Rural Access and Usage of Patient Portals: A 2019 Health Information National Trends Survey Analysis

Abstract

To examine differences in rural and urban respondents' use of and access to patient portals in the United States, this study used the 2019 National Cancer Institute's Health Information National Trends Survey (HINTS) 5, Cycle 3. A cross-sectional secondary data analysis utilizing jackknife weighting procedures was used to generalize the findings to be nationally representative. Despite similar rates of providers maintaining an electronic medical record system, adjusted analyses found that rural respondents had lower odds of being offered access to a patient portal by their healthcare provider (OR: 0.60; 95 percent CI: 0.39-0.91) and accessing their patient portals in the last 12 months (OR: 0.62; 95 percent CI: 0.43-0.91) when compared to their urban counterparts. Additional research is needed to determine effective strategies for overcoming geographic and structural barriers to adoption of this technology by rural residents.

Introduction

Healthcare continues to experience a need to meet the ambitious systematic approach of the Institute for Healthcare Improvement's (IHI) Triple Aim initiative: to improve the health of populations, reduce per capita cost of healthcare, and improve the patient experience.¹ In an effort to meet the needs of improving the patient experience, collaboration between healthcare organizations, physicians, and the engagement of patients is fundamental.² While initiatives foster collaboration at professional levels, patient engagement tends to be a more challenging aspect of individualized, family, and population healthcare. The use of electronic medical records (EMR), personal communication technologies, and patient portals provide a means to increase patient engagement and thus improve quality care and patient satisfaction.

Patients are showing an increased need to access their health information since the “Big Push” for EMR systems in 2009.³ A patient portal, the primary method of accessing one's own health information, is “a secure online website that gives patients convenient, 24-hour access to personal health information from anywhere with an internet connection.”⁴ Patient portal use is primarily used to securely message physicians, schedule appointments, request prescription refills, and review lab results and educational materials.⁵, ⁶ The benefits of patient portal use have shown to increase patient satisfaction,⁷, ⁸ disease awareness,⁹ communication with providers,¹⁰ and support medication management and adherence.^11-14 Studies on the association of patient portal use and improved patient outcomes, however, have mixed results.¹⁵, ¹⁶, ¹⁷, ¹⁸

Despite these benefits and an increase in the use of other technologies in healthcare, access and use of patient portals remains low in the United States, increasing from 25.6 percent in 2014 to 31.4 percent in 2018.^19-21 Barriers to patient portal adoption include lack of provider and patient buy-in, lack of awareness of portal functions, and negative patient experiences using portals.²² Additional reasons for not using patient portals include patients' desire to speak directly to healthcare providers and privacy and security concerns.²³, ²⁴

The lack of patient portal use is particularly concerning for residents of rural communities, who generally experience reduced access to healthcare at all levels, as well as a greater travel burden to access care.²⁵ This population stands to benefit considerably from increased utilization of patient portals, as they provide a salient means of direct engagement between patients and providers. Despite recent studies finding healthcare providers in rural areas maintaining EMRs at similar rates to urban areas, rural patients are less likely to access online medical records, email providers, and manage personal health information online.²⁶, ²⁷ This study examines disparities in both: being offered access to and utilization of patient portals, as well as patient-reported explanation of utilization or non-utilization, by rurality.

Methods

Data for this analysis was drawn from the National Cancer Institute's 2019 Health Information National Trends Survey (HINTS) 5, Cycle 3. The annual survey provides a nationally representative sample of civilian, non-institutionalized adults aged 18 and older and monitors changes in health communication, with a specific interest in use of cancer-related information. Data collection for this particular cycle of the survey took place from January through May 2019 via a mailed questionnaires and web pilot and collected a total of 5,438 responses. Specific details on the sampling and data collected protocols can be found in the dataset's methodology report.²¹ Respondents that reported their provider did not maintain an electronic medical record were excluded from the analysis (n = 1,053).

This analysis examined two dependent variables: whether the respondent was offered access to the patient portal (“Have you ever been offered online access to your medical records by your healthcare provider?” dichotomized as yes versus no/don't know) and whether the respondent accessed the patient portal within the last 12 months (“How many times did you access your online medical record in the last 12 months?” dichotomized as yes [1 or more times] versus no [0 times]). The term “online medical record” in the survey was appropriately interpreted as a patient portal in this analysis. Using the United States Department of Agriculture's (USDA) 2013 Rural-Urban Continuum Codes, respondents were categorized as urban (1-3) and rural (4-9) as the independent variable. The control variables included gender (male, female), age (18-34, 35-49, 50-64, 65-74, 75-plus years), race and/or ethnicity (non-Hispanic white, non-Hispanic Black, Hispanic, non-Hispanic Asian, non-Hispanic other), education (less than high school, high school graduate, some college, college graduate or more), health insurance (yes/no), regular healthcare provider (yes/no), and use of internet (yes/no).

Using SAS 9.4, this cross-sectional analysis utilized jackknife weighting procedures for complex survey design to generalize the findings to be nationally representative. Additional information about the weighting procedures can be found in the HINTS methodology report.²⁸ Descriptive statistics were performed to describe the sample characteristics. In bivariate analysis, chi-square tests at the 95 percent confidence level (α = 0.05) were conducted to examine rural or urban differences in being offered access to a patient portal and accessing the patient portal. An adjusted logistic regression was performed to examine potential predictors of the dependent variables, controlling for respondent demographics.

Results

There were no differences in providers maintaining electronic medical records (79.1 percent urban versus 76.5 percent rural; p = 0.4607; analysis not shown). The remaining analyses included only the providers that were maintaining an electronic medical record (n = 4,293). Of the 4,293 respondents, 3,826 (89.1 percent) were urban residents and 467 (10.8 percent) were rural residents. Respondents were predominately female, non-Hispanic (NH) white, 50-64 years old with some college education, health insurance, a regular healthcare provider, and access to internet (Table 1).

Table 1.

Respondent Demographics by Rural/Urban Designation for 2019 Health Information National Trends Survey

Characteristic*	Urban (n = 3,826)	Rural (n = 467)	Total (n = 4,293)	p-value
Gender				0.9056
Male	47.0 (1533)	47.5 (196)	47.0 (1729)
Female	53.0 (2221)	52.5 (265)	53.0 (2486)
Age				0.1787
18-34	24.2 (477)	22.4 (42)	24.0 (519)
35-49	23.4 (674)	25.6 (79)	23.7 (753)
50-64	31.9 (1190)	26.6 (133)	31.2 (1323)
65-74	12.3 (864)	13.6 (119)	12.5 (983)
75-plus	8.23 (549)	11.7 (82)	8.7 (631)
Race/ethnicity				<0.0001
NH White	63.2 (2192)	84.5 (339)	65.9 (2531)
NH Black	11.2 (479)	3.5 (33)	10.2 (512)
Hispanic	16.4 (507)	7.4 (29)	15.3 (536)
NH Asian	6.3 (176)	-	5.7 (182)
NH Other	3.0 (111)	2.9 (10)	3.0 (121)
Education				0.0017
Less than high school	5.4 (174)	5.3 (25)	5.4 (199)
High school graduate	19.9 (557)	23.8 (106)	20.4 (663)
Some college	41.5 (1142)	49.5 (147)	42.5 (1289)
College graduate or more	33.2 (1881)	21.4 (175)	31.7 (2056)
Health Insurance				0.3446
Yes	94.7 (3664)	96.4 (450)	95.0 (4114)
No	5.3 (114)	3.4 (11)	5.0 (125)
Regular Provider				0.7125
Yes	71.2 (2846)	69.6 (364)	71.0 (3210)
No	28.8 (919)	30.4 (96)	29.0 (1015)
Use Internet				0.0737
Yes	87.1 (3184)	82.2 (349)	86.5 (3533)
No	12.9 (612)	17.8 (114)	12.5 (726)

Column percentages are weighted to reflect the U.S. population.

All differences were assessed using Wald chi-square tests.

- = Sample size below 10.

^*Respondents were missing in each category: gender (n = 78), age (n = 84), race and/or ethnicity (n = 411), education (n = 86), health insurance (n = 54), regular provider (n = 68), use of internet (n = 34).

Chi-squared tests found significant differences in race and/or ethnicity and education by rurality. A larger percentage of rural respondents were NH white (84.5 percent versus 63.2 percent), while a smaller percentage were Hispanic (7.4 percent versus 16.4 percent) when compared to urban respondents. Fewer rural respondents were college graduates (21.4 percent versus 33.2 percent; p = 0.0017). In bivariate analysis, there were significant differences in whether providers offered access to and whether the respondent accessed a patient portal (Table 2). Rural respondents were less likely to report being offered access to a patient portal (56.9 percent versus 66.9 percent; p = 0.0397) and less likely to access a patient portal (35.9 percent versus 48.4 percent; p = 0.0046).

Table 2.

Patient Portal Access Among Those With Providers That Maintain Electronic Medical Record By Rurality (n = 4,293)

	Urban (n = 3,826)	Rural (n = 467)	Total (n = 4,293)	p-value
Accessed patient portal*				0.0046
Yes	48.4 (1831)	35.9 (172)	46.8 (2003)
No	51.6 (1955)	64.1 (289)	53.2 (2244)
Offered access to patient portal*				0.0397
Yes	66.9 (2586)	56.9 (275)	65.6 (2861)
No or Don't Know	33.1 (1217)	43.1 (190)	34.4 (1407)

Column percentages are weighted to reflect the U.S. population.

^*Respondents were missing in each category: accessed patient portal (n = 46), offered access to patient portal by provider (n = 25).

No differences in primary use of patient portals or reasons for not accessing patient portals were found among rural respondents when compared to their urban counterparts (Table 3). The majority of patient portal users found it easy to understand (90.6 percent) and useful (86.7 percent). The primary use of patient portals was to view test results (86.6 percent), securely message a healthcare provider and staff (52.7 percent), and request a refill of medication (46.7 percent). The primary reasons for not accessing a patient portal were that respondents preferred to speak to a healthcare provider directly (73 percent) and they did not perceive a need to use the patient portal (59.1 percent). Again, no differences in reasons for lack of accessing a patient portal were found between rural and urban respondents.

Table 3.

Reasons for Accessing or Not Accessing Patient Portal Among Those With Providers That Maintain an EMR By Rurality

Use of patient portals	Urban (n = 1,831)	Rural (n = 172)	Total (n = 2,003)	p-value
Used to request a refill of medication	47.3 (846)	43.9 (66)	46.9 (912)	0.6057
Used to view test results	87.2 (1562)	81.1 (144)	86.6 (1706)	0.3431
Used to request correction of inaccurate information	9.7 (154)	12.0 (15)	9.9 (169)	0.5986
Used to securely message provider and staff	53.4 (992)	46.4 (77)	52.7 (1069)	0.3363
Used to download health information	28.1 (468)	18.5 (32)	27.1 (500)	0.1138
Used to add health information	25.6 (466)	18.6 (26)	25.0 (492)	0.2925
Used to help make a decision about how to treat an illness or condition	25.1 (485)	15.7 (35)	24.2 (520)	0.0533
Electronically sent medical information to another healthcare provider	14.6 (237)	17.5 (21)	14.9 (258)	0.5108
Electronically sent medical information to a family member	7.1 (108)	-	6.7 (116)	0.0530
Electronically sent medical information to a service or app to help manage and store health information	4.6 (83)	-	4.7 (91)	0.8537
Ease of understanding patient portal1	91.0 (1632)	87.0 (150)	90.6 (1782)	0.5029
Usefulness of patient portal2	86.3 (1566)	90.2 (147)	86.7 (1713)	0.2079

Reasons to not access patient portal	Urban (n = 2,586)	Rural (n = 275)	Total (n = 2,244)	p-value
Prefer to speak to healthcare provider directly	72.3 (1323)	76.7 (192)	73.0 (1515)	0.3614
Does not have a way to access the website	21.8 (404)	21.4 (63)	21.7 (467)	0.9217
Does not have a need to use patient portal	58.8 (958)	60.7 (140)	59.1 (1098)	0.7397
Concerned about privacy or security of website	22.5 (428)	19.0 (56)	22.0 (484)	0.3399
Does not have a patient portal	24.0 (367)	19.4 (55)	23.3 (422)	0.2483
Found it difficult to log in	21.2 (406)	18.2 (44)	20.7 (450)	0.4786
Not comfortable or experienced with computers	23.6 (480)	20.4 (72)	23.1 (552)	0.4332

Column percentages are weighted to reflect the U.S. population.

- = Sample size below 10.

* Respondents were missing in each category: Refill medications (n = 33); View results (n = 30);

Request correction (n = 52); Message providers (n = 30); Download health records (n = 30); Add health information (n = 29); Make decision (n = 34); Sent to other healthcare provider (n = 28); Sent to family member (n = 33); Sent to health app (n = 35); Ease of understanding (n = 36) Usefulness (n = 26); Prefer to speak with provider directly (n = 224); No way to access website (n = 259); No need to use (n = 260); Concern with privacy or security (n = 257); Does not have patient portal (n = 306); Difficult to log in (n = 286); Not comfortable with computers (n = 1984).

¹“Easy to understand” is a combination of “very easy” and “somewhat easy,” compared to “difficult” which is a combination of “somewhat difficult” and “very difficult.”

²“Useful” is a combination of “very useful” and “somewhat useful,” compared to “not useful” which is a combination of not “very useful,” “not at all useful,” and “I do not use my patient portal to monitor my health.”

Adjusted logistic regression (Table 4) estimated differences in accessing a patient portal in the last 12 months (Model 1) and being offered access to the patient portal by a healthcare provider (Model 2). In Model 1, identifying as female (OR: 1.52, 95 percent CI: 1.18-1.97) and having an education of high school graduate or greater was associated with greater odds of having accessed a patient portal in the last 12 months. While living in a rural area (OR: 0.62; 95 percent CI: 0.43-0.91), identifying as NH Black (OR: 0.56; 95 percent CI: 0.36-0.89), not having a regular healthcare provider (OR: 0.49; 95 percent CI: 0.35-0.70), and having no internet use (OR: 0.28; 95 percent CI: 0.15-0.51) were associated with lower odds of having accessed a patient portal in the last 12 months.

Table 4.

Adjusted Multiple Logistic Regression Model Predicting Access and Offer to Access Patient Portals Where Provider Does Maintain EMR (OR; 95% CI)

Characteristic	Model 1 Accessed patient portal in last 12 months (n = 3,687)	Model 2 Offered access to patient portal by provider (n = 3,695)
Rurality (Ref: Urban)
Rural	0.62 (0.43-0.91)	0.60 (0.39-0.91)
Gender (Ref: Male)
Female	1.52 (1.18-1.97)	1.83 (1.39-2.40)
Age (Ref: 18-34)
35-49	1.17 (0.79-1.73)	1.28 (0.84-1.97)
50-64	1.24 (0.83-1.87)	1.31 (0.86-2.02)
65-74	1.11 (0.74-1.68)	1.27 (0.78-2.07)
75-plus	0.72 (0.44-1.16)	1.06 (0.63-1.76)
Race/ethnicity (Ref: NH White)
NH Black	0.56 (0.36-0.89)	0.63 (0.40-1.01)
Hispanic	0.82 (0.59-1.14)	0.66 (0.48-0.90)
NH Asian	0.74 (0.43-1.30)	0.46 (0.27-0.77)
NH Other	1.74 (0.6-4.90)	1.60 (0.54-4.72)
Education (Ref: Less than high school)
High school graduate	2.99 (1.46-6.12)	1.59 (0.98-2.59)
Some college	4.95 (2.44-10.03)	2.08 (1.27-3.40)
College graduate or more	7.06 (3.52-14.14)	2.63 (1.63-4.25)
Health Insurance (Ref: Yes)
No	0.94 (0.39-2.30)	1.06 (0.56-2.01)
Regular Provider (Ref: Yes)
No	0.49 (0.35-0.70)	0.40 (0.29-0.56)
Use Internet (Ref: Yes)
No	0.28 (0.15-0.51)	0.30 (0.20-0.45)

Abbreviations: OR: odds ratio; CI: confidence interval; NH: Non-Hispanic; Ref: Reference. Bold indicates significantly different from the reference group (α = 0.05).

Model 2, which demonstrates the odds of being offered access to a patient portal by a healthcare provider, shows that living in a rural area (OR: 0.60; 95 percent CI: 0.39-0.91), identifying as Hispanic (OR: 66; 95 percent CI: 0.48-0.90) or NH Asian (OR: 0.46; 95 percent CI: 0.27-0.77), not having a regular healthcare provider (OR: 0.40; 95 percent CI: 0.29-0.56), and not having internet use (OR: 0.30; 95 percent CI: 0.20-0.45) were associated with lower odds of having been offered access to a patient portal.

Discussion

Coinciding with other recent studies, this study found no differences in providers maintaining electronic medical records whether in rural or urban areas.²⁹, ³⁰ Despite providers using electronic medical records at similar rates, the findings from both models in this study indicate rural residents have lower odds of being offered access to a patient portal and are subsequently less likely to actually access a patient portal. Adjusted models are consistent with rural, racial, and education disparities found in the literature on access of and being offered access to patient portals.^31-40 Additionally, these findings confirm previous studies on reasons patient portals were utilized⁴¹ and further confirm that there are no rural or urban differences in reasons for not wanting to access a patient portal.

This analysis demonstrates that despite continued attempts to focus on patient engagement, low patient portal use continues. Without providers encouraging and supporting use of patient portals, low adoption rates are likely to persist and potential benefits and improvements to the patient experience will go unrecognized.⁴², ⁴³ Previous studies have found that provider encouragement to use patient portals increases the odds of a patient using the portal by almost 10 times.⁴⁴ Helping patients understand the benefits of accessing their patient portal is critical to ensuring this technology is used to improve health outcomes for all. However, provider encouragement is only a piece of the puzzle to increasing patient portal use. Other studies have pointed to effective strategies that can be implemented at an organizational level, such as: providing ongoing technical assistance or continuing to monitor usage and satisfaction.

Limitations

Several limitations should be considered when evaluating the results of this research. The first is that the HINTS survey response rate was approximately 30 percent, leading to the potential that the findings of this study are not representative of the study population as a whole due to nonresponse bias. The survey is also limited to those living the United States and, therefore, cannot be generalized to other countries. In addition, as this study analyzes cross-sectional data, we cannot infer a causal relationship between rurality and odds of use of and access to patient portals. Further, given the limitations of the survey data, we were unable to control for a number of provider-level factors that could help to elucidate likelihood of patient exposure to patient portals or effectiveness of portals at the provider level. While additional data on patient preferences could be used to explain some of the reasons for the limited access of patient portals by rural populations and assist vendors in designing more useful patient portals, this analysis was limited to the data available in the HINTS survey. Finally, as HINTS data is self-reported, there is a strong potential for inaccurate recollection of past offers of access to such portals. Given the relative lack of assignment of social desirability value of actually accessing patient portals, there is little concern as to potential for social bias.

Conclusion

Patient portals hold the potential of further engaging patients in their care by connecting them to their own health information. However, our results indicate that there remain significant differences in the access and usage of patient portals by rural populations. Continued support in the form of financial incentives and legislative provisions must focus on addressing geographic and structural barriers faced by rural communities in adopting and embracing technology, such as patient portals. In the meantime, it is imperative to recognize that the digital divide continues to exist and increasing the availability of the internet and using provider encouragement as a tool to promote the use of patient portals continue to be necessary. Additional research must be conducted to determine effective strategies to overcome barriers to patient portal adoption among rural populations and determine specific features within patient portals that lead to consistent improvements in care processes and health outcomes.

Author Biographies

Grishma P. Bhavsar (grishma.bhavsar@csun.edu) is an associate professor in the Department of Health Sciences at California State University Northridge.

Dalton Pena (depena@uab.edu) is a PhD student in administration-health services at the University of Alabama, Birmingham.

Ashley S. Robertson (ashsrobe@iu.edu) is an assistant professor of health services management at Indiana University Purdue University Indianapolis/Indiana University Fort Wayne.