Data used in this study were from the HINTS 5-Cycle 3 [28]. HINTS is a nationally representative survey designed to understand American adults’ knowledge of, attitudes toward, and use of cancer- and health-related information [29]. HINTS 5-Cycle 3 used a single-mode mail survey, with a 2-stage sample design, including a stratified sample of addresses and a selected adult within each sampled household [28]. The data were collected from 5438 respondents from January to May 2019 (English version only), with an overall 30.3% response rate [28]. Comprehensive reports on the sampling design for the HINTS survey have been published elsewhere [28-30]. The survey data were deidentified and are publicly available; institutional review board approval was not applicable.

All analyses considered the complex survey design of the HINTS 5-Cycle 3 sample by using the HINTS-supplied final weights to estimate population estimates and 50 replicate weights to compute the standard errors with the jackknife replication approach [29]. Specifically, descriptive statistics were used to describe the prevalence and the characteristics of e-communication users and nonusers. Multiple logistic regression analyses were used to assess the association of sociodemographics and comorbidities (Model 1) and sociodemographics, comorbidities, plus patient-clinician relationship factors (Model 2) with e-communication use. Multiple linear regression analyses were used to examine the association between e-communication use and patient-perceived quality of care with the control of sociodemographics and comorbidities (Model 3) and the control of sociodemographics, comorbidities, plus patient-clinician relationship factors (Model 4). Missing data pattern analysis indicated that most variables had missing data <5% (see Table S1 in Multimedia Appendix 1). Multiple imputation was performed, and the pooled results of model 3 and model 4 were based on 50 imputed data sets using multiple imputation by chained equations. All analyses were conducted using Stata software (version 14; StataCorp). Results were reported as weighted point estimates and 95% CIs. The level of significance was .05.

The overall prevalence of the use of e-communication was 60.3%. Most American adults who used e-communication with clinicians in the past 12 months were younger than 65 years, as older adults only accounted for 16.6% of e-communication users but 25.7% of nonusers (see Table 2). Table 2 also displays that most e-communication users were females (53.9%), had at least some college (41.7%), and 36.4% college graduates or more, were White people (65%), currently married (59.9%), with a household income ≥US $50,000 (63.9%), and did not live alone (85%) or in rural areas (89.5%). e-Communication users and nonusers were significantly different in all person-related characteristics. In addition, significantly more e-communication users had a regular health care provider than e-communication nonusers (72.9% vs 51.4%, respectively; P<.001) (see Table 2). All percentages are weighted; therefore, absolute values are not shown.

Table 3 presents the results of multiple logistic regression analyses on the sociodemographics, comorbidities, and patient-clinician relationship factors for e-communication use. In model 1, where only sociodemographic factors and comorbidities were considered, age (odds ratio [OR] 0.87, 95% CI 0.66-1.14), female (OR 1.44, 95% CI 1.17-1.77), education level (eg, for college graduates or more, OR 4.78, 95% CI 2.63-8.68), household income (OR 1.77, 95% CI 1.34-2.34), rural residency (OR 0.62, 95% CI 0.44-0.87), and number of comorbidities (OR 1.33, 95% CI 1.16-1.52) were associated with e-communication use (see Table 3). In model 2, after adding the relationship factors to the model, people who were females (OR 1.47, 95% CI 1.18-1.82), high school graduates (OR 1.95, 95% CI 1.14-3.34), having some college (OR 3.34, 95% CI 1.84-6.05), and college graduates or more (OR 4.89, 95% CI 2.67-8.95), with a household income at or greater than US $50,000 (OR 1.63, 95% CI 1.23-2.16), with more comorbidities (OR 1.22, 95% CI 1.07-1.40), or having a regular health care provider (OR 2.62, 95% CI 1.98-3.47) were more likely to use e-communication, whereas those who were older adults (OR 0.42, 95% CI 0.31-0.57) or rural residents (OR 0.61, 95% CI 0.43-0.88) were less likely to use e-communication.

Table 4 displays the results of the association between e-communication use and patient-perceived quality of care among American adults. After controlling for sociodemographic factors (age, gender, education, income), comorbidities, and patient-clinician relationship factors (having a regular health care provider, trust a doctor), the use of e-communication was statistically associated with better quality of care (β=.12, 95% CI 0.02-0.22; see Model 4 in Table 4).

This study examined the prevalence of and factors associated with e-communication use and the potential association between e-communication use and patient-perceived quality of care in a nationally representative sample of American adults. To the best of our knowledge, this study is the first to explore the association of e-communication use with patient-perceived quality of care at the population level. Several important findings emerged in this study.

First, the majority of American adults (60.3%) used some forms of e-communication with clinicians throughout 2019, which was significantly higher than the reported 7% in 2003, 10% in 2005 [9], and 31.5% in 2014 from the previous HINTS [31]. This finding indicates that e-communication use has become increasingly popular for adults to interact with their clinicians. The increased prevalence rate can be attributed to the increased availability and popularity of electronic health devices [32,33] and supportive policies (eg, promoting patient access to their electronic medical records) [34]. Although our data showed an overall growing trend in the use of e-communication, it is important to note that older adults’ use of e-communication still remained relatively low, and this rate was not much improved from that in 2003 and 2005 [12]. Literature indicates that older adults usually prefer direct in-person interactions with their clinicians [12], while there are increasing reports about older adults’ positive attitude toward e-communication and their preference for email and messaging communication with clinicians that is similar to that for younger adults [35,36]. Our finding suggests that there is still a gap in the actual use of e-communication between older adults and young adults [31,37,38]. More studies are needed to explore the practical challenges that older adults may encounter in the use of e-communication. Older adults are potentially the major users of e-communication, considering their high level of health care needs. It is important to develop appropriate e-communication support for this population for their better health outcomes.

In addition to age, we also found that the use of e-communication varied by gender, education, income, and residency, indicating that individuals who are females, with higher education, higher income, and more comorbidities, or who reside in nonrural areas were more like to use e-communication with their clinicians. This finding is congruent with reports of the general adoption of eHealth in literature [39-42]. Consistent with our finding, the positive association between education and e-communication usage was reported in previous studies [31,39], which can be interpreted as individuals who have higher education might have more eHealth literacy skills and technological capabilities [43] to help them better use electronic forms of information [31,39]. However, Senft and Everson’s recent study [44] reported that individuals who had lower levels of education and had negative care coordination experiences are more likely to use eHealth activities to communicate with clinicians [44], indicating that personal health care experiences can possibly interplay with education and thus influence the use of e-communication. However, it is unclear whether the limited use of e-communication among rural residents is related to lack of internet connectivity or awareness of e-communication services [45]. Additional studies can be conducted for further exploration.

Compared to those who did not use e-communication in the past year, in this study, e-communication users were more likely to have a regular health care provider and reported better trust in information from a doctor. However, trusting a doctor was not an independent predictor of e-communication use when having a regular health care provider was controlled for in the model. A previous qualitative study has indicated that a trusting relationship between patient-clinician is a significant contributor to better online patient-clinician interactions [20,46]. Even those who tend to frequently seek web-based health information are more willing to use the information provided by their trusted clinicians for their health decision-making [17,47]. Our findings suggested that patients with a regular health care provider had the greatest association with their use of e-communication. It is possible that patients who have a regular health care provider have already built a trusting relationship with their clinicians. Given the importance of trust in a provider in the patient-centered care process, future research directly examining possible confounding of this factor using longitudinal data is recommended.

Finally, it is not surprising that this study found that the use of e-communication was an independent predictor of patient-perceived quality of care. In 2001, the Institute of Medicine suggested that e-communication could improve the quality of care [48]. The previous literature review demonstrates that e-communication provides a convenient way of patient-clinician interaction, has a positive impact on patient satisfaction while saving time for patients and clinicians, and has the potential to extend health care efficiency [21,49]. The benefits and challenges of e-communication have been well addressed in the literature, while its benefits for the quality of care may not have been clearly quantified previously. The measures of quality of care can vary by the dimensions of care and care processes [50]. However, this study focused on the measure of the patient-perceived quality of care, which solely reflected patients’ perceptions of health care services received based on their experiences of care [24]. It did not mean to measure any technical clinical quality, for example, cholesterol screening [51]. There is increasing interest in patient-reported measures, as experiences with care are more easily understood by patients. In addition, previous literature demonstrated that the measure of patient experiences of care was related to measures of the technical quality of care, which can serve as valid summary measures of hospital quality [52]. These study findings were based on the analysis of nationally representative survey data, which should be generalizable to all American adults. The positive association between the use of e-communication and perceived quality of care confirms that e-communication can serve as an important tool to improve patient satisfaction and their perceptions of quality of care. This finding is particularly significant and applicable in the current COVID-19 pandemic when traditional in-person communication is less feasible. It is expected that e-communication will continuously replace an adequate portion of traditional face-to-face encounters and has the potential to transform the health care system [21]. Future research can be conducted to explore the sustainable long-term effects of e-communication on patient-centered care outcomes.

This study has a few limitations. First, data were mainly based on self-reports, which might have introduced recall bias. Second, the survey questions regarding the use of e-communication did not specify the frequency of use; therefore, they did not accurately reflect responders’ experiences of using e-communication and might affect their perceptions of quality of care. Third, a binary measure of e-communication use (yes/no) was used, which might result in the loss of information or power. However, considering the conceptual overlaps across 6 questions about e-communication behaviors in the survey, a combined continuous assessment for the number of e-communication behaviors would be conceptually inaccurate. Fourth, the e-communication was between patients and clinicians. However, the survey only focused on the patient side and thus, it was not possible to know clinicians’ perceptions of e-communication use. Finally, the results could be underestimated by potential reverse causality owing to the nature of the study design. The prevalence of e-communication use was higher in our study than that that reported in previous studies. The difference may also be due to varying measurement methods across studies. In our study, we used 6 questions to measure e-communication, which are more than that used in other studies. Different measures might affect comparisons of the prevalence of e-communication use across studies.

American adults’ use of e-communication with clinicians has been significantly increased in the past decade, which may be due to increased patient needs and advanced support from technologies and policies. As a convenient way of patient-clinician interaction, the use of e-communication is significantly associated with patient-perceived quality of care. The findings of multiple factors associated with e-communication use and the positive association between e-communication use and patient-perceived quality of care suggest that policy-level attention is needed to engage the socially disadvantaged (ie, those with lower levels of education and income, without a regular health care provider, and living in rural areas) to maximize the use of e-communication and to support better patient-perceived quality of care among American adults.