This study addressed the question of who would be denied access to mHealth interventions if they were only available to patients who possessed their own smartphone. We found that one-third of patients being treated for TB by health departments in three USA metropolitan cities did not own a smartphone prior to the study. Older age, male sex, lower income, and lower education level were associated with higher odds of not owning a smartphone. Importantly, however, we also found that smartphone ownership was unassociated with participants’ ability to learn or use the VDOT application.
Our finding that there is less ownership among older patients is consistent with other studies [13]. This is likely due to a combination of factors that includes the increased availability of smartphones in the last decade and the increased earning potential of younger participants that would allow them to purchase a smartphone [14]. We also found that men were more likely to not own a smartphone than women. While this was a surprising finding based on prior studies showing that women are less likely to own a smartphone than men, similar findings were observed in other studies [4,5,6, 15,16,17]. Although this finding was unexpected, it is reassuring to see that women, who historically had less access to technology, appear to have greater smartphone ownership in this cohort. Furthermore, we did not observe a significant difference in education level or income between male and female participants. Our results, while unexpected, are consistent with current trends showing that the gender gap in smartphone use is reversing in both high and low resource settings [16]. This could be due in part to the steadily decreasing wage-gender gap in the USA [18]. Women were found to make 80–90% per dollar compared to their male counterparts, up from 70% in the 1990s and early 2000s [18]. This rise is occurring in conjunction with the increasing labor market value of women and the increase in full-time workers [18]. Together these findings support the notion that more women would have a greater use for smartphones as well as the means to purchase them.
Not surprisingly, lower educational attainment and income were independently associated with non-smartphone ownership, as other studies have shown that smartphone ownership is more prominent among individuals with higher socioeconomic status [7, 19, 20]. When we analyzed education level by annual income (<$10,000 vs. ≥$10,000), we found that only 42.7% of participants with a high school education or below earned more than $10,000 a year compared to 60.3% of participants who completed more than a high school education. These results are consistent with other studies, including one that analyzed cellphone ownership among persons who inject drugs (PWID) in Tijuana, Mexico. The authors acknowledge that marginalized individuals were less likely to have the resources necessary to own and maintain a cellphone [21, 22]. While PWID cannot be directly compared to patients with TB, it is reasonable to expect that members of both groups have overlapping sociodemographic characteristics given that TB is often associated with poverty in high-burden settings [23].
Although most participants perceived their care to be good, we observed that participants who did not own a smartphone were more likely to report that in-person DOT made them feel “cared for” and that the amount of contact with healthcare workers during VDOT was “not enough”. There are multiple possible explanations for these findings. The observed associations could be because smartphone ownership is a proxy for having access to healthcare and patients with less access to healthcare (i.e., non-smartphone owners) might value the time spent with a healthcare provider during DOT visits more than patients who are accustomed to having access to their doctor (i.e., smartphone owners). Consistent with previous studies, those without smartphones were more likely to be older and belong to a low-income group, which could also contribute to greater social isolation [20, 24]. The preference for in-person DOT and more contact with treatment staff could therefore be valued for its effect on reducing isolation [20, 24]. Additionally, older or less educated patients might associate the in-person component of DOT with more attention from their provider and better care.
Despite these findings, differences in perceptions between the two groups did not affect participants’ ability to use VDOT. Notably, both groups predominantly found VDOT very easy to use and required less than two practice days to learn how to use the application. Moving forward, this should reassure designers of mHealth applications that baseline smartphone ownership is not necessary for successful use of the device with appropriate training. These results are important for designing future mHealth interventions in general and for understanding patient preferences in the methods used to monitor TB treatment adherence.
Some limitations should be considered when interpreting these findings. Since all study participants had used DOT before taking part in the study, it is possible that individuals who were willing to switch to VDOT were more likely to own a smartphone and more inclined to view VDOT as a positive intervention. However, only five patients in San Diego and three patients in San Francisco refused to participate, making selection bias and overestimation of smartphone ownership unlikely. It should also be noted that this study was limited to USA cities. Most of the global burden of TB is in other more resource-limited settings that may also have different levels of cellular phone and internet infrastructure as well as cultural and societal factors. Therefore, the results of this study might not be generalizable to other parts of the world. The study’s sample size potentially limited our ability to identify additional factors associated with smartphone ownership. Familiarity and comfort with using features on a smartphone could have varied among participants who reported owning a smartphone. For example, some participants might have only used it as a telephone, while others used it for calling, texting, emailing, Internet browsing, and engaging with other applications. The dependent variable (i.e., smartphone ownership) did not specify how participants used their smartphones. Finally, all study participants were TB patients receiving treatment through large urban public health departments, which might not generalize to other patient populations.
mHealth interventions requiring patients to use their own smartphone in the USA have the potential to disproportionately exclude patients who are male, older, and less educated. It is important to recognize that changing economic and societal norms have made access to smartphones more prevalent across the general population, thereby reducing the potential for the observed disparities. However, until these inequities are eliminated, there is still a concern that certain groups of patients could be systematically excluded if personal smartphone ownership is required. While recent studies in other parts of the world have suggested that mobile phone usage is prevalent enough to justify the use of mHealth interventions, it is not yet clear if or how outcomes will differ between VDOT and DOT, or whether providing smartphones to low-income patients will affect the impact of VDOT [8]. Similar studies conducted in rural areas and with other patient populations are also needed to determine whether these results are generalizable to the broader population. Healthcare providers proposing adoption of mHealth interventions should consider the availability of smartphones among the populations they serve and include contingencies to accommodate patients who lack smartphones to avoid creating or perpetuating health disparities.
