Background

Nowadays, the world’s increasing aging population implies growing needs for caregiving. Aging groups are usually diagnosed with medical conditions related to aging such as dementia, cancer, stroke, and chronic conditions [1]. A developing number of older people are leaving nursing home and being cared by families as a means of cutting down costs for caregiving. Family caregivers were regarded as an important extension of the aging healthcare system [2]. Informal caregivers, also known as family caregivers, are unpaid individuals (family members, friends) who provide most of the required assistance or supervision [2, 3]. Caring for older persons usually led to personal sacrifices, and caregivers produced goods to their loved ones and society at considerable costs to themselves [4]. It has been well documented that most caregivers undertook heavy burden, led to negative objective and subjective consequences resulting from the provision of care, encompassing physical, psychological, emotional, financial, and social problems [3]. Especially, during the COVID-19 pandemic, family caregivers experienced changed caregiving tasks and additional caregiving challenges within unusual circumstances and changes to the caregiving routines [5].

Globally, needs were urgent for enhanced capacity for caregivers of older adults and home care, which many believed could be addressed via technology at least in part. Gerontechnology as an expanding and novel field dedicated to the development and utilization of technological devices to meet the demands of aging groups [6]. Mahoney and colleagues [7] detected the effects of the first computerized workplace-based intervention for caregivers directly. In a study on an interview study of around 1500 caregivers, 53% of them used Internet sources of information upon caregiving [2]. A National Academy of Sciences report pointed out technology might be helpful for family caregivers in ways typical for the general population and specific functions for the caregiving role [8]. And previous studies were mostly focus on TBI used to support dementia caregiving [9,10,11,12,13,14,15,16,17,18,19,20,21] and by means of Internet [1, 9, 12, 13, 15, 16, 18, 19, 22,23,24,25]. Yet, there were not acknowledged definitions and clarifications on those applied technologies. And publications even used extensively variated terminologies [26]. For this research, we divided technologies for caregiving into several main formats: assistive devices (e.g., helping devices for transportations), tele-series devices (e.g., telephone etc.), Information and Communications Technologies (ICTs) (e.g., Internet etc.), smart home technology, and artificial intelligence and big data [21, 26]. Usage of technology applications were categorized as listed: for decreasing caregiving tasks and compensating for needy care-recipients; for delivering psycho-social interventions, mainly include psychosocial/cognitive behavioral therapy and social supports; for providing information and problem-solving strategies, coordination of care, and managing a new caregiving routine; for social and family communications and leisure; and for activities and behavioral training via devices [5, 8, 14, 18, 27].

TBI had the potentials of removing the barriers of traditional social services (e.g., logistics) and facilitating the utility of demanding supports for caregivers [2]. Findings recognized that TBI had the strengths of being convenient, practical, cost-effective, low-cost, and contained various formats and interactive between actors [13, 26, 28,29,30]. Technology has helped caregivers increase knowledge and skills, sense of control, confidence, care self-efficacy, and enjoyment; and improve their adherence to treatment, support from professionals and other caregivers; and connect more with distant family members [2, 29, 31, 32]. Overall, caregivers who received TBI reported a decrease in workload, pressure and burden in caregiving [2, 6, 30]. However, a substantial number of caregivers have experienced obstacles in using supportive technologies, such as limited accesses, availability of technology, cost, time, less willingness, problems of affordability, retrofitting complications, potential inappropriate use of the technology, and other usability problems [13, 27, 30, 33]. Participants indicated concerns about technology, like digital divide, considering of standardization, technicalities, surveillance, skepticism, and security [26, 27, 34]. Thus, it is essential to illustrate the roles of TBI to support informal caregivers of older adults generally.

Multiple systematic review and meta-analysis of technology interventions to support caregivers are increasingly noted in the literature [35]. While, most of reviews discussed the positive aspects of technologies on dementia patients and their family caregivers, and most of them focused on computer and Internet-based intervention [28, 36, 37]. Several reviews were about the effects of assistive technologies and mobile apps on caregivers [23, 38].

Accordingly, the effectiveness of TBI on the burden of caregivers of the elderly is inconclusive. Less is known about the effects of TBI on caregiving burden within aging care-recipients with various specific medical issues, and about the effects of diverse modalities of technologies. This review is to define the utilization of technologies applied to decrease caregiving burden, to detect their effects in practice, and to distinguish the influential elements of TBI on caregiving burden of aging people via several group disparities.

Methods

This systematic review and meta-analysis comprehensively incorporated RCTs. The systematic review was registered with the International Prospective Register of Systematic Reviews at the Centre of Reviews and Dissemination in the UK (CRD42021277865). It was reported according to the 2020 edition of Preferred Reporting Items for Systematic Review and Meta-analyses (PRISMA) guidelines [39].

Search strategy

The complete literature search was carried out on 11 databases and registry, containing Web of Science, PubMed, EMBASE, Scopus, CINAHL, PsycINFO, WANFANG, CNKI, CQVIP databases, Cochrane Library Trials, and ClinicalTrials.gov. We selected all articles and trials published between January, 1990 and October, 2022. Our search keywords consisted of three subsets: participant (“caregiving”, “caregiver”, “carer”), intervention (“tele”, “big data”, “mobile”, “internet”, “robot”, “technology”, “artificial Intelligence”, “e-health”, “m-health”, “e-medicine”, “m-medicine”), and outcomes (“burden”, “stress”). We connected them with two Boolean operators (AND and OR) to search for relevant researches in English and in Chinese.

Study inclusion and exclusion criteria

Two teams of researchers screened and selected these studies independently. Studies were chosen in accordance with the inclusion criteria as follows: (1) Study designs were RCTs; (2) Informal caregivers of elderly people included family members, relatives, friends, and volunteers; (3) Interventions in trial groups were provided using supportive technology such as Internet, apps, network, and other technology products; (4) Articles were published between January, 1990, and October, 2022; (5) Interventions aimed at improving the burden of caregivers, measured by burden scales; (6) Care recipients were people aged 55 years or above. We excluded studies if they met the following criteria: (1) were published in neither Chinese nor English; (2) did not report results data, or could not be inferred by contents.

Data extraction

All search results from databases were exported to EndNote X9.3.3 and Excel. We generated a data extraction table to collect the basic information of the included studies, containing title, author, country, year of publication, abstract, purpose of study, journal and so on. Also, the coding scheme extracted contents including subjects, intervention designs, outcomes, measurement tools and duration of follow-ups. We extracted the data mainly by quantitative data: mean and standard deviation (SD). Some of the literature did not report these data directly. The mean and SD of these studies were calculated according to the standard error (SE), D effect size, P value, 95% confidence interval (CIs) and other information given in the literature. Two dependent teams extracted the information separately. It was verified by another reviewer, to ensure the reliability of data extraction. All reviewers agreed on the content of the final data extraction table after discussion. The intricate particulars of data extraction were delineated in Additional file 1.

Assessment of risk of bias

Risk of bias of each eligible study was further assessed by two independent reviewers following the Cochrane Systematic Review Handbook. This research quality assessment evaluated the levels of bias in the following: (1) selection bias (sequence generation and allocation concealment); (2) performance bias (blinding of participants and personnel); (3) detection bias (blinding of outcome assessment); (4) attrition bias (incomplete outcome data); (5) reporting bias (selective outcome reporting); and (6) other biases [40]. The specifics of the risk of bias assessment could be referenced within Additional file 1.

GRADE assessment

We evaluated the quality of evidence for each outcome using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) framework with four possible levels: high, moderate, low, very low [41]. Five factors could decrease the quality of the evidence: risk of bias; inconsistency; indirectness; imprecision; and publication bias [41]. Detailed assessments were presented in Additional file 2. Any difference of evaluation results was discussed between the two teams of reviewers. A third reviewer participated to recheck the assessment and helped reach a consensus.

Data analyses and synthesis

We applied the data (mean, SD) in the meta-analysis on the impacts of TBI. To mitigate inconsistencies across diverse scales and facilitate data amalgamation, we employed Standardized Mean Difference (SMD) in constructing the forest plot. I2 was used to evaluate the heterogeneity among the studies in quantitative statistics. When I2 was reported lower than 50%, a fixed-effects model was chosen in the meta-analysis [42]. We conducted sensitivity analysis to test the stability of outcomes. In order to prevent publication bias, we carried out a funnel plot test to predict the bias. In particular, according to the differences in participants, control groups, and intervention factors, we conducted a subgroup analysis to determine which components of TBI were playing key roles in mitigating burden on caregivers. This meta-analysis was done by RevMan 5.3.

Results

Selection of studies

As shown in Fig. 1, a total of 11,095 research studies and trials were identified from 11 databases and registry. After removing 3647 duplicated literature, 7421 records remained. Following the inclusion and exclusion criteria, we excluded 7330 studies as non-related, and 45 articles and 46 trials were further screened and reviewed in full text. We excluded 77 records: non-RCT studies (n = 7), Repeated experiment (n = 7), Inconsistent participants (n = 7), Inconsistent interventions (n = 8), Inconsistent controls (n = 12), Inconsistent outcomes (n = 10), Uncompleted trials and data unreported (n = 26). Finally, we identified 14 eligible RCTs for our meta-analysis.

Fig. 1
figure 1

PRISMA diagram of included studies in the meta-analysis

Full size image

Characteristics of included studies

Across 14 trials, the sample size ranged from 10 to 237, and the total sample size was 1010. By group, 534 participants were in the experimental groups, and 476 participants were in the control groups. Caregivers’ average age was calculated as 60.65. The ages of the people under care were reported in nine studies. The mean age of care receivers ranged from 59 to 83.27 [10, 15, 16, 18,19,20, 22, 31, 32]. One study reported age as categorial variable and ages of all the caregivers were above 55 years [11]. Participants in the studies were all caregivers of older adults. Five studies did not report the ages of care-recipients, but all identified the care recipients as living with diseases of elderly people, namely dementia [9, 11, 12, 14, 43]. The details of the studies are presented in Table 1.

Table 1 Key Information about participants and RCTs in included studies
Full size table

Risk of bias

According to Fig. 2, all studies have been rated as having a low or moderate level of risk of bias. Two studies reported a low level of risk of bias on all seven aspects [12, 15]. One study was rated with a high risk of selection bias on allocation concealment [16] and six unclear [10, 11, 14, 19, 31, 43]. Four studies were evaluated as having a high risk on performance bias [16, 20, 31, 32]. Seven trials did not report the status of blinding of participants and personnel [9,10,11, 14, 19, 22, 43]. Moreover, three research studies were assessed as having a high risk on detection bias [9, 20, 31]. And seven studies did not report the conditions of blinding of outcome assessment [10, 11, 14, 16, 18, 19, 22]. Except one study by Gustafson and collegues [11], the other trials studies were considered to have a low risk of bias in random sequence generation. And all trials were assessed with a low risk of bias measurement results, data reporting, and other bias.

Fig. 2
figure 2

Summary of Assessments of Risk of Bias

Full size image

Results of overall effects and subgroup analyses

Overall effects of TBI

A meta-analysis of 14 trials indicated an overall reduction in burden of informal caregivers of older adults. As demonstrated in Fig. 3, TBI resulted in a reduction in scores of burden scale for caregivers, which dropped an average of 0.13 points (95% CI − 0.25 to − 0.00). There was statistical significance in the overall combined effects corresponding to TBI (p ≤ 0.05) in the sample of 1010 caregivers. It was found out that there was a low heterogeneity between the studies (Chi2 = 19.75, p > 0.05, I2 = 34%). The results revealed moderate quality evaluated by GRADE (Table 2).

Fig. 3
figure 3

Forest plot of total effects of TBI

Full size image
Table 2 Summary of the results
Full size table

Results of subgroup analysis

Subgroup analysis of specific medical conditions

Most of the trials focused on caregivers of older adults with dementia [9,10,11,12, 14,15,16, 18,19,20, 43]. Two studies had care recipients with disability [32] and chronic diseases [31]. One study did not categorize the data on medical conditions of the care receivers [22].

And the subgroups based on care recipients’ medical conditions were significantly different (Chi2 = 8.86, p ≤ 0.05, I2 = 66.1%) (Fig. 4). The subgroup with older adults having chronic diseases was statistically effective (SMD = − 0.43, 95%CI − 0.84 to − 0.01, p < 0.05), as was the groups without specific medical conditions (SMD = − 0.72, 95%CI − 1.27 to − 0.17, p ≤ 0.01). Evaluated by GRADE, the results were proved with moderate and low quality (Table 2).

Fig. 4
figure 4

Effects among subgroups on medical conditions of older adults

Full size image

Subgroup analysis of formats of technologies

There were three kinds of technology used in intervention: (1) ICTs, including Internet-based intervention [9, 11, 12, 14,15,16, 18, 19, 22] and app-based intervention [31]; (2) telephone [10, 20]; (3) assistive technology [32]; and (4) smart home technology [43].

The subgroups categorized on formats of technology, including ICTs, telephone, assistive technology, and smart home technology were not significantly different (Chi2 = 4.31, p > 0.05, I2 = 30.5%) (Fig. 5). ICTs-based interventions had a significantly positive effect on reducing burden (SMD = -0.24, 95%CI − 0.42 to − 0.06, p < 0.01). According to GRADE, the results were proved in moderate and low levels of quality (Table 2).

Fig. 5
figure 5

Effects among subgroups on formats of technologies

Full size image

Subgroup analysis of usage of TBI

Five studies offered participants psychosocial interventions [9, 10, 16, 20, 22]. Seven studies provided participants with TBI for problem and coping strategies to coordinate care [11, 12, 15, 18, 19, 31, 32]. One study offered technology-based behavioral training for caregivers of elderly people [14]. And one trial provided safe and supportive home environment [43].

The subgroups with groups were not significantly different (Chi2 = 0.55, p > 0.05, I2 = 0%) (Fig. 6). Also, each group did not show significantly beneficial effect on burden (p > 0.05). The results were proved with moderate and low quality by GRADE (Table 2).

Fig. 6
figure 6

Effects among subgroups on usage of TBI

Full size image

Subgroup analysis of controlled intervention

In the control groups, five studies provided usual care or standard care for participants [9, 11, 20, 31, 32]. The control intervention in seven trials was delay intervention [12, 14,15,16, 18, 22, 43]. And controlled arms did not receive any intervention in two studies [10, 19].

The subgroups with interventions on control group were also significantly different and highly heterogeneous (Chi2 = 5.86, p ≤ 0.01, I2 = 76.8%) (Fig. 7). Delay intervention as a comparator had a statistically significant effect on reducing burden on caregivers (SMD = -0.30, 95%CI − 0.49 to − 0.11, p < 0.01). The results were shown in moderate and low quality via GRADE (Table 2).

Fig. 7
figure 7

Effects among subgroups on controlled interventions

Full size image

Subgroup analysis of burden instruments

Six trials measured caregivers burden levels with Zarit Burden Inventory (ZBI) [9, 14, 15, 19, 31, 43]. Three studies used The Caregiver Burden Inventory (CBI) [10, 22, 32]. One research applied Burden Scale for Family Caregivers (BSFC) [18]. And the other four studies used self-made scales (SMS) to test burden status [11, 12, 16, 20].

The subgroups were divided by instruments of burden were not significantly different (Chi2 = 1.68, p > 0.05, I2 = 0%) (Fig. 8). Also, each group did not show significantly beneficial effect on burden (p > 0.05). Evaluated by GRADE, the results were proved in moderate and low quality (Table 2).

Fig. 8
figure 8

Effects among subgroups on burden instruments

Full size image

Subgroup analysis of cultural context

The studies were conducted in European and North American cultural contexts. Nine studies were set in Europe [9, 10, 12, 16, 18,19,20, 22, 31] and five in North American countries [11, 14, 15, 32, 43].

The subgroups categorized on cultural context were not significantly different (Chi2 = 0.74, p > 0.05, I2 = 0%) (Fig. 9). Among European context setting, the intervention had a significantly positive effect on reducing burden (SMD = -0.17, 95%CI − 0.32 to − 0.01, p < 0.05). The results were proved with moderate quality by GRADE (Table 2).

Fig. 9
figure 9

Effects among subgroups on cultural context

Full size image

Results of sensitivity analysis

To verify the stability of the meta-analysis results, we excluded a single study with the largest sample size to test its impact on the overall results. We excluded the study of Wilz et al. [20], an RCT from Germany with 121 participants in the experimental group and 116 participants in the control group. As shown in Fig. 10, the overall effect was still statistically significant (p < 0.05), indicating that the overall result of the combination was stable. The results also revealed moderate quality evaluated by GRADE (Table 2).

Fig. 10
figure 10

Forest plot of sensitivity analysis

Full size image

Publication bias assessment

Figure 11 depicted a funnel plot of effect sizes on the overall meta-analysis. No study fell outside the confidence interval (95%CI), indicating a low heterogeneity in this study. The plot presented symmetry, and all of the scattered points fell to the left of OR = 1, proving that there was little or no publication bias in this review.

Fig. 11
figure 11

Funnel plots of effect sizes for TBI

Full size image

Discussion

Many reviews have revealed positive effects of TBI on mental health outcomes of caregivers [37]. However, this is the first systematic review and meta-analysis to include RCT researches aimed at reducing the burden on older adults’ caregivers via various technologies generally. The overall results of these studies indicated that TBI indeed ameliorated the burden on the caregivers of older adults. The results are consistent with previous studies which revealed a significant post-intervention effect of TBI on caregiver burden [2]. It could be explained technology helped maintain or improve individual capability to do things in daily life and assisted persons in coping with a range of difficulties, such as mobility [8, 17]. Caregivers often sought information and support on the web, which helped reduce the burden by caregiving [25].

Based on the findings from subgroups analysis, we’ve distinguished interventions in control groups and medical conditions of care-recipients were both determinate factors on caregivers’ burden in TBI. It was found out that the delayed intervention with TBI was significantly more beneficial to these caregivers, compared with usual care or non-intervention. It supported the opinion that TBI had the potential to reduce the burden of caregivers of older adults. As suggested previously, face-to-face delivery of interventions was not always optimal or practical for caregivers [44]. Likewise, TBI had a positive effect on caregivers of persons with chronic diseases and without specific medical conditions compared with those caregivers on persons with dementia and disability. This was in accordance with the majority of reviews that web-based intervention programs had positive effects on reducing strain on caregivers of adults with a chronic disease [25]. For utilizing TBI to reduce caregiving burden, there were no significant disparities in term of formats and usage of technology, burden instruments, and cultural context unexpectedly.

Meanwhile, TBI has revealed advantages in reducing caregiving burden of older adults with ICTs (Internet and apps) compared with telephone, assistive technology, and smart home technology. This finding was supported by an explanation that caregivers selected Internet as convenient access to individualized practical advice and emotional support to contact with professionals online, as well as helpful information about the disease, coping skills, and other information to support caregiving [36]. Apps became indispensable and complementary facilitators to health care [23]. Moreover, different from trials in North American countries, TBI conducted in European countries displayed significant effects. It might be explained technologies had more advantages in application within European humanistic or geographic conditions. Similarly, in this review BSFC was proved as an instrument with advantages in measuring caregiving burden of informal caregivers, which was might for the reason on the construction of the instrument. However, usage of technology in intervention consisted of psychosocial intervention, problems and coping strategies, behavioral training, and home environment did not show any disparities and advantages in TBI. On this sense, we did not have preferable contents in using technology to reduce caregivers’ burden.

Strengths and limitations

This review has several strengths. It includes the most recent evidences in exploring the potential effects of TBI on the burden of caregivers of older adults. All the selected trials have been published since 2014. And it ensures literature diversity and comprehensiveness for including peer review articles, grey literature, and registry of RCTs. Conclusively, it proves the positive effects on caregiving burden of older adults’ carers via TBI generally. Via subgroup analyses, this review has also distinguished the most effective factors and advantages in exploring the mechanism of TBI to work on caregiver burden. Furthermore, sensitivity analysis reveals the results of meta-analysis are stable, and the results of publication bias test suggest less publication bias. Also, this review searches both Chinese and English studies to avoid bias in cultural diversity.

However, there remain some limitations. Even included RCTs are mostly conducted with rigorous experimental designs, several of the items are evaluated with high or unclear risk. Evaluations by GRADE assert moderate quality of overall effects and sensitivity analysis, and moderate and low quality on results of subgroup analysis, which are due to risk of bias and imprecision. In one study, the final standard deviation is unavailable, and it was be substituted by the baseline SD following Cochrane Systematic Review Handbook [11]. Also, one trial was excluded for lack of key data. And a number of the trial protocols and reports have not been completed or published.

Implications

Technology for caregivers of older adults is in a rapid pace of changes and developments. It is a policy issue for both of the public and private sectors tend to utilize technologies to facilitate aspects of caregiving to address caregiving needs. The developments of TBI for caregiving demands urgent policy advocacy. Government or organizations like nursing associations are expected to make standards, including notions, categories, and agendas on technologies for caregiving. Nursing professionals should participate in cross-disciplinary dialogues on caregiving-related models and measurements which could be processed creatively in technology for caregiving. It is promising to apply technologies in reducing burden and supporting informal caregivers of aging persons. Yet, there were not that much studies on practice of TBI to help caregivers. It still needs to raise awareness and perceptions about utilization of technologies for caregiving on older caregivers and could refer to European experience. Clinical professionals should pay attention on the factors of caregivers’ adoption of technology, containing social/demographic factors (e.g., socio-economic status), attitudinal elements (e.g., computer anxiety), and component abilities (e.g., cognitive abilities) [24, 27, 33], in order to make caregivers feel more efficacy to use TBI as much as possible. Meanwhile, most of published researches are about TBI in dementia caregiving and by means of ICTs. We advocate its application in caregiving on older adults with various medical conditions, and more kind of modalities of technologies could be tried to support them. It has been proved tailored systems gained more chance of acceptance by the target population [6]. And TBI should be matched to caregivers’ needs seamlessly [21]. As discussed previously, rigorous evaluations were lacking to evaluate the effects of TBI [30]. Yet, as referred in included studies, it is impossible to blind participants when they are exposed online during intervention [16, 20]. We should consider preventing the risks in RCT research designs for TBI. Also, additional methodologies, such as Egger regression, could be implemented to enhance result reliability and precision in accordance with the demands of future investigations. Further studies are warranted to elucidate the associations between intervention effect sizes and various characteristics of RCTs, thus elucidating the efficacious mechanisms of TBI more comprehensively. Furthermore, while insights from European and North American studies are valuable, caution is advocated when applying these findings directly to the Asian region. We suggest more researches with high-quality on TBI to reduce the burden of caregivers published in Eastern contexts.

Conclusion

Highly reliable evidence about applying TBI for reducing caregiving burden remains limited. This research is the first review to discuss on the benefits of a variety of technological interventions on the burden of caregivers, devoted to the caregiving of older adults with dementia, disability, chronic diseases, and without medical conditions. TBI has been elaborated as a positive intervention to alleviate the burden on these caregivers. Controlled intervention and medical characteristics of care recipients are both influencing factors on effects of TBI to reduce burden of caregivers, among which we could focus more on caregivers of persons with chronic diseases and without medical conditions, and delayed TBI as controls. It also enlightens us to pay more attention on the advantages of ICTs-based interventions, BSFC as a burden measurement, and the application of TBI in European context. Moreover, more RCTs with high-quality, different groups of participants, modalities of TBI, and caregiver outcomes are expected in future studies, in order to enrich the evidence of reduction of burden by TBI for caregivers of older adults.