Telehealth Interventions for Family Caregivers of Persons with Chronic Health Conditions: A Systematic Review of Randomized Controlled Trials

Objective The purpose of this study was to provide an in-depth analysis of the components and outcomes of telehealth interventions for family caregivers of individuals with chronic health conditions. Methods A systematic review of 17 databases was conducted for randomized controlled trials published between January 2002 and January 2017. Interventions were analyzed based on type of telecommunication modality, caregiver and care recipient characteristics, intervention components, and caregiver outcomes. Results A total of 57 articles met criteria for inclusion. Telephone was the most frequently used mode of telehealth delivery and focused primarily on caregivers of older adults with dementia and stroke. Skills training was the most prevalent treatment strategy across telephone, web, and combined telephone and web modalities. Improved psychological functioning was reported most frequently across telehealth modalities. Conclusion Telehealth is an effective tool in delivering caregiver interventions and leads to significant improvement in caregiver outcomes. Telephone was used most often to deliver cognitive-behavioral and psychoeducational strategies as compared to web and combined telephone and web modalities. Further research is needed to examine the effects of telehealth interventions on caregiving skills and self-efficacy, as well as health outcomes.


Introduction
Approximately 44 million family members and significant others residing in the United States provide unpaid assistive care to children and adults with chronic health conditions [1]. The responsibilities of family caregivers vary considerably depending on the age of the care recipient, severity of the disability, and availability of resources [2]. As a consequence of caregiving activities (e.g., providing assistance with bathing and managing medications) and other responsibilities (e.g., work commitments), caregivers must perform a difficult balancing act. It is commonplace for family caregivers to prioritize the needs of their loved ones with chronic illnesses above their own personal concerns and lifestyle preferences. Although they often verbalize the importance of health-promoting activities, many caregivers find it challenging to create time for self-care opportunities, to attend doctors' appointments, and to ask for caregiving assistance from other family members [3,4]. Caregivers often report substantial declines in social activities, disrupted familial relationships, financial problems, and deterioration in physical and mental health [5].
In an effort to mitigate these difficulties, governmental and healthcare organizations have called for research on the effects of telehealth interventions in meeting the informational, psychosocial, and healthcare needs of family caregivers. Several of these entities have emphasized the importance of ensuring ease of access, uptake of services, and cost-effectiveness [6][7][8]. Telehealth services for family caregivers facilitate access to interventions in convenient locations and in turn reduce costs of transportation and respite care. Caregivers with partor full-time employment also are able to receive needed services during off-hours, thus reducing work disruptions and potential loss of wages. Furthermore, caregivers residing in rural areas or crowded urban communities may find telehealth modalities more convenient, reducing travel costs, parking restrictions, and/or risks to self and property [9,10].
Although research on the outcomes of telehealth interventions with family caregivers has been slower in evolving compared to research focusing on individuals with chronic illnesses, this trend has changed markedly over the past decade. Prior systematic reviews [2,11] on telehealth intervention provided useful background information about caregiver and care recipient characteristics, types of telecommunication modalities and intervention approaches, and their overall pattern of findings. These reviews supported the effectiveness of telehealth interventions in improving the emotional and psychosocial functioning of family caregivers of persons with chronic health conditions. Although these reviews [2,11] represent important initial efforts in evaluating the outcomes of caregiver telehealth interventions, a notable shortcoming limited the validity of their findings. Both reviews [2,11] incorporated studies with nonexperimental research designs, ranging from case studies to quasi-experiments, thus weakening the causal conclusions that could be drawn about the efficacy of telehealth caregiver interventions.
To address this limitation, this review examined the results of randomized clinical trials assessing the effects of telehealth interventions on the psychosocial functioning, health status, and caregiving abilities of family caregivers. The decision to focus on the results of randomized controlled trials was based on prevailing standards for judging treatment efficacy [12]. Special attention also was given to the types of telecommunication modalities deployed in delivering caregiver telehealth interventions, user applications and options incorporated during the course of treatment, and their relationship to caregiver outcomes.

Identification of Studies.
A literature search was performed to identify studies examining the efficacy of randomized controlled trials of telehealth interventions for family caregivers of individuals with chronic conditions published in English-language peer-reviewed journals from January 2002 to January 2017. Seventeen interdisciplinary scholarly databases were searched (Table 1) using the following terms in various combinations: "caregiver," "telemedicine," "telehealth," "video," "computer," "Internet," "conference," "remote consultation," "train," "intervention," "randomized," "controlled," and "clinical trial." These keywords were selected based on their capacity to retrieve a wide range of pertinent articles and to ensure identification of new articles, as well as those not fully indexed. Strict inclusion/exclusion criteria also were adopted to enhance the relevance of articles for study inclusion ( Table 2).
The initial search yielded 1621 potentially relevant studies archived for further review. After the removal of duplicate articles (n = 280), the remaining articles (n = 1341) were screened for eligibility. Ineligible studies were excluded (n = 899) (e.g., not written in English, feasibility studies, quasi-experimental design, and qualitative studies). Abstracts of the remaining studies (n = 442) underwent further evaluation, with additional studies excluded (n = 358) due to failure in meeting the inclusion criteria. Next, the full texts of the remaining articles (n = 84) were reviewed by two independent coders to confirm eligibility, resulting in the exclusion of 27 additional articles. A total of 57 articles met all inclusion criteria and were included in this systematic review ( Figure 1). Due to the heterogeneity of study populations, interventions, and outcomes examined, we opted not to pursue a formal meta-analysis of the included studies.

Identification and Operationalization of Intervention
Characteristics. Each study was examined to identify telecommunication modalities used during treatment, caregiver and care recipient sample characteristics, outcome measures, intervention characteristics, and primary findings. Description of these key characteristics, organized by delivery modality, can be found in Tables 3-5 (telephone, web, and combination web and telephone interventions). The interventions were further inspected to classify intervention components, caregiver applications, and user control ( Table 6). Caregiver intervention components consisted of group or individual skills training, psychoeducational and resource materials, self-monitoring/tracking, reminders, and group or individual counseling. Caregiver applications included the use of email, discussion forum/online chat, online journaling, text messaging, telephone calls (independent of intervention), and provision of useful links and resources. Regarding user control, interventions were categorized as either facilitator-guided or self-guided. Facilitator-guided interventions used healthcare professionals to lead sessions or maintain the flow of participation during the intervention. Self-guided interventions relied exclusively on participants to complete all phases of the program.

Coding of Intervention Characteristics.
Two researchers coded eligible studies independently on caregiver and care recipient demographics, intervention design, telehealth modalities utilized, methodological characteristics, and study outcomes. Operational definitions were summarized in a codebook to ensure coding procedures were performed consistently and accurately. Coder agreement was assessed on quantifiable factors across a subset of categories. Findings from both reviewers were compared, and areas of disagreement were discussed until consensus was achieved. (1) Were telecommunication-based randomized controlled trials which focused on at least one caregiver outcome (i.e., health behaviors, psychosocial functioning, health service utilization, and cost or clinical outcomes) (2) Used at least one telehealth component in 50% of caregiver training sessions (e.g., service delivery over the Internet, telephone, videophone, point-to-point videoconference, and/or chat) (3) Were primarily focused on family caregiver intervention (may also include a care recipient training component) (4) The care recipient had a chronic medical condition (e.g., HIV/AIDs, progressive dementia, heart failure, psychiatric disabilities, and traumatic brain injury) (1)Were written in languages other than English (2) Were duplicate articles (3) Were proposals or written descriptions of intervention programs, usability studies, quasi-experiments (i.e., used control groups, but not randomized), feasibility studies, or largely qualitative studies that lacked quantitative measures of caregiver health behavior or clinical outcomes (4) Used telehealth technologies solely to evaluate data on physical signs and symptoms of a condition, physiological indicators of health, and/or to provide emotional support (5) Focused exclusively on improvements in caregiver's knowledge and attitudes Records excluded for following: a) were not randomized controlled trials (RCTs), b) did not have at least one telehealth delivery component, (c) reported less than 50% of training sessions that were technology-mediated, (d) used outcome measures focused exclusively on knowledge and/or attitudinal measures, and (e) were not focued on training the primary family caregiver of a care recipient with a chronic medial contion (n = 358) Full-text articles excluded due to the following: : (a) primary usability studies, (b) proposals or written descriptions of an intervention program, (c) quasiexperiments, (d) feasibility studies with less than 15 participants, or (d) qualitative studies that did not report quantitative for health behavior or clinical outcomes; (e) used telehealth methods to evaluate data on physical signs and symptoms of the conditon, and /or physiological indicators of health; and (f) used telehealth technologies to collect data on symptoms and signs of illness only and/or for emotional support but not for formal caregiver intervention          Caregivers' levels of depressive symptoms decreased significantly more for those in the intervention group, compared to control. Caregivers of male patients in the intervention group experienced a greater decrease in anxiety from presurgery to 3 weeks after discharge than did caregivers of male patients who were in the control group.
There was no evidence that the intervention had a greater impact than control on patient anxiety from presurgery 11 International Journal of Telemedicine and Applications

12
International Journal of Telemedicine and Applications Caregivers showed a significant decrease in negative experience and a significant increase in positive experience postintervention, which was maintained at the one-month follow-up. 14 International Journal of Telemedicine and Applications

18
International Journal of Telemedicine and Applications

20
International Journal of Telemedicine and Applications  about and attitudes toward dementia were statistically significant across groups among laypeople or professionals. In prepostcomparisons, participants in the intervention group reported less distress in tense situations, more empathy and concern for the well-being of other people and feeling better able to 22 International Journal of Telemedicine and Applications   an initial 2-day intervention training program was provided to participating therapists. Family caregiver participants received an initial 1day in-home training session to introduce the family intervention which focused on problem-solving training, communication skills, self-regulation, and anger management and to identify family goals to be addressed in the intervention. Participants received a detailed manual and completed a series of online modules and Skype sessions with a licensed therapist. Up to 4 supplemental The intervention was effective in reducing parent distress but had minimal effect on depression and selfefficacy. The intervention was more effective in reducing distress in low-income families at 12-and 18month follow-up than the control group. While depression was greater for caregivers of adolescents with severe TBI versus adolescents with moderate TBI, these differences did not continue through the 12-and 18-month follow-up. 24 International Journal of Telemedicine and Applications

26
International Journal of Telemedicine and Applications 27 International Journal of Telemedicine and Applications 2.4. Data Analytic Strategy. Analysis of systematic review findings was organized based on the primary telecommunication modalities used in delivering caregiver interventions. Studies were placed into three different categories: (a) telephone interventions, (b) web interventions, and (c) combination of telephone and web interventions. This data analytic strategy was implemented due to limited variation in caregiver populations. Over half of included studies targeted family caregivers of older adults with dementia or stroke. Two advantages emerged from this approach: (a) optimal differentiation across caregiver telehealth outcome studies and (b) substantive methodological and conceptual comparisons. Only a small percentage of included studies reported caregiver-care recipient dyadic results. Thus, this review addressed caregiver outcomes only.
Data analysis focused primarily on counts and percentages across the major elements of the review (i.e., characteristics of caregiver samples, intervention strategies, outcomes, and caregiver applications and user controls). First, descriptive statistics on caregiver sample characteristics were calculated for both the overall sample of caregivers and separately for each telecommunication delivery mode. Similarly, the number and percentage of specific types of intervention strategies were assessed for the overall sample of studies and separately for each of the three categories of delivery modalities.
Second, vote counting (i.e., number and percentage) of statistically significant caregiver target outcomes was conducted for both the overall sample and separately for each of the three telecommunication modalities. The term, improved outcome, was used only if the target outcome of the intervention group was statistically significant compared to the comparison/control groups (Tables 3-5). Target outcomes included psychological well-being, physiological functioning, caregiving skills, social functioning, problemsolving/goal setting, and quality of life. Studies were included in this count once. The rationale for this procedure was to facilitate accurate and consistent percentage calculations of the total number of significant studies versus the total number of studies and for each of the three modality categories.
Last, the number and percentage of caregiver applications and user control features were assessed for both the overall sample and separately for each of the three telecommunication modalities (Table 6).

Characteristics of Telehealth Randomized Controlled Trials across Modalities
3.1.1. Samples, Journal Outlets, and Study Locations. A total of 57 randomized controlled trials evaluating the efficacy of telehealth interventions for caregivers of persons with chronic health conditions were included in this systematic review (Figure 1). Tables 3-5 summarize the key characteristics and outcomes of these studies, including telehealth delivery modalities, type of caregiver sample, primary outcome measures, intervention description, and key findings. Of the

28
International Journal of Telemedicine and Applications  Length: 4 to 6 months. Description: caregivers were stratified into 2 cohorts based upon use (n = 155) or nonuse (n = 74) of the Internet, then randomized. Intervention: (n = 107) electronic content (i.e., video vignettes, written information, brief assessments) on dementia progression, caregiving skills, health topics, and caregiver health were accessed 3 days per week with care manager oversight and follow-up electronically. Comparison: (n = 122) printed educational materials and a DVD covering content like that introduced electronically were provided, in addition to monthly telephone calls from a care manager for support.

International Journal of Telemedicine and Applications
Among home Internet users, receiving the Internet plus case management intervention predicted a decrease in reported feelings of isolation, compared to receiving the telephone only intervention. Among non-Internet users, receiving telehealth and case management predicted a decrease in reported family hardship associated with caregiving, compared to those receiving the telephone only intervention.

30
International Journal of Telemedicine and Applications  Intervention: (n = 49) received a computermediated system using interactive voice response (IVR) with a telephone interface. When caregivers called, the IVR queried about problem behaviors and responded with targeted information. Study personnel were alerted when a caregiver reported continuing problem behaviors and increasing stress levels during any 3week period. The system also provided voice-mail linkage to experts in Alzheimer's and related dementias, a voice-mail telephone support group, and a distraction call for care recipients. Usual care: (n = 51) standard care and educational materials.
No main effects were found on any measure. In post hoc analyses, among caregivers who reported low-to mid-level mastery at baseline (N = 32), those in the intervention group showed a decrease in bother, depression and anxiety relative to the comparison group. participants received weekly automated selfcare support telephone calls with notifications about problems sent to Caregivers in the intervention group reported significantly less caregiving strain than those in the control at 6 and 12 months. This effect plus improvements in depressive symptoms were noted in caregivers 31 International Journal of Telemedicine and Applications Intervention: (n = 77) self-guided web-based caregiving and TBIspecific online modules, plus concurrent biweekly telephone calls to discuss problems and solutions.
Modules were not presented in a specific order, rather caregivers were directed to relevant modules during telephone calls as indicated by their answers to a brief interview at the onset of the call. Usual care: (n = 76) standard care.
Caregivers in the intervention group showed significantly higher quality of life and emotional well-being. They also showed significantly more active coping strategies and less emotional venting coping styles at follow-up. Caregivers also tended to report more support from others, feeling more competent at obtaining information from medical professionals, and taking care of themselves than usual care. International Journal of Telemedicine and Applications 57 included articles, only 16 (28%) were caregiver-care recipient dyad studies. The majority of intervention studies (n = 32, 56%) centered on family caregivers of older adults with neurological conditions, such as dementia and stroke. A substantially smaller percentage of studies focused on interventions for caregivers of persons with cancer (n = 6, 11%) and caregivers of children, adolescents, and young adults with chronic health conditions (n = 6, 11%). Only a few intervention studies targeted caregivers of persons with heart failure (n = 2, 4%) and persons receiving hospice or palliative care (n = 2, 4%).
Of the 57 included studies, 58% (n = 33) were telephonemediated, 25% (n = 14) used web technologies, and the remainder (n = 10, 17%) used a combination of telephone and web modalities (Tables 3-5). Last, the 57 included studies were published in 42 different journals, primarily psychological and medical in nature. The randomized controlled trials were conducted in nine different countries, with the United States accounting for 80% of publications.

Intervention Components.
Of the 57 studies, the majority of interventions (n = 37, 65%) were ≤ six months long, included participant follow-up for ≤ six months (n = 17, 30%), and did not incorporate booster sessions (n = 53, 93%). Furthermore, most studies (n = 47, 82%) did not collect postintervention data. Across the 57 studies, interventions included at least one of five different components, including skills training (group, individual, or both), psychoeducational and resource materials, self-monitoring/tracking, reminders, and counseling (group or individual). Skills training was conducted most often (n = 43, 75%) either individually (n = 32, 56%), in a group (n = 9, 16%), or using a combination of individual and group training (n = 2, 4%). Topics covered in the skills-building sessions varied across telehealth modalities and included training on effective communication, diseasespecific caregiving skills, care recipient safety, social skills, and problem-solving. Psychoeducational and resources materials also were commonly used (n = 40, 70%) to supplement intervention components and covered similar topics across modalities (e.g., problem-solving, stress management, and coping skills). Other less frequently used strategies consisted of individual (n = 8, 14%) or group (n = 7, 12%) counseling and the use of self-monitoring/tracking (n = 14, 25%) and reminders (n = 4, 7%) ( Table 6).
3.1.3. Outcomes. Forty-two of 57 studies (74%) reported significantly greater improvement on at least one target online video-based modules with coping skills training related to dementia caregiving, with an accompanying workbook, and 5 weekly telephone coaching sessions. Wait list control: (n = 57) received the intervention following the study.
Outcomes for the intervention and wait list control groups significantly differed on depression, perceived stress, and blood pressure.
Differences in the predicted direction were sustained through followup assessments at 5 months postintervention, but interpretability of these findings is limited by baseline differences in some variables.      Tables 3-5). Significant differential improvement in psychological well-being was noted in the intervention conditions for the majority of studies (n = 36, 63%), followed by enhanced health and physiological functioning (n = 8, 14%), as well as improved caregiving skills (n = 8, 14%). Less common findings included significant improvements in social functioning (n = 6, 11%), quality of life (n = 5, 9%), and problem-solving/goal setting (n = 3, 5%) for the intervention groups as compared to comparison/control group counterparts. Fifteen of the 57 (26%) included studies did not show significant improvement in caregiver outcomes following intervention compared to the comparison/control groups (Tables 3-5).

Intervention Components.
Among the five types of components implemented in the 33 caregiver telephone interventions, skills training procedures (n = 27, 82%) were deployed most often, the bulk of which were delivered to individual caregivers (n = 22, 67%), followed by the use of group format (n = 3, 9%) or a combination of group and individual training (n = 2, 6%). The foci of skills training included communication enhancement techniques, disease-specific caregiving methods, monitoring and management of care recipient's physical and psychological well-being, care recipient safety, and fostering caregiving self-care activities.

Outcomes.
Eleven of 14 (79%) web studies reported significant improvements on target outcomes in the intervention groups as compared to little or no change in the comparison/control groups. Three (21%) of 14 web studies reported nonsignificant intervention effects on target outcomes (Table 4). Differential increases in psychological well-being (e.g., decreases in depression, anxiety, and stress) were the predominant outcomes in the intervention arms across the 11 studies (79%). Other significant outcomes were obtained for caregiving skills (i.e., improved caregiving selfefficacy; n = 1, 7%) and quality of life (i.e., increased positive appraisals of caregiving; n = 1, 7%; improved quality of life; n = 1, 7%) ( Table 4).

Applications and Use
Control. Eight studies (57%) provided useful web links and resources (e.g., online educational materials, tip sheets [54] and asynchronous modules [55]) to supplement the intervention. Email was incorporated in five studies (36%) to enhance communication with the research team [48,52,54,56] and to provide materials to participants [47]. Five studies (36%) used discussion forums/online chat to facilitate communication and to encourage peer support [48,49,51,52,56]. Telephone calls independent of the intervention were used to follow up with participants in one study (7% [54]). Eight of the web interventions were self-guided (57%) and six (43%) were facilitator-guided (Table 6).

3.4.3.
Outcomes. All ten combined telephone and web studies (100%) reported significant caregiver improvements on target outcomes in the intervention conditions as compared to little or no change in the comparison/control conditions (Table 5). Improvements in psychological well-being (e.g., decreased depression and stress; increased coping) in the intervention conditions were reported across 9 of 10 studies (90%). Enhanced social/family functioning (e.g., increased perceived social support) also was found among intervention participants in two studies (20%). Significant gains in caregiving skills (e.g., increased caregiver self-efficacy) were reported by intervention caregivers in two studies (20%). Similarly, greater improvement in well-being (e.g., increased participation in self-care activities) also was obtained in the intervention conditions as compared to comparison/controls counterparts in one study (10%). Finally, improvement in quality of life was found among intervention participants for 1 of 10 studies (10%) ( Table 5).

Applications and Use
Control. Discussion forums/online chat were used in four studies (40%) to foster communication and provide support [60,61,63,67]. Four studies (40%) provided useful web links and resources to supplement the intervention [60,62,65,67]. Three studies (30%) included telephone calls with caregivers that were independent of the intervention [62,65,67], whereas two studies (20%) used email to communicate with caregivers [61,65]. Online journaling was used in one study (10%) to facilitate documentation of psychological and physical challenges related to caregiving and care recipients' health [67]. Six studies (60%) were facilitator-guided and four (40%) were self-guided (Table 6).

Discussion
In examining the descriptive findings across modalities, telephone delivery was the most used approach. One possible explanation for the disproportionate use of telephone across studies lies in the demographic characteristics of caregivers of persons with dementia and stroke, who were more likely to International Journal of Telemedicine and Applications be older adults and in turn, slower adopters of web technologies. In contrast, caregivers of children/adolescents were generally younger parents who incorporated web technologies into their daily lives and thus more likely to participate in web intervention. Of the six studies with caregivers of children, adolescents, and young adults, web modalities were used most often (n = 5, 83%), followed by combined telephone and web modalities (n = 1, 17%).
Next, over 70% of included studies focused exclusively on caregiver intervention rather than the caregiver-care recipient dyad. This decision may have been predicated by the diminished capacity of the care recipient to benefit from intervention, logistical issues, and limited exposure to dyadic treatment. First, over 50% of care recipients across the 57 studies were older adults with neurological disorders. Cognitive deficits, such as short-term memory loss and difficulties with verbal comprehension and expression, may pose obstacles in care recipient engagement in telehealth intervention [2,11]. Second, logistic difficulties may have led to limited implementation of interventions involving both caregivers and care recipients. For example, organizing dyadic sessions may be problematic due to scheduling conflicts, competing job demands and collateral family obligations. Last, dyadic intervention requires specialized therapeutic skills many telehealth providers may not possess.
Regarding caregiver intervention strategies, skills training was used in 75% of studies across modalities. In contrast to previous research [11], both telephone (n = 27, 81%) and web approaches (n = 11, 78%) showed similar utilization rates in deploying skills training strategies. Prior research [11] reports disproportionately lower integration of skills training approaches in web interventions as compared to telephone modalities.
The format and methods of delivery of psychoeducational and resource materials varied considerably across modalities. In telephone interventions, caregivers received information booklets, manuals, or videos through the mail [14,21,27]. In web and combined telephone and web modalities, psychoeducation and resource materials were delivered online using modules, interactive platforms, and video vignettes [47,48,62]. Relatively high utilization rates of psychoeducation and resource information were noted across telephone, web, and combined telephone and web interventions (i.e., 61%, 79%, and 90%, respectively). This pattern was similar to that previously reported [11]. Overall, the technologies used in the interventions included in this review were basic, easy-to-use, and low-cost, such as telephone, videoconferencing [51,58], DVD, web-based modules [33,46,50,62], and interactive web-based platforms [47,61]. However, one intervention incorporated a computer-mediated interactive voice response system [63] that required more sophisticated programming than other technologies highlighted in this review.
Self-monitoring/tracking was used infrequently across all modalities (n = 14, 25%). Limited incorporation of this component was difficult to interpret, particularly in light of its importance in guiding judgments about effectiveness of the treatment process. Furthermore, a more pronounced pattern of disuse was found for reminders. Only 7 of 57 studies included reminders to enhance attendance and adherence to intervention. A plausible explanation may be self-monitoring, and tracking systems were used in these studies, but not reported.
Next, caregiver applications were integrated into intervention for 21 of 57 studies (37%). Of the telephone interventions, only the independent phone call application was incorporated during intervention. However, the utilization rate was low (15%). Surprisingly, none of the telephone studies used text messaging. This finding may have its origin in lower adoption rates of digital technologies and/or use of text-based communication among caregivers of persons with dementia and stroke, a large proportion of whom were older adults.
In contrast, four applications were used in the web interventions and five with the combined telephone and web interventions ( Table 6). Except for online journaling [67], both web and combined telephone and web interventions adopted similar caregiver applications. Useful links and resources were used in 57% of web applications and 40% of combined telephone and web interventions. Discussion forums were implemented in web and combined telephone and web interventions at similar rates (36% and 40%), respectively. In contrast, email was used more often in the web studies (36%) compared to combined telephone and web modalities (20%). This finding suggests telephone may have replaced email use in the latter telecommunication approach.
Last, a discrepant pattern for user control was noted across the three intervention modalities. Ninety-four percent of telephone and 60% of combined telephone and web interventions were facilitator-guided. In contrast, most web interventions (57%) were self-guided. Web technologies may be more appropriate for caregivers who are technologically savvy and require less guidance in assimilating information acquired through this modality.
Regarding intervention outcomes, significant improvements in at least one target caregiver outcome were found in 74% of the studies across all three modalities. Consistent with prior research [2], improvement in caregiver psychological well-being was the most frequently reported outcome (48%) in the telephone interventions. Significant improvements in psychological well-being also were obtained in the web (79%) and combined telephone and web interventions (90%), corroborating prior findings [11]. However, it is difficult to interpret the underlying reasons for increasing percentages of positive outcomes in psychological well-being across categories. This issue requires investigation in future studies.
Next, a discrepant pattern of results was found for physiological outcomes. Statistically significant improvements in health and physiological functioning were evidenced in both telephone (18%) and combined telephone and web methods (10%), but little or no change on these variables was found in the web interventions. This discrepancy may have been due to the self-guided nature of web modalities. In the absence of facilitator feedback, tailoring of intervention to bolster health and physiological functioning was precluded.
While this review provides an excellent summary of telehealth research through January 2017, this review does not 41 International Journal of Telemedicine and Applications capture more recent telehealth studies. However, since our review, there have been no published systemic reviews of telehealth interventions for caregivers of the chronically ill which encompasses a variety of study populations, interventions, and outcomes, as does this review. Further, although the current systematic review consisted of a variety of telehealth interventions, four major limitations were noted in the included studies. First, the overall pattern of intervention outcomes may have been skewed by the incorporation of telephone in two of three intervention categories. Although telecommunication technologies are increasing in popularity, a large population of caregivers continues to face considerable barriers in accessing interventions using these modalities, particularly older caregivers and those who reside rural geographical locations. Unmet technology training needs, high cost of equipment, and barriers in gaining access to Internet service, especially among older adult caregivers, may have accounted for the disproportionate use of telephone in caregiver telehealth studies [70]. Second, web and combined telephone and web interventions were conducted exclusively with caregivers of children, adolescents, and young adults. Telephone interventions were not performed with this caregiver population. Therefore, our knowledge of the efficacy of telephone interventions across the age continuum was restricted in this review. Third, although telehealth modalities hold considerable promise in increasing access to hard-to-reach populations, the majority of studies in this review did not specify caregivers' and care recipients' primary area of residence (e.g., rural or urban). This shortcoming limited evaluation of the impact of geographic location on treatment outcomes. Last, a large proportion of studies incorporated interventions to enhance caregiving and selfcare skills. However, the impact of these activities on psychological functioning and health status as well as their level of implementation was not measured.

4.1.
Research and Practice Implications. The overall findings of this review provided considerable support for the efficacy of telehealth interventions across a variety of psychosocial and emotional outcomes in diverse populations. A major strength of the current systematic review was the selection of studies using randomized controlled designs. This decision was predicated not only on the need for increased rigor in drawing causal conclusions but also on the growing use of telehealth interventions in caregiver research [2,11].
Despite these encouraging developments, only a limited number of studies compared the effects of telehealth intervention versus an in-person comparator. For example, Davis et al. [18] compared the effects between telephone caregiver coping skills training and in-person intervention using the same protocol on changes in psychosocial functioning for dementia caregivers. The investigators found positive and equivalent gains in caregiver burden and distress at three months across both treatment conditions. In contrast, none of the web or combined telephone and web studies tested the differential impact of these modalities against similar in-person interventions. Thus, more research is needed to evaluate the efficacy of telehealth modalities versus in-person intervention delivery. Further, research is needed to test the equivalence or noninferiority of the effects of telecommunicated intervention versus traditional in-person interventions. The findings of such investigations will provide valuable information about the strengths and limitations of deploying specific telehealth delivery methods across different caregiver populations, caregiving challenges, and geographic and socioeconomic conditions. This review found that telehealth modalities can be used successfully in delivering a variety of skills training and psychoeducational interventions for family caregivers. With recent advances in telehealth technologies, clinicians can conduct a wide variety of psychosocial and behavioral interventions that support caregivers' emotional functioning regardless of location and proximity to healthcare providers. Clinicians also have access to a wide array of online resources, videos, interactive modules, and other healthpromoting technologies to enhance caregivers' skills in stress management, emotional well-being, and their health status. Although useful for all caregivers, telehealth modalities may be particularly beneficial in reaching rural caregivers, who are less likely to attend caregiving workshops due to financial or transportation issues.
Furthermore, telehealth intervention offers an effective method for promoting increased self-care in distressed caregivers. In a study by Piette et al. [65], caregivers received weekly, automated support calls with the option of obtaining guidance about care recipients' problems from an interdisciplinary team coupled with email suggestions for enhancing caregiver's self-care. Findings showed that caregivers who received the support calls reported significantly less caregiving strain compared to caregivers in the control conditions. Telehealth intervention holds considerable promise in addressing efficiently and effectively both caregiving challenges and the self-care concerns of family caregivers where and when assistance is needed. Therefore, clinicians should consider incorporating telehealth modalities for caregiver correspondence to assess emotional health and provide an avenue for caregivers to discuss support needs between scheduled healthcare visits.

Conclusion
The advancement and increased utilization of telehealth technologies offer unique opportunities to improve caregiver emotional functioning, health status, and caregiving skills. Findings of this review suggest that telehealth modalities are effective tools in delivering a variety of caregiver interventions, with significant improvements on several target caregiver outcomes across studies. Telehealth has been used most often to deliver skills training strategies to improve caregiver psychological well-being and enhance caregiver support. However, further research is needed to evaluate the efficacy of telehealth interventions in improving caregiving skills and caregiver self-efficacy, as well as caregiver health outcomes. Likewise, future studies need to examine the benefits of telehealth interventions with underserved, rural-residing caregiver populations. 42 International Journal of Telemedicine and Applications