Introduction

As mortality rates continue to decrease among cancer patients in the USA, health care systems are faced with a growing population with complex medical, psychological, and social needs [1]. This has caused comprehensive care that extends years beyond treatment initiation. However, lapses in care continue to persist and much progress is needed to improve quality among cancer survivors [2]. Mobile health (mHealth) technology has served as a tool that can potentially address these needs and encourage active participation among patients in a variety of care settings including cancer-related care [3,4,5].

The integration of mobile health (mHealth) applications in cancer care have been widely discussed and proposed roles include augmenting clinical processes, managing chemotherapy-related side effects, and supporting drug adherence [6, 7]. The current role of mHealth in oncology has been limited to self-management, and the quality of cancer-related mobile applications varies drastically [8, 9]. A recent study found that while the acceptance of mHealth technology was high among cancer patients, adoption of such technology is low. Factors predictive of the use of mHealth in this population included rural residency and a new diagnosis of cancer, indicating that geography and education are driving motivators for mHealth utilization [10].

Mobile health has also been proposed as a potential tool to reduce health disparities among underserved populations through improved health literacy, medication adherence, and clinical outcomes [11]. Low health literacy has been associated with several poor health outcomes and partially explains the racial and socioeconomic disparities in care [12, 13]. While the data regarding the efficacy of implementing mHealth initiatives are low quality and scarce, the current data that exists is promising [5, 8]. A recent meta-analysis of mHealth-based interventions among breast cancer patients found that interventions consistently improved quality-of-life measures among these patients, albeit the data on the psychological effect of these interventions is less conclusive [14]. However, another meta-analysis reported in pain, fatigue, and other psychological outcomes among cancer survivors utilizing mHealth-based interventions [15]. This suggests that the data on the efficacy of mHealth on both cancer survivorship and those undergoing treatment is still an emerging and exciting area of research.

The purpose of this study was to investigate factors associated with a willingness to utilize a mHealth application among a socioeconomically diverse and medically under-served cohort of cancer patients. This study was initially launched as a quality improvement project seeking new patient engagement platforms. We hypothesized that while the acceptability of a mHealth application for cancer-related health information will be quite high, it will be limited by education-related factors and current technology use patterns. Specifically, we believe that (1) patients without current daily internet will be less accepting of a mHealth application and (2) patients that currently lack a smartphone will be less accepting of a mHealth application than their counterparts with this technology.

Materials and methods

Study design and participants

This is a cross-sectional study of cancer patients attending the outpatient clinic and infusion center at an academic medical center in Philadelphia. Inclusion criteria consisted of (1) age greater than 18 years, (2) diagnosis of malignancy, and (3) willingness to participate. Before administration, the survey was evaluated by the institutional review board (IRB) at the host institution and was deemed IRB exempt. Data was collected over a 1-month period in 2018.

Survey development

This survey was designed by one of the authors (R.P.). The survey consisted of basic demographic information followed by a series of questions regarding access to the internet and mobile technology. Previous use of mHealth applications as well as desired use of mHealth applications was also evaluated. All questions were arranged in yes/no format for simplicity. Also, the participants were asked to answer the questions irrespective of current technology use (i.e., using a mobile application on a smartphone in participants currently without a smartphone). The complete survey can be found in Supplementary Files.

The survey was initially tested among a pilot population of 76 participants and subsequently expanded to a target of 150 participants. An error in the data collection process resulted in the loss of demographic information related to race in the expanded cohort. However, no changes or adaptations were made to the study protocol. The sample size was determined using an a priori power analysis estimating a 20% mean difference in mHealth acceptability based on both age and educational achievement with a Type I error rate of 0.05% and Type II error rate of 20%. Analysis indicated that 99 participants would be required to detect this difference.

Data collection

Eligible participants were approached and given explanations regarding the purpose and significance of the study. A trained interviewer performed structured interviews using the survey as a guide. Study data were collected and managed using the Research Electronic Data Capture (REDCap) tools hosted at the study site [16].

Statistical analysis

The data were analyzed using the Statistica® Version 13.0 (TIBCO Software Inc.) and Graphpad® PRISM 6.0 (Graphpad Software Inc.). Demographic variables were characterized by descriptive statistics. Chi-Square test for association was utilized for categorical variables. Statistical significance was defined as p < 0.05 on a two-tailed distribution. Logistic regression models were fit to identify associations between participants’ sociodemographic characteristics and their willingness to use mobile health applications for monitoring their health condition. Adjusted odds ratio reported as (OR mean, CI95%).

Results

A total of 151 participants were approached and all participants consented to the survey (100% response rate). The demographic information of participants is summarized in Table 1. The mean age of participants was 62.3 ± 14.4 years of age and consisted of 35.8% male. Race demographics were not collected after the pilot population was expanded, but the pilot population consisted primarily of minority groups (83.1%). While 60.3% of participants had a high school degree or less, 39.7% obtained at least an undergraduate degree with 5.9% of total participants achieved additional education (Table 1).

Table 1 Demographic information of participants (N = 151)
Full size table

The response distribution of selected survey items stratified by age is outlined in Table 2. Only 73.5% of survey respondents have daily access to internet services with significant differences in age groups. Individuals without internet tended to be older (mean age of 71.9 ± 11.2 vs. 58.8 ± 13.9) and less educated (12.5% with a college degree vs. 49.5%) than their counterparts with internet. Similar characteristics were seen in individuals that lacked either a mobile phone or smartphone. Among the 106 individuals that currently own a smartphone, 60.4% use phone that operates iOS software (iPhone®), and 39.6% use a phone that operates Android software (Samsung®, LG®, etc.). The frequency of previous, defined as over the past 6 months, mobile access to general health and personal health information was similar (53% and 51.7%, respectively). However, 59 participants (39.1% of total) neither accessed general nor personal health information, while 66 participants (43.7% of total) utilized both forms of health technology (Table 2).

Table 2 Comparison of positive responses to selected survey items stratified by age
Full size table

Future-oriented survey items (items 12 through 16) evaluated different aspects of integrating mHealth technology into practice and were met with modest support overall. However, acceptance was significantly different based on current technology utilization. Among individuals without access to the internet, the mean acceptance rate of future-oriented items was 7.5%. This pattern was also demonstrated in individuals lacking a mobile phone (4.2%) and smartphone (2.3%) (data not shown).

Regarding demographic factors influencing the willingness to utilize a mobile application for daily health monitoring, acceptance was significantly associated with both age and educational achievement (Table 3). Willingness to utilize a mHealth application was highest in individuals < 50 years of age (83.3% favorable) and lowest in individuals > 80 years of age (82.4% unfavorable). Logistic regression revealed that age groups 61–70 (OR 0.24, CI95% 0.07–0.90), 71–80 (OR 0.05, CI95% 0.01–0.23), and > 80 years (OR 0.04, CI95% 0.01–0.22) were significantly less likely to utilize a daily mHealth application than individuals < 50 years (Fig. 1A). 71.7% of individuals with at least a college degree were favorable of a mHealth application compared to 41.8% of individuals without a college education (OR 2.78, CI95% 1.25–5.88) (Table 3 & Fig. 1A). No significant differences in race, gender, or marital status were identified. However, differences among age groups and education were eliminated when adjusting for current smartphone use (Fig. 1B).

Table 3 Comparison of demographic information to the willingness to use an application for daily health monitoring (Survey Item #16)
Full size table
Fig. 1
figure 1

Factors associated with willingness to utilize a daily mobile application for health monitoring among all participants. *Reference group for education, gender, and age was high school or less, male, and < 50 years, respectively

Full size image

A desire to increase cancer-related knowledge was associated with an increased odds of utilizing a daily mHealth application (OR 261.5, 10.13–6748.71, p < 0.01). Interestingly, no other features of a mobile application exhibited signficant associations despite being similar in nature (Fig. 2B). Evaluation of the frequency of each response revealed that of the 81 participants that were willing to use a daily mobile application, and 80 (98.8%) participants also desired an application to increase cancer-related knowledge. Conversely, of the 70 individuals not willing to utilize a daily application, 63 (90.0%) also did not desire an application to increase their cancer-related knowledge (data not shown).

Fig. 2
figure 2

Willingness to utilize a daily mHealth application among smartphone users based on previous and desired mHealth formats reported as odds ratios. *Reference value for each activity was the negative response (no)

Full size image

Discussion

This study aimed to investigate the factors associated with acceptability of a mHealth application among a population of cancer patients in a socioeconomically diverse and medically underserved community. Our study revealed that the acceptability of mHealth applications is mediated by age and educational-related factors, as well as access to internet. This suggests that past technology use is highly indicative of adoption of future technology such as a daily mobile application for cancer-related health.

Several studies have identified gaps in internet access along sociodemographic boundaries such that individuals with lower income, lower education, and identify as a racial minority are less likely to have access to internet services [17, 18]. These findings, coupled with the abundant literature linking the prevalence of inadequate health literacy and poor clinical outcomes among African-Americans, further highlight the need for mHealth applications among minority populations in the USA [11, 12, 19].

Our study both confirms and expands upon the previous literature. Previous literature has reported both high acceptability and positive outcome measures among cancer patients utilizing mHealth-based interventions [9, 10, 14, 20]. However, our study has investigated both the acceptability and feasibility of mHealth-based interventions among a population of predominately racial minorities and of lower socioeconomic status. While our study reported an overall acceptance of a mHealth application for health information among cancer patients, we found that our population faces technology-related barriers at a higher prevalence than the rest of the USA. The rate of daily internet access among our patient (73.5%) is much lower than a recent Pew Research estimate of 90% [21]. Our population also had a much lower rate of both a mobile phone (84.1%) and smartphone ownership (68.2%) than a recent national estimate of 96% and 81%, respectively [22]. This highlights a concern regarding implementation of smartphone-based patient engagement platforms because a large percentage of patients would be excluded from participation. Considering the existence of many cheap smartphones within the USA, future studies should investigate both monetary and non-monetary factors related to smartphone adoption among low-income communities.

We found that the desire to increase cancer knowledge was significantly associated with an increased likelihood of utilizing a daily mHealth application for cancer-related information among current smartphone users. Interestingly, we found that past mHealth-related use patterns (accessing personal or general health information on a smartphone) was not associated with the willingness to utilize a daily mHealth application. This suggests that although current smartphone users may not have accessed personal or general health information in the past, future mHealth applications could gain acceptance among patients. This offers an exciting avenue for future patient engagement initiatives.

There are several limitations in this study. This study investigated only a small subset of patients within a population of predominately low socioeconomic status, so the results are likely not generalizable to the overall population. While a post hoc power analysis indicated that our study was adequately powered to detect the observed differences, there is always the possibility of detecting a difference that is not present in the overall population given our sample size. Our study also only evaluated the willingness of a theoretical mobile application, therefore further research is required to develop and test a mobile application within this population. Future studies should investigate whether or not access to technology facilitates health-promotion activities among a predominate minority and socioeconomically diverse population.

Conclusions

The potential use of mHealth-based interventions in a socioeconomically diverse population offers several barriers and opportunities. The relatively lower rates of daily internet access and smartphone use in a low-income patient population appear to be the largest barrier to mHealth acceptance, while education and age-related barriers are eliminated when adjusting for current smartphone use. The desire to increase cancer knowledge was significantly associated with an increased likelihood to utilize a daily mHealth application and offers an opportunity to integrate patient education with future mHealth-based interventions.