Description of Included Studies by Focused Area of Research

In the results, we organize the findings for the 211 references by the 7 areas of research shown in Table 2. Please refer to the reference list for full study citations and to Appendix C for study characteristics of all included primary studies. Appendix D lists the primary outcomes by focused area of research.

Table 2

Studies by Focused Area of Research.

Study Characteristics and Demographics

Of the 81 studies in the maternal health area of research, 21 focused on prenatal health (eg, labor preparation and management and reducing substance use during pregnancy),^{22,37,39,41,42,48,55,57,64,70,72,73,76,79,80,85,86,91,106,108,109} 16 on lactation,^{21,35,43–45,50,51,56,59,60,67,74,90,102,105,110} 14 on mental health issues related to pregnancy (eg, postpartum depression and parenting stress),^{52,62,63,65,66,75,77,78,83,84,88,89,104,107} 9 on issues during the postpartum period (eg, maternal and infant care and maternal physical activity),^{36,40,47,61,68,81,92,100,101} 8 on smoking cessation during the peripartum period,^{33,46,54,69,71,82,87,103} 8 on weight management during or after pregnancy,^{38,53,94–99} and 5 on issues related to gestational diabetes.^{17,49,58,93,111}

Most studies (n=78) were RCTs. The single SR¹¹² included 2 RCTs and 2 nonrandomized trials. The greater proportion of maternal health studies included 500 or fewer participants (n=54; 66%); 14 studies (17%) had 501 to 1,000 participants, and 13 studies had more than 1001 participants (16%). Most studies in this focus area included women of a mean age of 20-39 years (n=65); 10 studies reported age without a measure of central tendency (12%), and age was not reported in 6 studies (7%). Just over half of included studies (n=44, 54%) did not report the racial/ethnic make-up of participants. Twenty-one studies were majority white (26%), 6 were majority black (7%), and 10 had another race as the most common (12%) (Figure 2, panel I). Most studies were conducted in World Bank high-income countries (n=66), with 37% (n=30) conducted within the United States specifically; 13 studies were completed in middle-income countries and 2 in low-income countries. Sixty-two studies recruited patients from the outpatient or community setting (77%), while 17 studies recruited from inpatient areas (mostly labor and delivery floors); 2 studies recruited from other settings (eg, military base and national database) (Figure 1, panel II). The single study that recruited from a military base was conducted within a hospital for active military addressing smoking cessation during pregnancy.⁸²

Intervention Details

We also categorized studies based on the prominence of the role of telehealth in the study (central or non-central). Most studies employed a telehealth modality as a central feature of the intervention (n=51, 63%); and for most of the studies the telehealth modality chosen was telephone (n=70, 86%) (Figure 2, panel III). Of those studies that did not use telephone as the primary modality, 5 used text messaging or SMS,^{74,76,79,87,92} 3 used an interactive website,^35,37,55 and 1 each used Facebook^©,⁹⁰ a mobile application,¹⁰⁶ and videoconferencing.¹⁰⁰ Seven studies reported using a secondary telehealth modality. Five studies added telephone communication as a second modality: 1 combined phone with Facebook^©,⁹⁰ 2 with text messaging,^79,111 1 with an interactive website use,³⁷ 1 with videoconferencing,¹⁰⁰ and 1 with computer algorithm-delivered communication.⁹² Only 8 studies used asynchronous communication (eg, email exchanges and online discussion boards) and the rest were synchronous (eg, telephone counseling). Most studies used telehealth modalities to communicate between a health care team member and the patient (n=72, 89%) compared with only 7^{35,55,71,74,76,87,92} that used a computer algorithm to tailor communication to individual patients and 2 that used multiple communication strategies (Figure 2, panel III).

There was a wide variety of credentials and expertise among the interventionists in this area of study. The 3 most common types of interventionists were registered nurses (n=20; 25%), midwives (n=13; 16%) and peer or lay health workers (n=12, 15%) (note that midwives and peer/lay workers are included in the “other” category in Figure 2, panel IV). The most common intervention length was 12 weeks or less (n=34; 42%), 31 studies were 13-52 weeks, and only 1 study was longer than 52 weeks and addressed lactation and infant feeding.⁷⁴ Fifteen studies did not report the intervention length. Just under half of telehealth studies about pregnancy and maternal health reported adherence to the intervention (n=39, 48%). Twenty studies for this focus area reported using a theoretical framework for their intervention. Of those theories reported, the most common was Social Cognitive Theory²²⁸ (n=6)^{33,47,71,85,87,93} and the Transtheoretical Model²²⁹ (n=2).^71,86

Outcomes Evaluated

Sixty-nine studies (85%) designated a primary outcome (Appendix D). No studies reported a primary outcome that was focused on provider issues (eg, provider satisfaction). Common specific primary outcomes at the patient level were breastfeeding (n=13), mental health symptoms (n=14), pregnancy outcomes (n=12), and smoking cessation (n=8). The system-level primary outcomes were utilization, quality-of-care indicators, economic outcomes, and access to care. Of all outcomes measured and reported, the most common type was patient-level outcomes (n=65, 80%) followed by patient- and provider-level outcomes (n=3) (Figure 2, panel V).

Figure 2

Evidence Map for Maternal Health Area of Research.

Systematic Review Findings

The single relevant SR that focused on maternal mental health evaluated web-based treatments with interventionist support for perinatal mood disorder.¹¹² Of the 4 studies included in the SR, 2 RCTs met our inclusion/exclusion criteria (n=1138). Kersting et al (2013; n=228) recruited German participants of mean age 34 years after a loss of pregnancy. Response to a 5-week intervention consisting of ten 45-minute web-based writing assignments (with therapist feedback) led to a significant decreases in grief, depression, and anxiety compared with a waitlist control (with continued improvement in depression scores at 3- and 12-month follow-up). In the second study, O’Mahen et al (2013; n=910) conducted a study in the United Kingdom with postnatal women of mean age 32 years. The 15-week intervention included eleven 40-minute online sessions with homework focused on behavioral activation, and also provided access to weekly text-based synchronous chat with an interventionist. Depressive symptoms decreased for more participants in the treatment group when compared with those receiving treatment as usual.

Subcategories Within Maternal Health Studies

To further elucidate studies focused on maternal health, we examined the studies by smaller areas of focus within this category. Specifically, we characterized interventions for the following 7 subcategories: prenatal care, lactation, mental health, postpartum, smoking cessation, weight management, and gestational diabetes (Figures 3–6).

Gestational Diabetes

Gestational diabetes was addressed in 5 trials (some trials also included women with type 1 diabetes, and one included women with gestational diabetes or type 2 diabetes).^49,111 Two studies intervened to reduce postpartum consequences for women with gestational diabetes,^58,93 2 addressed diabetes management during pregnancy,^49,111 and 1 aimed to prevent the development of gestational diabetes among women at risk.¹⁷ Four of these 5 studies were RCTs and 1 was a nonrandomized trial.⁴⁹ Study sizes ranged from 100 to 2280 participants, and 2 were conducted in the United States.^58,111 All used telephone as the telehealth modality, and 2 studies used it as a central part of the intervention.^49,111 There was 1 study that used physicians to deliver the intervention,⁴⁹ which was relatively unusual across the larger field of women’s telehealth literature. Three of the 5 studies lasted 13 to 24 weeks,^17,49,58 and 2 were 12 weeks or less.^93,111 Primary outcomes for gestational diabetes studies were patient-level when reported (n=4)^17,58,93,111 and included weight,^17,58,93 fasting glucose,^17,93 and mean self-monitoring blood glucose compliance rate.¹¹¹

Figure 3

Evidence Map for Maternal Health Subcategories (Study Size, Age, Race).

Figure 4

Evidence Map for Maternal Health Subcategories (Modality, Centrality, Dyad).

Figure 5

Evidence Map for Maternal Health Subcategories (Income, Recruitment Location).

Figure 6

Evidence Map for Maternal Health Subcategories (Interventionist, Length of Intervention).

Study Characteristics and Demographics

The second largest area of telehealth interventions for women focused on prevention (n=56). These included 18 studies on increasing physical activity,^{18,30,118,126,127,129,132,133,137,138,140–143,149,150,155,158} 14 on cancer screening,^{29,32,115,120,123–125,130,134,135,148,151,157,162} 11 on weight management,^{113,114,117,122,128,131,136,139,152,159,163} 10 on smoking cessation,^{116,119,121,144,146,147,156,160,161,164} and 3 on diet.^145,153,154 Nearly all of these studies were RCTs, with only 1 nonrandomized trial, a controlled before-and-after study.¹⁵⁹ Most studies had 500 or fewer participants (n=36; 64%), yet 21% had more than 1000 participants. The most common age category for this focus area were women with mean ages from 50 to 59 years (n=21; 36%), with only 4 studies focused on women with a mean age 60 years or older. Eight studies reported age with no measure of central tendency. In nearly half the studies, the recruitment populations were majority white (n=26; 46%), and 8 studies recruited populations that were majority black; 30% did not report race or ethnic data (Figure 7, panel I). Nearly all studies were conducted in countries categorized as high income by the World Bank (n=55), which included 43 from the United States, with only 1 study conducted in middle-income and none conducted in low-income countries. The most common recruitment site was the outpatient setting (n=25; 45%),^{29,115,117,121–124,128,129,131,133,134,136,140,144,145,148,150–152,154–156,163} followed by community settings n=23; 41%).^{18,30,113,114,118,120,122,125,127,130,132,137–139,141–143,149,153,157,161,162,164} One study recruited from a Navy base and addressed smoking cessation for new recruits¹¹⁹ (Figure 7, panel II).

Intervention Details

For prevention-focused studies, 61% used telehealth modalities as the central intervention strategy (n=34). The overwhelming majority used synchronous, person-to-person communication via telephone as the primary mode of telehealth delivery (n= 54; 93%). For the 4 studies that did not use telephone as the primary modality, 2 used texting,^127,143 1 used a mobile application,¹⁶⁴ and 1 used an interactive voice response (IVR) system¹⁶³ (Figure 7, panel III). The most common intervention length was 25-52 weeks (n=23; 40%). However, 18% of studies were only 12 weeks or less in duration (n=10). Conversely, 6 studies were 52 weeks or more (Figure 7, panel IV). We also categorized studies based on the participants involved in the telehealth communication (ie, the treatment dyad). For studies in this focus area, the majority (n=48; 86%) were focused on health care team member-to-patient communication. Five used telehealth interventions to direct strategies guided by computer algorithms to patients,^{121,127,133,135,164} 1 focused on communications from one provider to another provider,¹⁴⁶ and 2 were focused on a mix of provider-to-patient plus IVR communications¹⁶³ or provider to patient plus provider to provider communications.¹⁴¹ Many types of individuals served as interventionists in prevention-focused studies with a wide mix of backgrounds and credentialing. Many studies used nonmedical professionals such as community worker or peer support personnel (n=9), health educators (n=5), health coaches (n=3), and a combination of other nonmedical professionals with unspecified credentialing or training (eg, interventionist trained by study staff) (n=14) (Figure 7, panel IV). There was also a wide variety of other credentialed professionals such as dieticians (n= 6), behavioral health specialists (n=3), and registered nurses (n=6) that served as interventionists. However, no physicians, nurse practitioners, or physician assistants served as interventionist in any of the identified studies. We also assessed if studies provided data on adherence to the intervention; for prevention-focused studies, 53% did not provide this information (n=30). Also the majority of studies did not delineate a theoretical framework (n=33; 59%). Of those that did, the most commonly named models were the Social Cognitive Theory²²⁸ (n=14) and the Transtheoretical Model²²⁹ (n=9).

Outcomes Evaluated

For prevention-focused studies, all primary outcomes were on the patient level (Appendix D). No studies reported provider- or system-level outcomes as a primary outcome. Of those studies designating a primary outcome (n=49, 85%), the most common primary outcomes were cancer screening rates (n=12) and physical activity (n=11), followed by smoking cessation (n=8) and weight (n=7). When looking at secondary outcomes for the prevention-focused studies, there were 6 studies that included additional secondary outcomes at the provider (n=2) or system level (n=4) (Figure 7, panel V).

Figure 7

Evidence Map for Prevention Area of Research.

Study Characteristics and Demographics

Forty-three (20%) telehealth interventions were focused on disease management. The vast majority were RCTs (n=41; 95%); we identified only 2 nonrandomized studies. The majority of studies had 2 arms (n=26; 60%), with 10 studies having 3 arms and 7 having 4 arms. Sample size ranged from 100 to 6591, with most studies having 1000 or fewer participants (n=38; 88%), and only 5 studies having more than 1000 participants.^{19,168,185,186} The mean age ranged from 30 to 80 years. The most common age category was middle-aged women of mean age 40-60 years (n=24), with only 8 studies focused on women with mean age over 60 years (n=9, 20%); 8 studies did not report mean age and/or did not report age at all. Only 23 (53%) studies reported race and ethnicity, and across these studies the most commonly reported category was white (n=18, 41%) (Figure 8, panel I). Except for 1 study, which was conducted in Iran, a middle-income country,¹⁹⁷ all other studies were conducted in countries categorized as high income by the World Bank. Twenty-five studies (58%) were conducted in the United States.^{19,20,23,26–28,31,34,167–169,171,174–177,182–186,192–194,196} The most common recruitment site was specialty care settings (n=20, 45%) followed by community settings (n=9, 21%) and inpatient settings (n=3, 7%) (Figure 8, panel II).

Intervention Details

The most common intervention length was 12 weeks or less (n=20, 47%). Six studies were between 13 and 24 weeks,^{23,26,34,169,176,193} 4 studies between 24 and 52 weeks^{174,177,182,192} and 6 studies were 52 weeks or more.^{24,25,31,185,186,195} Six studies did not report intervention length. For disease management studies, 77% used telehealth modalities as the central intervention strategy (n=33). The most common mode of telehealth delivery was telephone (72%; n=31 studies). For the 9 studies that did not use telephone as the primary intervention modality, 7 used interactive website,^{23,172,173,177,179,189,191} 1 one used secure email,¹⁹⁰ and 1 used a mobile application.¹⁸⁰ Three studies use more than one modality: 1 added videoconferences to telephone communication as a second modality,²⁰ and 2 combined telephone with an interactive website^26,34 (Figure 8, panel III). Overall, 84% (n=37) used synchronous communication, while 5 used asynchronous,^{179,180,189–191} and 2 used a hybrid approach.^19,172

We also categorized studies based on the participants involved in the telehealth communication. Telehealth communication was most commonly conducted between a health care team member and a patient (n = 37; 84%), followed by mixed or multiple types of communication with patients (n = 6; 14%).^{19,23,26,165,179,189} One study used a computer algorithm in patient communication.¹⁷² The most common type of interventionist was registered nurse (n=15, 35%). Three studies used a mix of interventionist (nurse and/or behavioral health specialist,^31,180 nurse and computer algorithm¹⁷³), and 2 used computer algorithms^34,172 (Figure 8, panel IV). We also assessed if studies provided data on adherence to the intervention; for this focus area, only half the studies provided this information (n=21). Twenty-six studies endorsed using various theoretical frameworks for their intervention, including a health promotion model²³⁰ (n=1¹⁷⁷), acceptance and commitment therapy²³¹ (n=1¹⁸⁰), and social cognitive theory²³² (n=1¹⁷⁹), among others.

Outcomes Evaluated

Of the studies that reported a primary outcome, the majority had a patient-level primary outcome (n=35; 81%), 4 studies focused on system-level outcomes, and 4 studies did not report any specific primary outcome (Appendix D). Studies with a patient-level primary outcome mainly measured quality of life (n=13), clinical symptoms (n=14), and medication or treatment adherence (n=3, 7%).^19,174,195 Studies with a system-level primary outcome measured cost,¹⁸⁹ utilization,¹⁷⁶ quality-of-care indicators,¹⁶⁸ and economic outcomes.²⁴ When looking across all outcomes reported (eg, primary and secondary outcomes), the majority reported patient-level outcomes only, 2 reported system-level outcomes, 1 reported both patient- and provider-level outcomes, and 2 reported patient- and system-level outcomes (Figure 8, panel V).

Figure 8

Evidence Map for Disease Management Area of Research.

Study Characteristics and Demographics

Eleven studies focused on telehealth interventions for family planning, addressing issues related to choice of contraception and contraception adherence (n=5, 45%),^{198,199,201,206,207} post-abortion care (n=3, 27%),^202,203,205 infertility (n=2, 18%),²⁰⁰ and fertility and pregnancy among cancer survivors (n=1, 9%).²⁰⁴ Ten studies were RCTs, and 1 was a nonrandomized trial.²⁰¹ Study sizes ranged from 108 to 1,433 participants; 3 of the 4 largest studies, all with over 1,000 patients, addressed contraception choice and contraception adherence.^198,199,201 As expected, the most common age category for this focus area was 20-39 years (n=8, 73%); 2 studies did not report age in a comparable way.^198,205 Of note, 1 study had a mean age range between 51 and 60 years²⁰⁴ – this was a peer counseling study for African-American women survivors of breast cancer and addressed survivorship concerns across the reproductive spectrum from infertility and pregnancy to menopausal symptoms. Across the 10 studies, race was mostly not reported (n=8, 73%), while 1 study each was majority black,²⁰⁴ white,¹⁹⁹ and other¹⁹⁸ (Figure 9, panel I). The location in which the family planning studies were conducted was split between high-income (n=6, 55%)^{199,200,204,206–208} and middle-income countries (n=4, 36%),^{201–203,205} with only 1 study from a low-income country (Uganda).¹⁹⁸ Four studies were conducted in the United States.^{199,204,206,208}

Intervention Details

We also examined characteristics of the intervention type. As with other intervention focus areas, most of these studies used telephone as the mode of telehealth delivery (81%, n=9), while 1 used video conference to home,²⁰⁰ and 1 used an interactive website.²⁰⁶ Seven of the 11 studies featured a telehealth modality as a central feature of the intervention,^{199–201,205–208} while for 3 it was ancillary^198,202,204 and nominal in 1.²⁰³ The majority of family planning studies used telehealth for communication between members of the health care team and patients; only 1 study used a computer algorithm to communicate with patients²⁰⁶ (Figure 9, panel II). Most studies utilized synchronous communication between involved parties, while 1 used asynchronous communication²⁰⁰ and 1 used a hybrid approach.²⁰⁵

The interventionists varied widely. Two studies used a peer or lay health worker,^198,204 2 used a health educator,^201,206 1 used a social worker,²⁰⁸ and 1 study each used the following interventionist types: unspecified counselor,²⁰⁵ provider,²⁰⁰ registered nurse,²⁰² unspecified nonprofessional,²⁰³ and multiple types of interventionists.¹⁹⁹ The credentials of the interventionist were not reported in 1 study.²⁰⁷ Of those studies that reported the length of the intervention, the most common length was 12 weeks or less (n = 4),^{203,205,207,208} with 2 lasting 25-52 weeks,^201,204 and 1 lasting 13-24 weeks;¹⁹⁹ 4 studies did not report the length of the studied intervention^{198,200,202,206} (Figure 9, panel III). Of the studies using telephone as the telehealth modality, only one study reported the length of the telephone calls (30 minutes).²⁰⁴ Of the telephone-based studies that reported the number of phone contacts, 4 consisted of only 1 phone call,^{202–204,207} 1 had 2 phone calls,²⁰⁸ and 2 had 6 phone calls.^199,205 Most studies did not report adherence to the study protocol (n=7, 64%),^{199,200,203–205,207} and only 1 study¹⁹⁹ reported basing the intervention design on an established theory, the Health Belief Model.²³³ All recruitment for the family planning studies occurred in either the outpatient or community setting (Figure 9, panel IV). There were no VA-based studies.

Outcomes Evaluated

All family planning studies had a patient-level primary outcome (Appendix D). The most common primary outcome was medication adherence (n=4), followed by physical symptoms (n=2), contraceptive use (n=2), mental symptoms (n = 2), and patient satisfaction (n=1). One study included a system-level outcome as a secondary outcome (Figure 9, panel V).

Figure 9

Evidence Map for Family Planning Area of Research.

Study Characteristics and Demographics

We identified 7 telehealth interventions related to identifying and managing women at high risk for breast cancer.^209–215 These included 5 interventions designed to test phone-based versus in-person genetic counseling,^210–214 1 providing emotional support to women with known BRCA1/BRCA2 mutations,²¹⁵ and 1 targeting the sisters of young women with breast cancer to promote appropriate risk assessment, increase knowledge about risk, and reduce worry related to breast cancer risk.²⁰⁹ All of these studies were individual (n=3)^209,214,215 or cluster RCTs (n=4).^210–213 The size of studies ranged from 100 to 1,012 participants. The most common age category was 40-49 years (n= 4)^{209,211,214,215} and 50-59 years (n=3).^210,212,213 Most studies had a predominant racial group that was white (n = 6),^209–214 with 1 study that did not report racial composition²¹⁵ (Figure 10, panel I). All breast cancer risk studies were conducted exclusively among high-income countries. Most studies were set in the United States (n=6), while 1 was set in Australia.²¹⁵ Participants were recruited from a variety of settings including specialty care (n=3),^211,214,215 primary care (n=1),²¹⁰ community-based settings (n=2),^209,213 and a cancer registry²¹² (Figure 10, panel II).

Intervention Details

All studies were synchronous, telephone-based, and featured telehealth as the central intervention modality. Most studies were designed for information flow from the health care team to a patient, and one study sent information from a computer algorithm to a patient²¹⁴ (Figure 10, panel III). The interventionist varied across studies. Providers were most frequently genetic counselors (n=4),^{210,212–214} followed by behavioral health specialists,²⁰⁹ peer or lay health workers,²¹⁵ and physician or advanced practice nurse.²¹¹ Only 1 study clearly described the length of the study, which was 13-24 weeks; this intervention provided social support to women with known BRCA1/BRCA2 mutation.²¹⁵ Four studies did not specify the length of the intervention but focused on the number of telephone counseling sessions (ranging 1-3 calls),^{210–212,214} and 2 studies did not report the number of calls^209,213 (Figure 10, panel IV). Three of the 7 studies reported adherence to the intervention protocol.^210,214,215 We also evaluated whether studies included information about a theoretical framework. Most breast cancer risk studies did not provide a theoretical rationale (n=6); however, 1 study²⁰⁹ referenced the Health Belief Model,²³³ Self-Regulation Theory,²³⁴ and Transtheoretical Model.²²⁹

Outcomes Evaluated

For this focus area, 4 studies had primary outcomes at the patient level including mental health symptom assessments^209,211,215 and patient satisfaction,²¹⁴ while 3 studies had system-level primary outcomes including cost²¹⁰ and utilization.^212,213 No breast cancer risk studies included provider-level outcomes (Appendix D). However, when accounting for both primary and secondary outcomes together, 2 studies included patient-level outcomes only, 2 had system-level outcomes only, and 4 reported patient- and system-level outcomes (Figure 10, Panel V).

Figure 10

Evidence Map for High-Risk Breast Cancer Assessment Area of Research.

Study Characteristics and Demographics

Six studies focused telehealth interventions on women’s mental health outside the postpartum period. All of these studies were 2-armed, individual RCTs. Four studies had between 100 and 250 participants^{217,219–221} and 2 had 251-500.^216,218 Of those studies that reported age of participants, 2 studies had a mean age of 20-39 years,^217,220 1 study 40-49 years,²²¹ and 1 study 50-59 years.²¹⁸ Participants in 3 studies were predominately white,^218,219,221 and 3 studies did not report participants’ race^216,217,220 (Figure 11, panel I). Most of the studies were conducted in countries categorized by the World Bank as high income (n=5),^{216–219,221} with most of these being in the United States.^218,219,221 One study was conducted in a middle-income country.²²⁰ Further, most studies recruited women from specialty care settings (n=4),^216–219 while 2 studies recruited from the community (Figure 11, panel II).^220,221 One study compared the delivery of cognitive processing therapy via in-person versus video teleconferencing for posttraumatic stress disorder and recruited both Veteran and civilian women through VA and community settings.²²¹

Intervention Details

Telehealth played a central role in 4 mental health studies.^{216–218,221} Further, all studies were synchronous; 5 used telephone as the primary telehealth modality, and 1 used videoconferencing²²¹ (Figure 11, panel III). For studies in this focus area, all centered on health care team-to-patient communication. For studies using telephone as the intervention modality, most involved repeated telephone calls (n=4), while a few studies reported only 1 or 2 calls (n=2).^217,218 For the 1 study using video teleconferencing, it featured a total of twelve 90-minute sessions.²²¹ The health care team member who served as interventionist varied and included nurses (n=2),^216,217 peer or lay health workers (n=1),²¹⁹ behavioral health specialists (n=1),²²¹ health educators (n=1),²¹⁸ and 1 study with multiple types of interventionists²²⁰ (Figure 11, panel IV). We also assessed whether studies address adherence to the intervention and 4 did.^216,218,220 Intervention length ranged from 1 contact²¹⁷ to 2-5 years.²¹⁶ Only 1 mental health study²²⁰ described a theoretical framework, specifically Cohen’s Social Support.²³⁵

Outcomes Evaluated

All primary outcomes were assessed at the patient level (Appendix D). The most common outcomes were depression and anxiety. No secondary outcomes reported were on the provider- or systems-level (Figure 11, panel V).

Figure 11

Evidence Map for Mental Health Area of Research.

Study Characteristics and Demographics

We identified 5 studies that focused on IPV.^222–226 All but 1 were conducted among women with a previous history of IPV, with the goal of reducing the risk of future experience of violence through approaches such as promoting safety behaviors,²²² providing social support,^224,225 problem-solving training^224,225 and empowerment²²³ through motivational interviewing. The fifth study was a trial of providing support to women who were recent survivors of rape to promote adherence to HIV post-exposure prophylaxis.²²⁶ All studies were RCTs and involved 1189 patients, while individual study sample sizes ranged from 150-307. Studies in this focus area primarily included women of reproductive age, with 3 studies reporting a mean age of 20-39 years (2 studies did not report a measure of central tendency for age).^223,226 The predominant race/ethnicity was white for 2 studies,^223,224 Latina for 1 study,²²² and not reported in 2 studies.^225,226 One study for which race/ethnicity was not reported was conducted in Hong Kong,²²⁵ and the other in South Africa²²⁶ (Figure 12, panel I). Of the 4 trials conducted in high-income countries, 3 were in the United States^222–224 and 1 in Hong Kong.²²⁵ One study was conducted in a middle-income country (South Africa).²²⁶ Two studies recruited from the outpatient setting,^223,226 1 from a pediatric emergency department,²²⁴ 1 from the community,²²⁵ and 1 from a special family violence unit within a district attorney’s office²²² (Figure 12, panel II). None recruited patients from VA settings.

Intervention Details

We also categorized studies based on the prominence of the role of telehealth in the study (central or non-central). Four studies used telehealth modalities as the central intervention strategy.^{222,224–226} All studies used telephone as the delivery modality (Figure 12, panel III), and all studies used synchronous communication. We also categorized studies based on the participants involved in the telehealth communication. The interventionist for these studies included 1 study with a nurse,²²⁴ mixed types of interventionists with nurses and a case worker,²²² 2 used unspecified nonprofessionals,^223,226 and 1 did not clearly report the type of interventionist.²²⁵ For studies in this focus area, 4 were focused on health care team to patient communication,^{222–224,226} while the fifth study focused on communication from an unspecified individual to a patient.²²⁵ Three studies had an intervention length of 12 weeks or less,^222,225,226 and 2 had interventions lasting 13-24 weeks^223,224 (Figure 12, panel IV). We also assessed if studies provided data on adherence to the intervention, of which 3 of 5 did.^223,224,226 Two studies noted using theoretical frameworks: 1 study²²⁵ named the Dutton Empowerment Model²³⁶ and Cohen’s Social Support,²³⁷ and the other study²²² named Walker’s 3-Phase Cycle of Violence²³⁸ and Curnow’s Open Window Phase.²³⁹

Outcomes Evaluated

For these studies, all stated primary outcomes were at the patient level (Appendix D). The primary outcome of 2 studies was clinical assessment of mental health,^223,225 1 used experience of violence,²²⁴ 1 measured adherence to post-exposure prophylaxis,²²⁶ and 1 measured safety-promoting behaviors.²²² All reported outcome measures (including primary and secondary outcomes) across these 5 studies were at patient level. (Figure 12, panel V).

Systematic Review Findings

One trial was identified from an SR that was also found individually. The SR examined telephone interventions for preventing new HIV infection across care settings.²²⁷ The relevant telehealth study meeting criteria for our evidence map was the RCT²²⁶ mentioned above that included 274 female rape victims who had presented to community clinics in South Africa. Enrolled women were randomized to usual care versus telephone support; support was delivered by a skilled counselor, and the primary outcome was patient level (medication adherence).

Figure 12

Evidence Map for Intimate Partner Violence Area of Research.