Technology use to facilitate health care among young adult transgender women living with HIV

Abstract

Little is known about how young adult transgender women living with HIV use digital technologies to facilitate their health care. This study examined the use of digital technologies to search for health information and support HIV care among young adult transgender women living with HIV (N = 130) in Los Angeles County, California. About half used the Internet “all the time” to search for transgender-specific resources (TSR; 53.8%) and for sexual health information (SHI; 51.5%). Less than half (39.2%) received digital HIV care reminders and, of those taking ART medication (n = 63), 36.5% received digital medication reminders. Internet information search was associated with Hispanic/Latina ethnicity (TSR: OR=0.23, 95% CI [0.09, 0.58]; SHI: OR=0.29, 95% CI [0.12, 0.73]) and higher (≥ $500) past-month income (TSR: OR=2.67, 95% CI [1.13, 6.34]; SHI: OR=2.67, 95% CI [1.14, 6.26]); receiving digital medication reminders with post-secondary educational attainment (OR=5.70, 95% CI [1.04, 31.19]) and higher income (OR=6.73, 95% CI [1.52, 29.67]). Receiving analog, but not digital, HIV care reminders was associated with engagement in HIV care (OR=2.37, 95% CI [1.13, 5.00]) and ART uptake (OR=2.18, 95% CI [1.06, 4.48]. Digital technology use was common for health-related searches but not for supporting HIV care.

Introduction

Transgender (hereafter: trans) women in the United States (U.S.) experience challenges to good health, and many lack access to trans-specific resources and appropriate healthcare. For example, many trans women seek gender-confirming medical treatment yet the limited availability of trans-competent and trans-responsive providers can be a barrier to care (James, Rankin, Keisling, Mottet, & Anafi, 2016; N. F. Sanchez, J. P. Sanchez, & Danoff, 2009). Consequently, trans women frequently report that they use non-prescribed hormones and rely on sources other than a physician for information about hormones (Clark, Fletcher, Holloway, & Reback, 2018; de Haan, Santos, Arayasirikul, & Raymond, 2015; N. F. Sanchez et al., 2009).

Research on the sexual health of trans women has demonstrated high rates of sexual risk behaviors, including condomless anal sex, sex while intoxicated or high, and engagement in sex work (Herbst et al., 2008; James et al., 2016; Poteat et al., 2015; Reback & Fletcher, 2014). These high-risk sexual behaviors, in combination with knowledge deficits and inadequate risk perception regarding HIV transmission, contribute to alarmingly high rates of HIV infection among trans women (Centers for Disease Control and Prevention, 2018; de Santis, 2009; Operario, Soma, & Underhill, 2008; Reback & Fletcher, 2014). A meta-analysis of biologically confirmed infection rates estimated HIV prevalence to be 21.7%, rendering trans women 34 times more likely to be living with HIV than adults in the general population (Baral et al., 2013). In Los Angeles County, one study demonstrated an increase in self-reported HIV prevalence of nearly 60% over a 17-year period, from 22% in 1998–1999 to 35% in 2015–2016 (Reback, Clark, Holloway, & Fletcher, 2018).

Among cisgender (i.e., individuals who identify with the sex they were assigned at birth) populations, adolescents and young adults aged 13 to 35 have the highest incidence of HIV infection (Centers for Disease Control and Prevention [CDC], 2018). Comparable age-stratified epidemiological data is not available for trans women, but findings derived from small, non-probability samples of young trans women indicate a high risk of HIV infection and transmission, especially for trans youth who identify as African-American/Black or engage in sex work (Garofalo, Deleon, Osmer, Doll, & Harper, 2006; Reisner at al., 2017; Schulden et al., 2008; Wilson et al., 2009).

Optimal HIV medical care along the HIV Care Continuum entails a continuum of services from timely linkage to outpatient care after diagnosis, to early initiation of antiretroviral therapy (ART), retention in care, and ART adherence, all of which are necessary to achieve and maintain undetectable HIV viremia (Gardner, McLees, Steiner, Del Rio, & Burman, 2011; Ulett et al., 2009). However, engagement and retention in care are impeded for the estimated 1.2 million persons living with HIV in the U.S, and this is particularly true for youth and young adults. According to CDC estimates, among those diagnosed with HIV in 2017, linkage to HIV care within one month of diagnosis decreased with age. Those aged 55 years and older had the best linkage to care rate within one month of diagnosis (81.6%), followed by those aged 45–54 years (81.4%), and those aged 13–24 evidenced the lowest linkage to care rate within one month of diagnosis (75.0%; CDC, 2019). HIV treatment outcomes are further compounded for trans women. Previous studies have demonstrated that trans women are even less likely to receive ART, to be ART adherent, and to be virally suppressed than cisgender persons living with HIV (Baguso, Gay, & Lee, 2016; Kalichman, Hernandez, Finneran, Price, & Driver, 2017; Melendez et al., 2006; Mizuno, Frazier, Huang, & Skarbinski, 2015; Yehia, Fleishman, Moore, & Gebo, 2013). Younger age is associated with poorer ART adherence among both trans and cisgender populations (Zanoni & Mayer, 2014; Mizuno, Beer, Huang, & Frazier, 2017; Sevelius, Saberi, & Johnson, 2014). A study that compared trans women youth and cisgender youth living with HIV found commensurate rates of ART uptake and viral suppression, but trans youth were less likely to report ART adherence (Dowshen et al., 2016).

There is growing research interest in developing technology-based health interventions, with a special emphasis on improving health outcomes along the HIV Care Continuum (Muessig, Nekkanti, Bauermeister, Bull, & Hightow-Weidman, 2015; Simoni, Kutner, & Horvath, 2015). Mobile devices such as cellular phones and smartphones, in particular, provide a versatile platform for delivering health services and information to persons living with HIV. Mobile health (mHealth) interventions hold special promise for marginalized, hard-to-reach populations such as trans women who face multiple and co-occurring structural and individual barriers to quality HIV care (Reback, Ferlito, Kisler, & Fletcher, 2015; Reback, Rünger, & Fletcher, 2019; Sevelius, Patouhas, Keatley, & Johnson, 2014; White Hughto, Murchison, Clark, Pachankis, & Reisner, 2016). Young trans women are likely to be especially receptive to mHealth HIV care interventions in light of the near-ubiquitous smartphone accessibility and widespread use of social media and social network sites among young adults (Jiang, 2018). However, to date, there is a dearth of knowledge concerning the use and utility of digital technologies in meeting the health needs of trans women living with HIV (Allison, Adams, Klindera, Poteat, & Wolf, 2014). A better understanding of how this population uses digital technologies for health-related purposes in their daily lives can provide valuable insights for designing effective, culturally responsive technology-based interventions.

This study carried out an analysis of baseline data from a randomized controlled trial that utilized a theory-based text-messaging intervention to optimize HIV Care Continuum outcomes among young adult trans women living with HIV. Specifically, this study focused on Internet use for the purpose of searching for trans-specific resources (e.g., hormones) and sexual health information, and the use of digital communication technologies (i.e., text messaging, mobile app, social media messaging) as a tool for HIV care support in the form of digital reminders to seek HIV care and to take ART medications. In addition, the study assessed whether the use of digital technologies varied by sociodemographic factors, and whether recourse to digital HIV care support was related to participants’ placement along the HIV Care Continuum.

Materials and Methods

Participants

Participants (N = 130) enrolled from November 2016 through May 2018. Eligibility criteria were (1) gender identity as a trans woman; (2) assigned a biological sex of male at birth; (3) between the ages of 18–34 years; (4) confirmed HIV-positive serostatus; (5) ability to receive daily text messages on either a personal cell phone or via an email account; (6) either tested HIV positive for the first time within the last 12 months, or had not had an HIV care visit in the previous 6 months, or had a viral load (VL) of ≥ 200 copies/ml on the last lab test result, or not currently prescribed ART medication, or was currently prescribed ART medication but did not rate her ability to take all her medications as excellent; (7) willing and able to provide informed consent; and, (8) willing and able to comply with study procedures.

Procedure

Participants were recruited from a community-wide effort in Los Angeles County, California that included street- and venue-based outreach, print media, online banner ads and digital flyers, poster advertisement posted onsite and at collaborating community-based organizations, and participant-incentivized snowball sampling. Following screening and informed consent, participants completed a baseline Audio Computer Assisted Self Interview (ACASI) assessment that included modules on sociodemographics, HIV Care Continuum, and health-related technology use (below). The baseline ACASI took approximately 60 to 90 minutes to complete. Participants were compensated with a $50 gift card for completing all the baseline procedures. All study procedures were approved by Friends Research Institute’s Institutional Review Board.

Measures

Sociodemographics

Participants reported their age, race/ethnicity, housing status, educational attainment, and income in U.S. dollars in the past month from all legal and illegal sources combined.

Health-related Technology Use

Participants rated how often they searched the Internet for sexual health information and for transgender-specific resources, using a ten-point Likert-type scale ranging from “Never” to “All the time.” Participants were asked who encouraged or reminded them to go for HIV care and to take their HIV medications in the past six months and which of the following means were used in each case: text messaging, mobile app, social media messaging, face to face, telephone.

HIV Care Continuum

Participants reported how many times they received health care for HIV in a doctor’s office or clinic in the past six months, whether they were currently taking ART or HIV medication, and the results of their last VL test. As a measure of ART adherence, participants rated their ability to take all their HIV medications as prescribed in the past month on a six-point Likert-type scale ranging from “Very poor” to “Excellent.”

Statistical Analysis

Dollar-amount income responses were dichotomized based on a median split into “< $500” versus “≥ $500.” “Housing instability” was coded as “1” if a participant reported that they lived in one of the following places: hotel, motel, boarding house, halfway house, drug treatment center, independent living unit, shelter, on the streets, in a parked car, in an abandoned building; otherwise the variable was coded “0.” Ratings on the frequency of Internet searches for trans-specific resources and sexual health information were dichotomized, such that participants received a “1” if they responded with “All the time” and a “0” otherwise. The variable “HIV care reminders digital” was coded as “1” if participants reported at least one instance of being encouraged or reminded by means of text messaging, mobile app, or social media messaging; otherwise it was coded as “0.” The variable “HIV care reminders analog” was coded as “1” if at least one encouragement or reminder was received communicating face to face or via telephone, otherwise “0.” The variables “ART medication reminders digital” and “ART medication reminders analog” were constructed analogously. Of note, the use of medication reminders was only assessed for participants who reported that they were currently taking ART medication. The measure “Engaged in HIV care” was derived by dichotomizing responses on the number of times participants received HIV clinical care such that participants were given a “1” if they received care at least once and “0” otherwise. The variable “VL undetectable” was coded as “1” if a participant reported that the result of their last VL test was “undetectable,” “0” otherwise.

Associations between sociodemographic variables and health-related uses of digital technologies were examined in separate multivariable logistic regression analyses. For the predictor variable “Income,” the response category “Don’t know/not sure” was included in the analyses in order to prevent listwise deletion of the data of ten participants who gave this response, but the parameter estimates were not reported because the contrast between the low-income reference category and the “Don’t know/not sure” category lacked a clear interpretation. Moreover, any significant findings for this contrast were potentially spurious, given the small number of participants in the “Don’t know/not sure” response category (n = 10).

In a second set of logistic regression analyses, two types of HIV care support, analog and digital care reminders, served as predictors for the three binary HIV care outcomes (“Engaged in HIV care,” “ART uptake,” “VL undetectable”). In a logistic regression analysis carried out for the subgroup of participants who were taking ART medication, the outcome variable “VL undetectable” was regressed on the two types of medication reminders, analog and digital. The ordinal outcome variable “ART adherence” was tetrachotomized (1 = Very poor/Poor, 2 = Fair, 3 = Good, 4 = Very Good/Excellent) and related to analog and digital medication reminders as predictor variables in an ordinal logistic regression analysis. Intercepts (cutpoints) are not reported because they were not used in the interpretation of the results. The significance level for all statistical tests was set to α = .05. All analyses were carried out using the R language and environment for statistical computing, version 3.5.1.

Results

Sociodemographics

Table 1 demonstrates that participants were predominantly young trans women of color (89.2%). Forty percent (40.8%) of the sample had not attained a high school diploma or equivalent. Nearly half (46.2%) reported that their income in the past month was less than $500, and 43.8% reported that they experienced housing instability in the past six months.

Table 1.

Sociodemographics (N = 130)

Characteristic	N (%)
Age
18-24	16 (12.3%)
25-29	38 (29.2%)
30-34	76 (58.5%)
Race/ethnicity
Hispanic/Latina	56 (43.1%)
African-American/Black	49 (37.7%)
Caucasian/White	14 (10.8%)
Multiracial/other	11 (8.5%)
Education
< high school	53 (40.8%)
High school/GED	45 (34.6%)
> high school	32 (24.6%)
Income (past month)
< $500	60 (46.2%)
≥ $500	60 (46.2%)
Don’t know/not sure	10 (7.7%)
Housing instability	57 (43.8%)

Digital Technology Use for Trans-specific Resources and Sexual Health Information

About half of the participants reported that they searched online “all the time” for trans-specific resources (53.8%) and for sexual health information (51.5%; Table 2). Logistic regression analyses (Table 3) showed that Hispanic/Latina participants were significantly less likely than African-American/Black participants to search online “all the time” for trans-specific resources (OR = 0.23, 95% CI [0.09, 0.58]) and for sexual health information (OR = 0.29, 95% CI [0.12, 0.73]). Participants who identified as multiracial/other also were less likely to search for sexual health information “all the time” (OR = 0.20, 95% CI [0.04, 0.99]) and trended towards a lower likelihood of searching for trans-specific resources. A higher income (≥ $500) in the past month was also associated with greater odds of searching “all the time” for trans-specific resources (OR = 2.67, 95% CI [1.13, 6.34) and for sexual health information (OR = 2.67, 95% CI [1.14, 6.26]).

Table 2.

Internet Search, HIV Care Continuum, and Technology Use for HIV Care

Characteristic	N(%) or Med(range, IQR)
Internet search for …
Trans-specific resources	70a (53.8%)
Sexual health information	67a (51.5%)
Engaged in HIV care	81 (62.3%)
Currently taking ART	63 (48.5%)
ART adherence (n = 63)	4 (1-6, 3-4)
VL status undetectable	45 (34.6%)
HIV care reminders
Analogb	70 (53.8%)
Digitalc	51 (39.2%)
ART medication reminders (n = 63)
Analogb	40 (63.5%)
Digitalc	23 (36.5%)

^a“All the time” responses

^bFace to face, telephone

^cText messaging, mobile app, social media messaging

Table 3.

Digital Technology Use for Trans-specific Resources, Sexual Health Information, and HIV Care by Sociodemographics (N = 130)

	Internet search for …		Digital reminders to …
	Trans-specific resources	Sexual health information	Seek HIV care	Take ART medication (n = 63)
Model parameters	OR 95% CI	OR 95% CI	OR 95% CI	OR 95% CI
Age
18-24	1.31 [0.38, 4.51]	2.16 [0.61, 7.64]	1.46 [0.46, 4.66]	2.79 [0.49, 15.96]
25-29	1.03 [0.41, 2.55]	1.16 [0.47, 2.87]	1.59 [0.66, 3.78]	0.61 [0.13, 2.94]
30-34	Reference category
Race/ethnicity
Afr.-Amer./Black	Reference category
Hispanic/Latina	0.23 [0.09, 0.58]**	0.29 [0.12, 0.73]**	1.26 [0.53, 3.00]	1.54 [0.33, 7.32]
Caucasian/White	0.35 [0.09, 1.40]	0.62 [0.16, 2.49]	2.02 [0.56, 7.32]	0.37 [0.03, 4.22]
Multiracial/other	0.24 [0.05, 1.07]	0.20 [0.04, 0.99]*	1.13 [0.27, 4.76]	0.17 [0.01, 2.48]
Education
< high school	Reference category
High school/GED	0.64 [0.26, 1.62]	0.68 [0.27, 1.72]	0.94 [0.38, 2.29]	1.13 [0.22, 5.81]
> high school	1.30 [0.47, 3.59]	1.54 [0.56, 4.25]	1.82 [0.70, 4.76]	5.70 [1.04, 31.19]*
Income (past month)
< $500	Reference category
≥ $500	2.67 [1.13, 6.34]*	2.67 [1.14, 6.26]*	1.97 [0.85, 4.53]	6.73 [1.52, 29.67]*
Housing instability	0.79 [0.33, 1.84]	0.73 [0.31, 1.71]	1.36 [0.59, 3.13]	1.55 [0.33, 7.24]
Intercept	1.78 [0.62, 5.08]	1.20 [0.43, 3.35]	0.24 [0.08, 0.72]*	0.11 [0.02, 0.80]*

Note. Afr.-Amer. = African-American

^*p ≤ .05.

^**p ≤ .01.

HIV Care Continuum

As shown in Table 2, nearly two-thirds (62.3%) of participants were engaged in HIV care, but just less than half (48.5%) were currently taking ART and only one-third (34.6%) reported that they had reached an undetectable VL when last tested. The median ART adherence rating was 4 (“Good”).

Digital Technology Use for HIV Care

About half of participants (53.8%) reported that they received analog reminders to seek HIV care and 39.2% reported that they received digital reminders (Table 2). Of those who reported ART uptake (n = 63; 48.5%), 63.5% received analog ART medication reminders and 36.5% received digital ART medication reminders.

Regression analytic results (Table 3) did not reveal any significant associations between sociodemographic variables and the use of digital reminders to seek HIV care. By contrast, in the subgroup of participants who were taking ART post-secondary educational attainment (OR = 5.70, 95% CI [1.04, 31.19]) and higher income (OR = 6.73, 95% CI [1.52, 29.67]) were associated with an increased likelihood of receiving digital ART medication reminders.

Table 4 summarizes findings on the association between analog and digital HIV care support and HIV Care Continuum outcomes. Analog, but not digital, care reminders were associated with an increased likelihood of being engaged in HIV care (OR = 2.37, 95% CI [1.13, 5.00]) and ART uptake (OR = 2.18, 95% CI [1.06, 4.48]), and there was a nonsignificant trend towards a greater likelihood of reporting an undetectable VL. Similarly, receiving analog, but not digital, medication reminders trended towards an increased likelihood of reporting better ART adherence. Lastly, medication reminders were unrelated to reporting an undetectable VL.

Table 4.

Associations between the HIV Care Continuum Outcomes and Communication Technologies (N = 130)

	Engaged in HIV care	ART uptake	ART adherence (n = 63)	VL undetectable	VL undetectable (on ART med.; n = 63)
Model parameters	OR 95% CI	OR 95% CI	OR 95% CI	OR 95% CI	OR 95% CI
Care reminders
analoga	2.37 [1.13, 5.00]*	2.18 [1.06, 4.48]*	—	1.94 [0.90, 4.18]	—
digitalb	1.08 [0.51, 2.29]	1.07 [0.52, 2.23]	—	1.86 [0.86, 4.01]	—
ART med. reminders
analoga	—	—	2.51 [0.95, 6.61]	—	1.79 [0.62, 5.21]
digitalb	—	—	0.54 [0.21, 1.43]	—	2.03 [0.68, 5.99]
Intercept	1.03 [0.55, 1.93]	0.60 [0.32, 1.14]	—c	0.28 [0.14, 0.58]***	0.67 [0.26, 1.75]

Note. Med. = medication

^aFace to face, telephone

^bText messaging, mobile app, social media messaging

^cIntercepts/cutpoints are not reported

p ≤ .05.

^**p ≤ .01.

^***p ≤ .001.

Discussion

This study provided insight into the use of technology to access health information and to support HIV care among a sample of young adult, socioeconomically disadvantaged trans women living with HIV. Given the evidenced high rates of sexual risk behaviors (Herbst et al., 2008; James et al., 2016; Nemoto, Operario, Keatley, Han, & Soma, 2004; Poteat et al., 2015; Reback & Fletcher, 2014), high rates of hormone misuse (Clark, Fletcher, Holloway, & Reback, 2018), and high rates of HIV infection among trans women (Centers for Disease Control and Prevention, 2018), coupled with the widespread use of technology among this population (Reback, Clark, Fletcher, Holloway, 2019), it is critical to understand if and how young adult trans women use digital technologies for health-related purposes. This knowledge can help in the development of accessible, scalable and culturally responsive technology-based interventions for this high-priority population.

This sample of young adult trans women living with HIV was comprised predominately of trans women of color; many had limited educational attainment, very low income, and experienced housing instability. While nearly two-thirds of the participants were engaged in HIV care, only about half had initiated ART uptake. Participants frequently used the Internet to search trans-specific resources and sexual health information, with about half of the participants reporting that they did so “all the time.” However, Hispanic/Latina ethnicity and lower monthly income were associated with reduced Internet usage for these purposes. Less than 40% of the participants reported that they were encouraged or reminded to seek HIV care or to take ART medication by means of digital communication technologies (text messaging, mobile app, social media messaging). Receiving digital ART medication reminders was more common among participants with post-secondary educational attainment and higher income. Finally, analog HIV care support (i.e., face to face or by phone), but not digital support, was related to superior outcomes along the HIV Care Continuum, a finding that is consistent with a previously reported association between tangible/practical social support and HIV care outcomes (Kalichman et al., 2017).

However, it would be premature to conclude that analog support had a causative effect on HIV care outcomes, given the correlational nature of these findings. Conversely, the lack of significant findings for digital care support does not warrant the conclusion that the use of digital technologies is ineffective because the present study made no attempt to distinguish between different providers of digital care support (e.g., friends versus health care providers) or to quantify the frequency of digital support beyond its presence or absence in the past six months. Nonetheless, interestingly, the results of this study suggest that in participants’ everyday lives, digital care support was less effective or persuasive than analog support. Research on mHealth interventions has identified factors associated with greater intervention efficacy. A meta-analysis by Finitsis, Pellowski, and Johnson (2014) showed that text-messaging interventions were more successful in promoting ART adherence, if messages were sent less often than daily, communication was bidirectional, message content was personalized, and participants’ ART dosing schedule was taken into account. Moreover, superior efficacy has been demonstrated for text message content grounded in evidence-based theories of behavior change (Reback, Fletcher, Shoptaw, & Mansergh, 2015).

These findings must be interpreted in light of several limitations. The study relied on convenience sampling of young adult trans women living with HIV in a large urban area. Findings may not generalize to populations with different sociodemographic profiles and in different geographic locations. Laboratory biomarker and medical chart data were not available to locate participants on the HIV Care Continuum. Participants’ self-reported data may have under- or overestimated their utilization of HIV care and HIV health outcomes. The measures of Internet search for trans-specific resources and sexual-health information were construed broadly and participants may have differed widely in the specific types of information they searched for online, given that there was no operationalized definition for either trans-specific resources other than “hormones” or sexual-health information. Moreover, similar patterns of associations with sociodemographic variables suggest that the two measures might have reflected an underlying general tendency to search for information on the Internet.

Findings from this study were somewhat counterintuitive. Although approximately half of the participants reported using technology to access trans-specific resources and sexual health information “all the time,” only analog communication was related to HIV Care Continuum outcomes. In fact, a minority of participants received HIV care support via digital communication technologies, but this type of support was unrelated to HIV care outcomes. Taken together, the study findings provided instructive information for developing technology-based interventions that are specifically tailored to the needs of this population. The common usage of digital technologies among study participants indicates a receptiveness to technology-based interventions, but the absence of an empirical link between digital, as opposed to analog, communication and HIV care outcomes offers a cautionary note, suggesting that, for this population, health care reminders might be more favorably received via traditional, i.e., analog, forms of communication.