AustraliaRCTDepression or anxiety6661A four-module, self-paced, and online early intervention support program called “Minds Together” with a social forum component (resembled a Facebook feed where participants could post, comment, react, and upload photos)10 weeksEducational/Social networksUsual care3 months after the initial program accessZBI-7.57-1.72b-8.26-1.48-9.10-1.73-6.95-4.50No statistically significant difference in caregiving burden (p > 0.05)6Ghahramani et al. (2022)U.S.RCTCKD28Face-to-face group: 29Online group: 29A structured peer mentoring program, with two modes of delivery: face-to-face and online6 monthsMonitoring/Telephone-basedUsual care18-month after the intervention’s completionZBI22.1 (9.3)22.0 (9.9)Face-to-face group: 21.1 (9.4)Online group: 23.6 (12.1)Face-to-face group: 14.0 (7.8)Online group: 15.2 (9.9)Statistically significant decrease in the ZBI score (p < 0.05)7Griffin et al. (2024)U.S.RCTreceiving inpatient palliative care out of the hospital183184TPC program that involved 2 video visits between family caregivers and palliative care nursesTwo video visits during the first four weeks post-discharge, followed by one visit per week for the subsequent four weeks, totaling a maximum of 11 sessionsMonitoring/Web-basedUsual careweek 2 and week 8 after dischargeBCOS-R4.5 (0.6)0.10 (0.16)c4.4 (0.6)-0.002 (0.16)cNo differences in caregiver burden (P > 0.05)8Heckel et al. (2018)AustraliaMulticenter RCTCancer108108A telephone outcalls program (three times)4 monthsSupportive/Telephone-basedUsual care6 monthsZBI18.05 (1.17d)17.20 (1.33d)18.99(1.17d)18.46(1.31d)No effect on the caregiver burden (p > 0.05)9Hepburn et al. (2022)U.S.RCTAlzheimerAttention control group: 111Waitlist control group: 5496The intervention was Tele-Savvy: an online adaptation of the successful in-person psychoeducation program (seven weekly 75- to 90-minute synchronous Zoom-based meetings of small groups led by a facilitator)43 daysEducational/Social networksAttention control6 monthsZBIAttention control group: 35.78 (14.66)Waitlist control group: 35.94 (15.99)Attention control group: 35.91 (12.96)Waitlist control group: 37.48 (17.31)37.34 (14.66)35.54 (13.63)No significantly difference in caregivers burden (p > 0.05)10Hoek et al. (2017)NetherlandsRCTCancer3638A weekly and prescheduled tele-consultation with a hospital-based specialist palliative care consultation team (consisting of videoconferencing, initially using a desktop computer, later replaced by tablet computers using FaceTime)12 weeksCounselling/telephone-basedUsual care12 weeks after baselineEDIZ15.54 (6.46)15.00 (7.10)15.78 (5.87)15.95 (8.25)No effect on the caregiver burden (p > 0.05)11Khouban-Shargh et al. (2024)IranRCTCKD3131Stress management trainingconsisted of six online group sessions on the WhatsApp social network2 monthsEducational/Social networksUsual care2 weeks after completion of the interventionZBI49.1 (6.0)49.0 (5.8)50.8 (4.9)45.9 (4.1)Significant decrease in caregivers burden (p < 0.05)12Lech et al. (2023)GermanyCluster RCTDementia3454The DemTab App: a “tablet-based intervention to improve adherence to the German S3 Dementia Guideline in primary care9 monthsSupportive/Mobile appUsual care9 months after the intervention beganHPS12.9 (7.4)3.67-26.64b11.6 (7.6)-3.27-5.26bNo significant differences in caregivers burden (p > 0.05)13Montoro-Rodriguez et al. (2024)U.S.RCTSignificant neurocognitive disorders, such as dementia4760An online psychoeducational program called Caregiver TLCh (consisted of six sessions delivered in a virtual format via online meetings, each session lasted 90 min)8 weeksRehabilitative/Social networksUsual careafter 8 weeks when the intervention or time on the waitlist was completedZBI11.61 (4.60d)11.95 (4.40d)14.03 (4.82d)12.35 (4.89d)Significant changes in levels of burden (p < 0.05)14Mooney et al. (2024)U.S.Multisite RCTCancer173159An automated digital caregiver coaching tool known as Symptom Care at Home (SCH)until the patient’s death or for a maximum of 6 monthsEducational/Web-basedUsual care6 months after the patient’s death (week 1 and week 8)The Caregiver Burden Measure0.71-0.83 g0.78–0.950.71–0.830.47–0.80Significantly lower in caregivers burden (P < 0.05)15Mueser et al. (2022)U.S.RCTSchizophrenia7573A telehealth-based family psychoeducation and skills training (16 live web-based sessions via MyHealios, an online platform with live web-based sessions led by clinicians)6 monthsEducational/Web-basedUsual care12 months after the baseline assessmentIEQ29.7 (17.3)33.2 (18.2)30.8 (16.4)29.0 (16.1)No significant differences in caregivers burden (p > 0.05)16Piette et al. (2015)U.S.RCTHeart failure180189Mobile health services (mHealth) delivered via automated telephone calls and email summaries12 monthsSupportive/Telephone-basedAttention control12 months after the interventionCaregiver Strain Index3.1 (2.7)Not mentioned3.3 (3.1)-0.0970- -0.0204bSignificantly lower in caregivers burden (P < 0.05)17Schuit et al. (2022)NetherlandsRCTCancer3028an eHealth self-management application called “Oncokompas” (consists of three steps: measure, learn, and act)3 monthsEducational/Mobile appUsual care3 monthsCSI4.2 (3.2)4.3 (3.0)2.8 (3.4)2.5 (2.6)No significant effects on caregiver burden (p > 0.05)18Tinoco-Camarenaet al. (2023)SpainRCTChronic complex disease4343An online nursing intervention (through a video call using the Zoom platform) called “dialogue circles” consisted of three dialogue circles per group. Each circle lasted approximately 90 min30 daysEducational/Social networksUsual care30 days after the intervention’s completionZBI22.00 (5.00)25.00 (7.00)26.00 (4.00)20.00 (6.00)Significant improvement (decrease) in ZBI-7 scores (p < 0.05)19Tremont et al. (2015)U.S.RCTDementia11713316 telephone contactsThe initial telephone contacts had standardized durations of approximately 60 min, while follow-up contacts varied from 15–30 min6 monthsSupportive/Telephone-basedUsual care6 months telephone-based follow-upZBI38.82 (14.63)37.17 (13.93)38.61 (13.98)35.95 (14.34)No significant differences for caregiver burden (p > 0.05)20Wang et al. (2024)ChinaRCTStroke4545Multiple courses delivered via video: one course providing stroke-related information, two courses on daily life nursing, and six courses on stroke rehabilitation trainingFrom admission to dischargeEducational/Web-basedUsual care15 weeks after dischargeZBI41.56 (7.34)35.77 (4.67)39.23 (7.23)31.98 (6.43)Decreased significantly in caregivers burden (p < 0.05)21Welten et al. (2024)NetherlandsMulticenter RCTABIsuch as stroke and other injuries6159A blended care program called CARE4Carer (consisted of 9 online sessions and 2 face-to-face consultations with a social worker)20 weeksCounselling/Web-basedUsual care20 weeksSRB50.3249.0847.3743.78No statistically significant differences in caregivers burden(P > 0.05)22Xie et al. (2024)ChinaRCTDementia3636Web-based caregiver skill training(18 lessons (totaling 270 min) delivered in a web-based format)6 monthsEducational/Web-basedUsual care6 months after the interventionZBI19.85 (11.55)30.30 (13.05)24.48 (10.77)10.97 (5.43)Significant reducing caregivers burden (P < 0.05)23Zhou et al. (2019)ChinaMulticenter RCTRecent acute stroke128118A novel nurse-led, caregiver-delivered model of stroke rehabilitation (by a custom-designed smartphone application)8 weeks after dischargeRehabilitative/Mobile appUsual care6 monthsCBI45.2 (3.7)74.0 (4.6)42.8 (4.0)71.6 (4.7)No effect on the caregiver burden (p > 0.05)a Mean (Standard Deviation)b Confidence Intervalc Mean Difference (SD)d Standard ErrorCOPD: Chronic Obstructive Pulmonary Disease; ZBI: Zarit Burden Interview; iMCSP: individualized Meeting Centers Support Program; NPI: Neuropsychiatric Inventory; TPC: Technology-enhanced transitional Palliative Care; CKD: Chronic Kidney Disease; BCOS-R: Bakas Caregiving Outcomes Scale-Revised; EDIZ: Self-Perceived Burden from Informal Care; HPS: Burden Scale for Family Caregivers; IEQ: Involvement Evaluation Questionnaire; CSI: Caregiver Strain Index; ABI: Acquired Brain Injury; SRB: Self-Rated Burden; CBI: Caregiver Burden Inventory Table 1Key characteristics and findings of included studiesNo.Authors (year)CountryDesignParticipants(Family caregivers of patient with)Sample sizeInterventionIntervention durationIntervention content/typeControl conditionFollow-upOutcome measurement (Tools)Caregiver burdena(Control group)Caregiver burdena(Intervention group)Main resultControl groupIntervention groupBaselinePost testBaselinePost test1Bahadori et al. (2023)IranRandomized Controlled Field TrialCOPD3535A smartphone-based’ pulmonary rehabilitation education through an android application6 weeksRehabilitative/Mobile appUsual careImmediately after the study interventionZBI50.77 (10.64)51.57 (7.32)49.77 (7.65)37.31 (6.95)Significantly decreasing caregiver burden (p < 0.05)2Barnes et al. (2024)U.S.RCTCognitive impairment4354An online, livestream, mind-body group movement program called Moving Together1-hour sessions held 2 days per week for a total of 12 weeksSupportive/Web-basedWaitlist control24 weeks after the baselineZBI8.9(5.2)9.2(5.3)8.5(5.1)10.4(5.4)No differences in caregivers burden (P > 0.05)3Chae & Lee (2024)South KoreaRCTDementia3030An online comprehensive cognitive training program called Smart Brain5 sessions per week, delivered through a tablet application for cognitive training, emotional support, and healthcare activities for 8 weeksEducational/Mobile appUsual careweek 4 and week 8 after the interventionZBI38.50 (20.56)49.96 (23.20)32.38 (16.62)31.71 (14.16)Effective in improving caregiver burden (P < 0.05)4Dröes et al. (2019)NetherlandsRCTDementia189iMCSP group: 39MCSP group: 54iMCSP consisted of three components:1) DemenTalent, 2) Dementelcoach (telephone coaching), and 3) STAR e-Learning (an online course)6 monthsSupportive/Telephone-basedUsual care6 months after baselineNPI burden subscale4.61 (2.17)4.66 (2.19)iMCSP group: 11.57 (5.93)MCSP group: 9.29 (8.28)iMCSP group: 7.67 (5.77)MCSP group: 10.03 (8.13)No significant differences in caregivers burden (p > 0.05)5Fitzgeraldson et al. (2024)AustraliaRCTDepression or anxiety6661A four-module, self-paced, and online early intervention support program called “Minds Together” with a social forum component (resembled a Facebook feed where participants could post, comment, react, and upload photos)10 weeksEducational/Social networksUsual care3 months after the initial program accessZBI-7.57-1.72b-8.26-1.48-9.10-1.73-6.95-4.50No statistically significant difference in caregiving burden (p > 0.05)6Ghahramani et al. (2022)U.S.RCTCKD28Face-to-face group: 29Online group: 29A structured peer mentoring program, with two modes of delivery: face-to-face and online6 monthsMonitoring/Telephone-basedUsual care18-month after the intervention’s completionZBI22.1 (9.3)22.0 (9.9)Face-to-face group: 21.1 (9.4)Online group: 23.6 (12.1)Face-to-face group: 14.0 (7.8)Online group: 15.2 (9.9)Statistically significant decrease in the ZBI score (p < 0.05)7Griffin et al. (2024)U.S.RCTreceiving inpatient palliative care out of the hospital183184TPC program that involved 2 video visits between family caregivers and palliative care nursesTwo video visits during the first four weeks post-discharge, followed by one visit per week for the subsequent four weeks, totaling a maximum of 11 sessionsMonitoring/Web-basedUsual careweek 2 and week 8 after dischargeBCOS-R4.5 (0.6)0.10 (0.16)c4.4 (0.6)-0.002 (0.16)cNo differences in caregiver burden (P > 0.05)8Heckel et al. (2018)AustraliaMulticenter RCTCancer108108A telephone outcalls program (three times)4 monthsSupportive/Telephone-basedUsual care6 monthsZBI18.05 (1.17d)17.20 (1.33d)18.99(1.17d)18.46(1.31d)No effect on the caregiver burden (p > 0.05)9Hepburn et al. (2022)U.S.RCTAlzheimerAttention control group: 111Waitlist control group: 5496The intervention was Tele-Savvy: an online adaptation of the successful in-person psychoeducation program (seven weekly 75- to 90-minute synchronous Zoom-based meetings of small groups led by a facilitator)43 daysEducational/Social networksAttention control6 monthsZBIAttention control group: 35.78 (14.66)Waitlist control group: 35.94 (15.99)Attention control group: 35.91 (12.96)Waitlist control group: 37.48 (17.31)37.34 (14.66)35.54 (13.63)No significantly difference in caregivers burden (p > 0.05)10Hoek et al. (2017)NetherlandsRCTCancer3638A weekly and prescheduled tele-consultation with a hospital-based specialist palliative care consultation team (consisting of videoconferencing, initially using a desktop computer, later replaced by tablet computers using FaceTime)12 weeksCounselling/telephone-basedUsual care12 weeks after baselineEDIZ15.54 (6.46)15.00 (7.10)15.78 (5.87)15.95 (8.25)No effect on the caregiver burden (p > 0.05)11Khouban-Shargh et al. (2024)IranRCTCKD3131Stress management trainingconsisted of six online group sessions on the WhatsApp social network2 monthsEducational/Social networksUsual care2 weeks after completion of the interventionZBI49.1 (6.0)49.0 (5.8)50.8 (4.9)45.9 (4.1)Significant decrease in caregivers burden (p < 0.05)12Lech et al. (2023)GermanyCluster RCTDementia3454The DemTab App: a “tablet-based intervention to improve adherence to the German S3 Dementia Guideline in primary care9 monthsSupportive/Mobile appUsual care9 months after the intervention beganHPS12.9 (7.4)3.67-26.64b11.6 (7.6)-3.27-5.26bNo significant differences in caregivers burden (p > 0.05)13Montoro-Rodriguez et al. (2024)U.S.RCTSignificant neurocognitive disorders, such as dementia4760An online psychoeducational program called Caregiver TLCh (consisted of six sessions delivered in a virtual format via online meetings, each session lasted 90 min)8 weeksRehabilitative/Social networksUsual careafter 8 weeks when the intervention or time on the waitlist was completedZBI11.61 (4.60d)11.95 (4.40d)14.03 (4.82d)12.35 (4.89d)Significant changes in levels of burden (p < 0.05)14Mooney et al. (2024)U.S.Multisite RCTCancer173159An automated digital caregiver coaching tool known as Symptom Care at Home (SCH)until the patient’s death or for a maximum of 6 monthsEducational/Web-basedUsual care6 months after the patient’s death (week 1 and week 8)The Caregiver Burden Measure0.71-0.83 g0.78–0.950.71–0.830.47–0.80Significantly lower in caregivers burden (P < 0.05)15Mueser et al. (2022)U.S.RCTSchizophrenia7573A telehealth-based family psychoeducation and skills training (16 live web-based sessions via MyHealios, an online platform with live web-based sessions led by clinicians)6 monthsEducational/Web-basedUsual care12 months after the baseline assessmentIEQ29.7 (17.3)33.2 (18.2)30.8 (16.4)29.0 (16.1)No significant differences in caregivers burden (p > 0.05)16Piette et al. (2015)U.S.RCTHeart failure180189Mobile health services (mHealth) delivered via automated telephone calls and email summaries12 monthsSupportive/Telephone-basedAttention control12 months after the interventionCaregiver Strain Index3.1 (2.7)Not mentioned3.3 (3.1)-0.0970- -0.0204bSignificantly lower in caregivers burden (P < 0.05)17Schuit et al. (2022)NetherlandsRCTCancer3028an eHealth self-management application called “Oncokompas” (consists of three steps: measure, learn, and act)3 monthsEducational/Mobile appUsual care3 monthsCSI4.2 (3.2)4.3 (3.0)2.8 (3.4)2.5 (2.6)No significant effects on caregiver burden (p > 0.05)18Tinoco-Camarenaet al. (2023)SpainRCTChronic complex disease4343An online nursing intervention (through a video call using the Zoom platform) called “dialogue circles” consisted of three dialogue circles per group. Each circle lasted approximately 90 min30 daysEducational/Social networksUsual care30 days after the intervention’s completionZBI22.00 (5.00)25.00 (7.00)26.00 (4.00)20.00 (6.00)Significant improvement (decrease) in ZBI-7 scores (p < 0.05)19Tremont et al. (2015)U.S.RCTDementia11713316 telephone contactsThe initial telephone contacts had standardized durations of approximately 60 min, while follow-up contacts varied from 15–30 min6 monthsSupportive/Telephone-basedUsual care6 months telephone-based follow-upZBI38.82 (14.63)37.17 (13.93)38.61 (13.98)35.95 (14.34)No significant differences for caregiver burden (p > 0.05)20Wang et al. (2024)ChinaRCTStroke4545Multiple courses delivered via video: one course providing stroke-related information, two courses on daily life nursing, and six courses on stroke rehabilitation trainingFrom admission to dischargeEducational/Web-basedUsual care15 weeks after dischargeZBI41.56 (7.34)35.77 (4.67)39.23 (7.23)31.98 (6.43)Decreased significantly in caregivers burden (p < 0.05)21Welten et al. (2024)NetherlandsMulticenter RCTABIsuch as stroke and other injuries6159A blended care program called CARE4Carer (consisted of 9 online sessions and 2 face-to-face consultations with a social worker)20 weeksCounselling/Web-basedUsual care20 weeksSRB50.3249.0847.3743.78No statistically significant differences in caregivers burden(P > 0.05)22Xie et al. (2024)ChinaRCTDementia3636Web-based caregiver skill training(18 lessons (totaling 270 min) delivered in a web-based format)6 monthsEducational/Web-basedUsual care6 months after the interventionZBI19.85 (11.55)30.30 (13.05)24.48 (10.77)10.97 (5.43)Significant reducing caregivers burden (P < 0.05)23Zhou et al. (2019)ChinaMulticenter RCTRecent acute stroke128118A novel nurse-led, caregiver-delivered model of stroke rehabilitation (by a custom-designed smartphone application)8 weeks after dischargeRehabilitative/Mobile appUsual care6 monthsCBI45.2 (3.7)74.0 (4.6)42.8 (4.0)71.6 (4.7)No effect on the caregiver burden (p > 0.05)a Mean (Standard Deviation)b Confidence Intervalc Mean Difference (SD)d Standard ErrorCOPD: Chronic Obstructive Pulmonary Disease; ZBI: Zarit Burden Interview; iMCSP: individualized Meeting Centers Support Program; NPI: Neuropsychiatric Inventory; TPC: Technology-enhanced transitional Palliative Care; CKD: Chronic Kidney Disease; BCOS-R: Bakas Caregiving Outcomes Scale-Revised; EDIZ: Self-Perceived Burden from Informal Care; HPS: Burden Scale for Family Caregivers; IEQ: Involvement Evaluation Questionnaire; CSI: Caregiver Strain Index; ABI: Acquired Brain Injury; SRB: Self-Rated Burden; CBI: Caregiver Burden Inventory Key characteristics and findings of included studies Rehabilitative/ Mobile app Supportive/ Web-based 8.9 (5.2) 9.2 (5.3) 8.5 (5.1) 10.4 (5.4) Educational/ Mobile app iMCSP group: 39 MCSP group: 54 Supportive/ Telephone-based iMCSP group: 11.57 (5.93) MCSP group: 9.29 (8.28) iMCSP group: 7.67 (5.77) MCSP group: 10.03 (8.13) Educational/ Social networks Face-to-face group: 29 Online group: 29 Monitoring/ Telephone-based Face-to-face group: 21.1 (9.4) Online group: 23.6 (12.1) Face-to-face group: 14.0 (7.8) Online group: 15.2 (9.9) Monitoring/ Web-based Supportive/ Telephone-based 18.99 (1.17d) 18.46 (1.31d) Attention control group: 111 Waitlist control group: 54 Educational/ Social networks Attention control group: 35.78 (14.66) Waitlist control group: 35.94 (15.99) Attention control group: 35.91 (12.96) Waitlist control group: 37.48 (17.31) Counselling/ telephone-based Stress management training consisted of six online group sessions on the WhatsApp social network Educational/ Social networks Supportive/ Mobile app Rehabilitative/ Social networks Educational/ Web-based Educational/ Web-based Supportive/ Telephone-based Educational/ Mobile app Educational/ Social networks 16 telephone contacts The initial telephone contacts had standardized durations of approximately 60 min, while follow-up contacts varied from 15–30 min Supportive/ Telephone-based Educational/ Web-based ABI such as stroke and other injuries Counselling/ Web-based No statistically significant differences in caregivers burden (P > 0.05) Web-based caregiver skill training (18 lessons (totaling 270 min) delivered in a web-based format) Educational/ Web-based Rehabilitative/ Mobile app a Mean (Standard Deviation) b Confidence Interval c Mean Difference (SD) d Standard Error COPD: Chronic Obstructive Pulmonary Disease; ZBI: Zarit Burden Interview; iMCSP: individualized Meeting Centers Support Program; NPI: Neuropsychiatric Inventory; TPC: Technology-enhanced transitional Palliative Care; CKD: Chronic Kidney Disease; BCOS-R: Bakas Caregiving Outcomes Scale-Revised; EDIZ: Self-Perceived Burden from Informal Care; HPS: Burden Scale for Family Caregivers; IEQ: Involvement Evaluation Questionnaire; CSI: Caregiver Strain Index; ABI: Acquired Brain Injury; SRB: Self-Rated Burden; CBI: Caregiver Burden Inventory Caregiver burden in the included studies reflected self‑reported perceived caregiving strain, operationalized using validated burden instruments. The Zarit Burden Interview (ZBI) and its short forms were the most commonly used measures (13 studies, 56.52%). These instruments assess perceived role burden, emotional distress, and the impact of caregiving on personal life (Table 1). The remaining studies (43.48%) used other validated instruments to assess caregiver burden. These included the Caregiver Strain Index (CSI) [60, 61], Self‑Rated Burden (SRB) [68], Involvement Evaluation Questionnaire (IEQ) [59], and the Brief Caregiver Outcome Scale-Revised (BCOS-R) [54]. One study used disease‑specific burden subscales such as the NPI‑Burden [48]. These instruments differ in format, number of items, and domain focus. However, they share a common focus on the subjective and psychosocial impact of caregiving rather than objective caregiving workload. This indicates conceptual consistency across the measures [71]. Notwithstanding this conceptual alignment, variability in scale structure, domain coverage, and sensitivity to change was observed. ZBI‑based instruments typically yield a continuous composite score and represent a broader range of caregiver burden’ domains. In contrast, shorter tools such as the CSI or BCOS emphasize specific aspects of functional impact or emotional burden [72]. A wide range of outcome measures limits direct comparability at the descriptive level. Therefore, the standardized mean difference (SMD) was used for meta-analytic pooling of caregiver burden outcomes assessed with different instruments. Disorders The study population primarily comprised family caregivers of patients with neurologic disorders (11 studies, 47.82%) and cancer (4 studies, 17.39%). Additional studies were conducted in chronic kidney disease (CKD) (2 studies), mental disorders (2 studies), heart failure (1 study), and chronic obstructive pulmonary disease (COPD) (1 study). Two studies were conducted in caregiving contexts that did not fit within the predefined disorder‑specific classifications. Therefore, they were categorized as other chronic conditions. One study involved family caregivers of patients receiving palliative care [54], and another focused on caregivers of patients with chronic complex disease [62] (Table 1). Studies related to neurologic disorders were mainly conducted in cognitive disorders such as dementia and Alzheimer’s (8 studies, 34.78%). Other studies focused on stroke (2 studies, 8.69%) and acquired brain injuries (ABI) (1 study, 4.34%). Mental disorders also included two studies in the areas of anxiety, depression (1 study, 4.34%), and schizophrenia (1 study, 4.34%). Tele-interventions The content of tele-interventions in the 23 included studies was educational (10 studies, 43.47%), supportive (6 studies, 26.08%), rehabilitative (3 studies, 13.04%), monitoring (2 studies, 8.69%), and counselling (2 studies, 8.69%). Educational content facilitated caregivers’ understanding of the disease and their roles [64, 65], reducing uncertainty and increasing confidence [65]. Supportive and counselling content focused on emotional support [63], stress management [50, 54], and coping strategies [59, 67], aiming to lessen emotional distress and social isolation [60, 63]. Rehabilitative content strengthened caregiving skills and self‑efficacy [49, 69], while monitoring content provided continuous feedback, reassurance, and timely professional input [53, 54]. Although only some studies explicitly reported a theoretical framework, most interventions were broadly aligned with psychosocial stress and coping models [59, 63, 67]. Across studies, tele‑interventions aimed to reduce caregiver burden. They sought to improve caregivers’ understanding, coping resources, and perceived control, while increasing access to informational and psychosocial support. The type of tele-interventions were classified into four mutually exclusive categories based on their primary mode of delivery. When tele‑interventions incorporated multiple delivery modalities, they were assigned to the subgroup representing the predominant component in order to avoid double-counting and to preserve statistical power. The categories were defined as follows: (1) telephone-based, referring to standard audio-only telephone calls; (2) mobile applications, defined as standalone software installed on smartphones or tablets; (3) web-based platforms, referring to interventions delivered through web browsers or video-conferencing systems; and (4) social networks or messaging applications, which utilize third-party communication platforms (e.g., WhatsApp), regardless of the device used to access them. Across the included studies, these modalities comprised web-based platforms (7 studies, 30.43%), telephone-based interventions (6 studies, 26.08%), mobile applications (5 studies, 21.73%), and social networks (5 studies, 21.73%), with web-based delivery being the most common. Web-based platforms included video visits [54], online group exercises [50], training sessions via video [59, 64, 65, 67], and videoconferencing [56]. Telephone-based interventions were delivered through regular phone calls. These calls were used to provide education [66] and family support [60, 63]. Some interventions were implemented as hybrid programs combining telephone support with face-to-face sessions [53, 68] or with online support [48]. In the studies reviewed, mobile applications were developed for educational [51, 61], supportive [70], and rehabilitative [49, 69] purposes. These applications included the “Smart Brain” and “DemTab” apps for dementia, “Oncokompas” app for cancer and Android apps for COPD and stroke. Social networks like Zoom [55, 62], WhatsApp [57], and Facebook [52] have been utilized for both individual [58] and group training [52, 55, 57, 62]. Despite heterogeneity in delivery methods, the underlying mechanisms of action were largely consistent. These interventions primarily focused on modifying caregivers’ subjective appraisal of caregiving demands rather than reducing objective caregiving workload. The effect of tele-interventions Of the 23 included studies, 10 reported a statistically significant reduction in family caregiver burden following tele-interventions (p < 0.05). The remaining studies reported either non-significant or mixed findings. A quantitative synthesis was conducted based on 17 randomized controlled trials (RCTs) that reported caregiver burden outcomes and provided sufficient data for meta-analysis. Six studies were excluded from the quantitative synthesis because of incomplete reporting of caregiver burden outcomes. Specifically, mean values and/or standard deviations for the intervention and control groups were absent or insufficiently reported. These studies were kept in the qualitative synthesis and are summarized descriptively [52, 54, 60, 67, 68, 70]. Using a random-effects model, the pooled analysis indicated a greater reduction in caregiver burden in the intervention group compared with the control group (SMD = -0.39; 95% CI: -0.76 to − 0.01; p < 0.001). However, this finding was accompanied by high heterogeneity among studies (I² = 94.50%, τ² = 0.58), indicating substantial clinical and methodological variability across studies. Given the considerable heterogeneity, subgroup analyses were conducted using random-effects models. These analyses were based on the type of tele-interventions and the type of disorders to explore potential sources of variability (Fig. 4). A sensitivity analysis was conducted employing a leave‑one‑out approach, indicating that the overall pooled estimate was not influenced by any singular study. Omission of two studies Bahadori et al. [49] and Xie et al. [65] resulted in pooled confidence intervals with less overlap with the original overall estimate. This change showed a moderate shift towards the null. However, exclusion of these studies did not materially change the magnitude or direction of the overall effect [Bahadori et al.: SMD = -0.28, 95% CI: -0.62 to 0.06; Xie et al.: SMD = -0.29, 95% CI: -0.64 to 0.06] (Fig. 5). Fig. 4Forest plot of the overall meta-analysis Forest plot of the overall meta-analysis Fig. 5Forest plot of sensitivity analysis Forest plot of sensitivity analysis Subgroup analysis of disorders Figure 6 presents the subgroup analysis of caregiver burden by disorder type. Based on random-effects models, tele-interventions were associated with statistically significant reductions in caregiver burden among caregivers of individuals with neurologic disorders (I² = 94.46%, p < 0.001). Significant reductions were also observed among caregivers of individuals with cancer (I² = 93.07%, p < 0.001), and other chronic diseases (I² = 86.45%, p < 0.001). Heterogeneity remained high across all subgroups, and the overall heterogeneity across disorder groups was considerable (I² = 94.50%, p < 0.001). The standardized mean difference suggested a more pronounced reduction in caregiver burden within the subgroup of other chronic diseases (SMD = -1.03). In comparison, smaller effects were observed for neurologic disorders (SMD = − 0.30) and cancer (SMD = 0.17). However, these findings are based on few studies within each subgroup and substantial heterogeneity, limiting firm conclusions. Fig. 6Forest plot comparing disorders Forest plot comparing disorders Subgroup analysis of tele-interventions The subgroup analyses examined the effects of tele-interventions according to both the type of interventions (Fig. 7) and the content of interventions (Fig. 8). Fig. 7Forest plot comparing the type of interventions Forest plot comparing the type of interventions Fig. 8Forest plot comparing the content of interventions Forest plot comparing the content of interventions When categorized by type, tele-interventions delivered via mobile applications (I² = 89.83%, p < 0.001), web-based platforms (I² = 93.69%, p < 0.001), telephone-based (I² = 95.86%, p < 0.001), and social networks (I² = 77.65%, p = 0.01) showed varying effect sizes. Larger reductions in caregiver burden were observed for mobile app–based (SMD = -1.03), web-based (SMD = -0.47), and social networks interventions (SMD = -0.30), whereas telephone-based modalities showed a smaller effect (SMD = 0.26). A second subgroup analysis examined the content of interventions. Because some categories contained only a single study, counselling (n = 1) and monitoring (n = 1) were merged with conceptually related components (educational and supportive, respectively) to form three robust groups: educational–counselling, rehabilitative, and supportive–monitoring. This increased statistical power and allowed for comparison of heterogeneity across the content of interventions. The corresponding heterogeneity values were educational-counselling (I² = 80.31%, p < 0.001), rehabilitative (I² = 95.36%, p < 0.001), and supportive–monitoring (I² = 85.59%, p < 0.001). Greater reductions in caregiver burden were observed for rehabilitative (SMD = − 0.82) and educational–counselling (SMD = − 0.68) interventions, while supportive–monitoring showed a smaller effect (SMD = 0.41). Overall, tele-interventions were associated with reductions in caregiver burden compared with control groups; however, heterogeneity across studies remained substantial (overall I² = 94.50%, p < 0.001), limiting definitive conclusions regarding the comparative effectiveness of specific type or content. GRADE assessment The summary of findings alongside the certainty of evidence concerning caregiver burden is presented in supplementary material 3. The overall certainty of evidence was appraised as “Low”. The evidence was downgraded due to serious risk of bias (arising from outcome assessment and selective reporting), and serious inconsistency (indicated by high heterogeneity, I² = 94.50%). Consequently, our confidence in the effect estimate is limited, and the true effect may be substantially different from the estimate of the effect. Discussion This systematic review and meta-analysis examined the effects of tele-interventions on family caregiver burden across a heterogeneous body of randomized controlled trials. Overall, the findings suggest that tele-interventions may be associated with reductions in caregiver burden; however, the substantial heterogeneity across trials limits the strength of this conclusion. Only 10 of the 23 included trials demonstrated statistically significant between-group effects, and the pooled estimate was accompanied by substantial heterogeneity (I² = 94.50%), indicating marked variability across study contexts. The high level of heterogeneity suggests that the pooled effect represents an average across diverse caregiver populations, intervention designs, and clinical conditions rather than a consistently observed benefit. Differences in intervention modality, duration, outcome measures, follow-up periods, and caregiver characteristics likely contributed to this variability, limiting confidence in the generalizability of the pooled estimate. Although this review did not explicitly evaluate or test a specific theoretical framework, the intervention components reported across the included studies can be interpreted through established psychosocial stress and coping models. For example, Lazarus and Folkman’s transactional model conceptualizes stress as a process in which individuals appraise demands and draw on coping resources to manage them [73]. Educational and supportive tele-interventions appear to strengthen these coping resources by enhancing knowledge, perceived competence, and emotional support, thereby contributing to adaptive appraisal and reduced burden. Educational components may support caregivers’ cognitive appraisal by increasing understanding of the disease and caregiving tasks, while supportive or counselling elements may enhance coping resources and emotional regulation [74–78]. In addition, the Pearlin’s stress process model of caregiving suggests that access to informational and social support can buffer the impact of caregiving-related stressors [79, 80]. This Model explains how interventions that increase mastery and social support can interrupt the progression from role strain to caregiver distress [81–83]. Many of the tele-interventions in this review, through education, emotional support, and skill-building, appear broadly consistent with these mechanisms [76, 77]. Our subgroup analysis further supports this perspective. Interventions with rehabilitative and educational contents were more effective in alleviating caregiver burden compared to passive supportive-monitoring methods. Through their emphasis on skill development and problem-solving, such active approaches more successfully enhanced caregiver empowerment. This finding is consistent with contemporary evidence underscoring the substantial psychological advantages of actively engaging caregivers [28, 84]. Consistent with prior systematic reviews, tele-interventions have been associated with improvement in caregiver-related outcomes such as knowledge, self-management skills, access to informational resources, and psychosocial support [75, 85–87]. A recent systematic review and meta-analysis comprising 23 randomized controlled trials involving informal caregivers of patients with chronic diseases further supports these findings. This review reported that e-health interventions significantly reduced caregiver burden and were also associated with improvements in anxiety, stress, and quality of life [88]. Recent evidence from meta-analyses further supports the potential benefits of technology-based caregiver interventions while also highlighting considerable variability across studies. For instance, Lumini et al. reported that digital technology–based interventions produced a moderate short-term reduction in caregiver burden (d = − 0.65) and stress, although heterogeneity remained substantial (I² = 75%), suggesting that effectiveness may depend on intervention modality, delivery format, and contextual factors [34]. Such findings are consistent with the variability observed in the present review. Li et al. conducted a systematic review and meta-analysis on online psychosocial interventions for informal caregivers of people with dementia. The study found no statistically significant effect on caregiver burden but did report improvements in self-efficacy [8]. This suggests that tele-interventions may have more consistent effects on proximal psychosocial outcomes than on burden itself, underscoring the importance of intervention type, target population, and outcome selection when interpreting results. The divergence between the two studies likely reflects methodological and contextual differences. Li et al. focused only on dementia caregivers and online psychosocial interventions, whereas our review encompassed a broader range of tele-interventions across diverse caregiver populations and chronic conditions. In addition, Li et al. restricted their analysis to studies published after the COVID-19 pandemic, while our review included both pre- and post-pandemic evidence. Sensitivity analyses showed that excluding the studies by Bahadori et al. [49] and Xie et al. [65] did not change the direction or statistical significance of the pooled effects, indicating the robustness of the overall findings. Differences in follow‑up duration and intervention design, particularly the short‑term assessment in Bahadori et al. and the largely self‑directed, web‑based approach in Xie et al., may explain their limited comparability with predominantly multicomponent, interactive interventions, but did not materially influence the overall results. Substantial methodological and clinical heterogeneity persisted across studies and subgroups, reflecting variation in intervention design, caregiver populations, outcome measures, and clinical contexts [34, 89–92]. Accordingly, subgroup differences are exploratory and cannot be considered confirmatory, particularly given the limited number of methodologically homogeneous studies [85, 92]. While subgroup patterns may reflect differences in disease trajectories and caregiving demands across chronic conditions, neurological disorders, and cancer, these interpretations remain speculative and are not directly supported by the pooled data [41, 87, 91, 93, 94]. For example, previous meta-analytic evidence in caregivers of patients with chronic respiratory diseases has also demonstrated significant reductions in caregiver burden following supportive or educational interventions, but with considerable heterogeneity across studies (SMD = − 0.65; I² = 82.7%) [95]. This pattern further suggests that intervention effectiveness may vary depending on caregiving context, intervention components, and population characteristics. Mobile application–based interventions appeared to show more favorable pooled effects than web-based or telephone approaches; however, direct comparisons remain limited due to substantial heterogeneity and differences in intervention intensity and design [96]. Overall, these findings are broadly aligned with recent systematic reviews and meta-analyses reporting modest reductions in caregiver burden following tele-interventions across diverse care contexts, including palliative care and dementia caregiving, despite considerable variability across studies [86, 97, 98]. Taken together, the current evidence suggests that while tele-interventions may offer meaningful support for family caregivers, their effectiveness is likely influenced by intervention characteristics, delivery modalities, and contextual factors. The findings of this systematic review and meta-analysis suggest that tele-interventions hold promise as supportive strategies for family caregivers; however, the substantial heterogeneity and methodological variability across studies significantly limit the strength of the conclusions that can be drawn. Future rigorously designed randomized trials with standardized outcome measures, clearly defined intervention components, and adequate follow-up periods are needed to establish more definitive evidence regarding the effectiveness and sustainability of tele-interventions in reducing caregiver burden. In addition, future research should explore intervention characteristics and contextual factors through well-designed subgroup analyses to better understand which tele-intervention approaches are most beneficial for specific caregiver populations. Strengths and limitations This systematic review and meta‑analysis synthesizes recent evidence published between 2015 and 2025 on the effects of tele‑interventions for family caregivers. A key strength of this review is the use of a transparent and rigorous methodological framework, including adherence to PRISMA guidelines, prospective protocol registration, quality appraisal using the Cochrane Risk of Bias tool, and the application of a random‑effects meta‑analytic model complemented by sensitivity and exploratory subgroup analyses. Additionally, the inclusion of randomized controlled trials across diverse caregiving contexts, intervention modalities, and outcome measures allowed for a broad and comprehensive overview of tele‑interventions supporting family caregivers. Manual reference checking further enhanced the completeness of the evidence identification process. This review exhibits several limitations. Substantial clinical and methodological heterogeneity was observed across studies, reflecting marked variability in caregiver populations, intervention characteristics, outcome measures, and follow‑up durations, which limits the certainty and generalizability of the pooled estimates. Caregiver burden was assessed using multiple instruments (e.g., ZBI, CSI, CBI, IEQ, EDIZ, BCOS‑R), differing in structure and sensitivity to change; despite the use of standardized effect sizes, this variability constrained comparability and secondary analysis. High heterogeneity persisted even after sensitivity and subgroup analyses, underscoring the complexity of synthesizing effects across heterogeneous tele‑interventions models and caregiving contexts. Excluding high–risk‑of‑bias trials improved internal validity but may have reduced the comprehensiveness of the evidence. Moreover, the overall methodological quality of the included trials was mixed, with frequent unclear risk of bias across key domains, suggesting that the current evidence base remains methodologically immature and highlighting the need for more rigorously designed and transparently reported trials. Although funnel plot asymmetry and Egger’s test suggested potential small‑study effects, the trim‑and‑fill method did not materially alter the pooled estimate; given the extreme heterogeneity, these findings do not rule out residual publication bias. Finally, several eligible studies were excluded from the meta‑analysis due to incomplete statistical reporting. In addition, the review was limited to English‑language publications from a restricted time frame and database selection, and none of the included studies reported cost or cost‑effectiveness outcomes, limiting conclusions regarding economic feasibility. Implications for practice While this review indicates tele‑interventions can support family caregivers, the considerable heterogeneity across studies limits the certainty of these findings. Clinically, these tools serve well as adjuncts to traditional care rather than standalone solutions [99, 100]. They are valuable in community health, home care, chronic disease management, palliative care, and rehabilitation nursing, where sustained connection is critical. In these contexts, tele‑interventions can enhance routine practice by substituting or optimizing selected in‑person contacts (e.g., follow‑up education, symptom check‑ins, or caregiver counseling), rather than introducing parallel tasks that exacerbate workload [101]. Nurses can utilize these tools to deliver education and support; but successful adoption depends on practical realities. The effective implementation of tele-interventions relies on robust organizational support, including protected time, platform usability, equity of access, digital training, and intuitive and user‑friendly platforms that streamline nursing workflows. Close collaboration with IT specialists, data security, and clear protocols are equally essential to prevent role strain and protect nurses from burnout [102]. Future research should focus on reducing the current variation across studies by using more standardized research methods. Using common and validated measures to assess caregiver burden would make it easier to compare results across different trials. Researchers should also clearly report the characteristics of the interventions and, when possible, standardize aspects such as their duration, frequency, and the type of tele-interventions used. This would help researchers better combine and interpret findings from different studies. In addition, future trials should include economic evaluations, such as cost-effectiveness or cost-benefit analyses, to provide evidence that can guide policy decisions and support wider implementation. Finally, more pragmatic trials conducted in real-world clinical settings are needed to determine whether tele-interventions remain practical, effective, and sustainable when applied in routine healthcare practice. Conclusion This systematic review and meta‑analysis suggests that tele‑interventions may help reduce family caregiver burden; however, the certainty of this evidence is limited by substantial heterogeneity across studies. As a result, the pooled estimates should be considered exploratory rather than conclusive. Nevertheless, the overall findings indicate that tele‑interventions represent a promising, flexible, and accessible means of supporting family caregivers. Future research should extend beyond the predominant focus on dementia and include caregivers of individuals with mental disorders, cancer, and a wider range of chronic conditions, including cardiovascular, renal, respiratory, and metabolic diseases. To better clarify when and for whom tele‑interventions are most beneficial, future trials should adopt rigorous study designs with clearly defined intervention components, standardized outcome measures, sufficient follow‑up durations, and diverse caregiver populations. If confirmed by well-designed trials, tele-interventions could provide a practical way to integrate caregiver support into routine nursing and community health services. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1: Search strategy for the main database. Supplementary Material 1: Search strategy for the main database. Supplementary Material 2: Funnel plot for assessment of publication bias. Supplementary Material 2: Funnel plot for assessment of publication bias. Supplementary Material 3: Summary of Findings (SoF) table based on the GRADE assessment (tele-interventions compared with usual care for caregiver burden). Supplementary Material 3: Summary of Findings (SoF) table based on the GRADE assessment (tele-interventions compared with usual care for caregiver burden).