• Center on Health Equity & Access
  • Clinical
  • Health Care Cost
  • Health Care Delivery
  • Insurance
  • Policy
  • Technology
  • Value-Based Care

Race/Ethnicity, Personal Health Record Access, and Quality of Care

Publication
Article
The American Journal of Managed CareFebruary 2015
Volume 21
Issue 2

Quality benefits were equal across racial/ethnic groups with equal personal health record (PHR) use, but nonwhite status and a preference for Spanish language predicted lower PHR registration.

ABSTRACT

Objectives

To estimate the impact of race/ethnicity and written language preference on registration for a personal health record (PHR) that included emailing providers, viewing lab results, refilling prescriptions, and other functionalities, and the impact of PHR use on quality across racial/ethnic groups with comparable access.

Study Design and Methods

Retrospective observational design among 3,173,774 adults. Factors affecting registration were assessed using logistic regression, and propensity score matching techniques assessed the impact of language preference on registration and PHR use on quality of care. Difference-in-differences methods assessed the significance of between-group changes in Healthcare Effectiveness Data and Information Set (HEDIS) scores, such as glycated hemoglobin and lipid screening and control.

Results

Race/ethnicity most strongly predicted PHR registration. After adjusting for multiple factors, Asian American, Latino American, and African American members remained 23%, 55%, and 62% less likely to register, respectively, than non-Hispanic white members. Preference for Spanish as a written language predicted poor PHR adoption. The probability of registration was 0.451 (95% CI, 0.449-0.453) for English language—preferring Latinos and 0.174 (95% CI, 0.173-0.176) for Spanish language–preferring Latinos. For non- Hispanic whites, Latinos, and African Americans using the PHR, HEDIS scores increased after PHR use by 1.3 to 12.7 percentage points, compared with differences of –1.1 to 8.1 percentage points among nonusers. All but 2 difference-in-differences between PHR users and nonusers were statistically significant.

Conclusions

Nonwhite race/ethnicity and Spanish language preference independently predict poor PHR adoption. PHR use is associated with higher quality healthcare, and when PHR use is equivalent across racial/ethnic groups, so is quality of care.

Am J Manag Care. 2015;21(2):e103-e113

Although race/ethnicity and language preference disparities exist in personal health record (PHR) registration, the quality benefits of PHR use are equivalent when racial/ ethnic groups have comparable access.

  • After adjusting for multiple factors, Asian Americans, Latino Americans, and African Americans remained 23%, 55%, and 62% less likely to register, respectively, for the PHR than non-Hispanic whites.
  • The probability of registration was 0.451 (95% CI, 0.449-0.453) for English language— preferring Latinos and 0.174 (95% CI, 0.173-0.176) for Spanish language– preferring Latinos.
  • PHR use is associated with higher quality healthcare; when PHR use is equivalent across racial/ethnic groups, so are quality gains.

Robust Internet-based patient portals and personal health records (PHRs) provide patients with access to personal health information and supporting patient education materials, provide secure messaging with providers, and offer functions to schedule appointments and refill prescriptions.1 In 2010, 7% of Americans used a PHR, although much higher registration rates occurred in integrated systems with robust and well-established PHRs, such as the Veterans Health Administration and Kaiser Permanente (KP).2-4 PHRs hold promise to enhance care quality, improve patient-provider relationships, and encourage healthy lifestyle behaviors.5,6 For instance, we previously demonstrated that among more than 35,000 patients with diabetes and/or hypertension, the use of secure e-mail with providers through a PHR was associated with improved Healthcare Effectiveness Data and Information Set (HEDIS) scores.7,8 The relationship between patient portal use and improved glycemic control and other quality measures among patients with diabetes has also been documented.9-12

Racial/ethnic disparities in PHR registration and use exist1,13-21; however, few recent studies designated Latinos or Hispanics as a racial/ethnic category and included patients whose preferred language was not English.1,13,21 Race/ethnicity is one factor of many that likely influences PHR registration; recent studies also uniformly examined age and gender. However, healthcare utilization was only measured in an underserved population,21 and health status was measured by the presence of a limited number of chronic conditions.1,21 To the best of our knowledge, no evidence exists as to the impact of written language preference on registration rates, which is logically more germane to using a PHR than is spoken language.21 Finally, no published studies document the quality impact of PHR use across racial/ethnic groups. An important question is, if patients with differing race/ethnicity have equivalent PHR use, do they experience similar clinical benefits?22

1. Registration rates among non-Hispanic whites would be higher than among members of any other racial/ethnic group. Other factors would affect registration rates, such as age, gender, disease burden, healthcare utilization, length of membership, and distance to the nearest medical office, but race/ethnicity would have the strongest influence on registration rates. To test this hypothesis, we examined registration rates and the impact of race/ethnicity relative to other factors.

2. Non-English language preference would reduce the likelihood of PHR registration, regardless of race/ethnicity. To test this, we assessed the impact of English and Spanish language preference among Latinos on registration.

3. Once barriers to access were removed, quality benefits of PHR use would accrue equally across racial/ethnic groups. To test this, we assessed the effect of PHR use on healthcare quality across racial/ethnic groups with comparable PHR access.

METHODS

We used a retrospective observational design to investigate the impact of race/ethnicity on PHR registration relative to other factors, and a propensity score—matched pair cohort design to investigate the impact of written language preference on registration and the effect of PHR use on quality of care in racial/ethnic groups with comparable PHR access.

Personal Health Record

Kaiser Permanente (KP) is the largest not-for-profit integrated healthcare delivery system in the United States, serving 9 million members. KP’s PHR, My Health Manager, is integrated with the electronic health record (EHR), KP HealthConnect. Features include access to portions of the medical record, test results, patient education, prescription refills, appointment scheduling, and the ability to securely e-mail providers. Patients must register for and activate an account to sign on to use these features. Access is free to all members who receive marketing information about the PHR, regardless of their race/ethnicity.

The PHR is currently available primarily in English. Although many patient education materials are available in Spanish, English is currently the only language option for complete navigation and for access to the health record, including test results, prescription refills, appointment scheduling, and secure e-mail with providers. As of March 2013, 65% of all age-eligible KP members were PHR-registered. The PHR has become a significant mode of care delivery; 28% of contacts between patients and primary care providers occur by secure e-mail.

Study Cohorts

To investigate the impact of race/ethnicity on PHR registration relative to other factors, we included all patients 18 years and older with active KP memberships on December 31, 2010, and with complete data for all variables of interest. To investigate the impact of written language preference on registration, we limited the population to Latino patients—the only group composed of enough individuals with varying self-reported language preferences (English and Spanish). To assess the impact of PHR use on quality, we included Latino, non-Hispanic white, and African American members with diabetes or hypertension from 1 region (KP Southern California, KPSC) who registered for the PHR and activated their accounts between January 1, 2006, and October 31, 2010, and who had active memberships between January 1, 2009, and December 31, 2010.

Outcome Measures, Covariates, and Data Sources

A main outcome measure was PHR registration. We defined members as registered if they had ever signed up for PHR access. Quality outcome measures were scores on 7 HEDIS effectiveness-of-care measures: 6 comprehensive diabetes control measures (lipid and glycemic screening and control, and retinopathy and nephropathy screening) and 1 control measure for diagnosed hypertension.23 Baseline HEDIS scores came from the calendar year before the one in which members registered for the PHR and follow-up scores came from the subsequent calendar year. Quality outcomes were examined in relationship to PHR use, which we defined as registration for and activation of an account.7

We included covariates that may have also affected registration and were used to create propensity score-matched pairs: race/ethnicity, age, gender, language preference, distance to the nearest medical office building, number of annual office visits, illness burden (measured by current DxCG risk scores derived from proprietary [Verisk Health, Inc] software forecasting risk-adjusted utilization24), and KP membership tenure. Data on age, gender, diagnoses, membership tenure, distance to the nearest medical office, illness burden, utilization of office visits, PHR registration, HEDIS effectiveness-of-care measures, and self-reported race/ethnicity and language preference were all obtained from the EHR.

Propensity Score Matching

Table 1

Impact of language preference on registration. We calculated a propensity score for each Latino member with a self-reported preference for English or Spanish as a written language, using SAS 9.2 (SAS Institute, Inc, Cary, North Carolina). The propensity score included the independent variables of age, gender, DxCG score, tenure, number of annual office visits, distance from nearest medical office building, and region. We used the greedy matching algorithm without replacement to create 209,873 matched pairs out of 211,995 Latino members preferring Spanish and 521,235 Latino members preferring English.25 Propensity scores were matched to the second decimal place ().

Table 2

Impact of PHR use on quality across racial/ethnic groups with comparable access. We used propensity score matching within Latino, non-Hispanic white, and African American populations because they yielded sufficient sample sizes. For each HEDIS measure, we calculated a propensity score for each member with diabetes or hypertension. The propensity score included the independent variables of age, gender, illness burden, tenure, number of annual office visits, distance from nearest medical office building, and baseline HEDIS score. We used greedy matching without replacement to pair each member who used the PHR to a member of the same race/ethnicity who had not used it.20 The number of propensity score-matched pairs within racial/ethnic groups varied by measure, from 4481 (A1C control among African Americans with diabetes) to 28,612 (blood pressure control among non-Hispanic whites with hypertension). Propensity scores were matched to the second decimal place, and no difference in baseline periods was allowed. provides an example of the pre- and post matching balance for 1 measure (A1C screening) among Latinos.25

Statistical Analysis

Logistic regression was used to assess factors affecting PHR registration for the entire population. Registration was the dependent variable, and race/ethnicity, age, gender, illness burden, number of annual office visits, distance from the nearest medical office, region, and tenure were independent variables. Modeling was performed with SAS 9.2.

To get the propensity score—matched data for the impact of language preference on registration, we calculated probabilities from the proportions of registered and unregistered members and bootstrapped the confidence intervals (CIs).26 Similarly, we used bootstrapping methods to compute CIs for HEDIS scores, and compared outcomes for propensity score-matched African American, Latino, and non-Hispanic white members with diabetes and/or hypertension who had and had not used the PHR. We assessed both the absolute value of changes before and after PHR use, and the difference-in-differences. We received an institutional review board waiver for this quality assessment project.

RESULTS

Table 3

Impact of race/ethnicity and other factors on PHR registration. Among 3,173,774 members meeting our inclusion criteria, 1,764,121 (56%) were registered to use the PHR (). Women, people 30 years or older, and non-Hispanic white members were more likely to register than men, younger individuals, and members of any other race/ethnicity. Of registered members, 57% were female, and 59% were aged 20 to 64 years. Among registered members, 10% made no office visits during the year, compared with 18% of unregistered members.

Table 4

The strongest predictors of PHR registration were race/ethnicity, number of annual office visits, age, and language preference, in decreasing order of strength. Other significant predictors—but to a lesser degree—were gender, illness burden, distance to the nearest medical office, and tenure. In terms of race/ethnicity, the odds of a member being non-Hispanic white and registered were greater than the odds of being any other race/ethnicity and registered. Asian/Pacific Islander members were 23% less likely, Hispanic members were 55% less likely, and African American members were 62% less likely to have registered than were non-Hispanic white members (). Preference for a written language other than English was also associated with substantial reductions in the likelihood of registration, and more annual office visits were associated with very substantial increases in the likelihood of registration. Age and gender also affected registration rates.

eAppendix Table

KP does not routinely collect data on income and education, and we did not include these variables in our analyses. However, to assess the extent to which excluding income and education might have affected our results, we repeated the logistic regression analysis using imputed data for these variables. The results varied insignificantly (, available at www.ajmc.com).

Impact of language on PHR registration. More than 90% of registered members preferred English as their primary language. Among 209,873 matched pairs of English language— and Spanish language–preferring Hispanic members, the probability of registration was 0.451 (95% CI, 0.449-0.453) for the former and 0.174 (95% CI, 0.173-0.176) for the latter.

Impact of race/ethnicity on quality measures. A total of 171,054 KPSC members with diabetes had complete data and were included; 41% registered for the PHR between January 1, 2006, and October 31, 2010. Similarly, 279,019 members with hypertension had complete data and were included; 42% registered during the same time period. The number of propensity score-matched pairs varied by measure, from 5098 (screening for retinopathy among African Americans with diabetes) to 28,061 (blood pressure control among non-Hispanic whites with hypertension).

Table 5

Table 5

Absolute increases in HEDIS scores before and after PHR use ranged from 1.3 to 12.7 percentage points (). Among nonusers, differences in HEDIS scores over the same time period ranged from a decrease of 1.1 percentage points to an increase of 8.1 percentage points. For users and nonusers, the largest improvements occurred in low-density lipoprotein cholesterol control and retinopathy screening among members with diabetes and in blood pressure control among members with hypertension. The difference-in-differences analysis revealed that greater improvements in HEDIS scores occurred across all racial/ethnic groups in PHR users, as compared with nonusers (). All but 2 of the differential increases were statistically significant.

DISCUSSION

Among more than 3 million KP members, racial and ethnic disparities in PHR registration existed after adjusting for potential confounding variables; nonwhite race/ethnicity independently predicted nonregistration. Spanish-language preference also independently predicted a substantially lower likelihood of PHR registration. However, quality benefits of PHR use accrued equally across racial/ethnic groups with comparable PHR access.

Our results, obtained from a very large and diverse population, confirm those of other studies documenting that minority status is associated with lower PHR registration rates.1,13-21 An association between higher use of clinical services and PHR use is also documented elsewhere.15,21,27 To the best of our knowledge, our study is the first to document the impact of written language preference on PHR registration among members of a single racial/ethnic group and equivalent quality benefits across groups with comparable PHR access.

Within a large, insured population, differences in PHR registration exist according to race/ethnicity and language preferences, but quality benefits of PHR use accrue equitably across racial/ethnic groups. The implications are clear: initiatives or organizations promoting the use of PHRs should recognize their varying adoption. Further research is needed to understand the barriers to PHR registration within and across racial/ ethnic/language-preference groups. In general, even among underserved populations, a lack of Internet or computer access is not to blame.28-30 An internal KP study also found that adjusting for income and education does not reduce disparities in access. Suggested reasons for observed disparities include varying preferences for communicating with providers, lower computer literacy, and distrust of Web-based communications for sensitive health information.28,31

Explanations have also been suggested for the impact of language preferences on PHR registration. The simplest explanation is difficulty in navigating an Englishbased PHR. Additional barriers may include a scarcity of bilingual physicians and frequent lack of language concordance between patient and physician.32 Unpublished results from KP focus groups exploring PHR registration among nonregistered Spanish-speaking members indicate that barriers unrelated to language include concerns about security of personal information; this is confirmed in the literature.2 The prevalence and force of these barriers needs further investigation, as does the relationship between registration and subsequent PHR use across a variety of populations.21,15

The finding that quality benefits associated with PHR use accrue equally across racial/ethnic groups requires confirmation in additional studies, but it is an important step toward building an evidence base documenting that the benefits of technology span racial/ethnic boundaries.22 Importantly, it suggests that differential registration rates may increase health disparities. A key implication of our project relates to the need for further work to ensure that access to health information technology ameliorates, rather than worsens, racial/ethnic healthcare disparities.

Limitations

Although we addressed some confounders, our findings may have been influenced by other unmeasured factors. For instance, we did not have data on Internet access, computer literacy, or cultural or other factors related to the use of PHRs. A limitation pertaining to our assessment of the impact of PHR use on quality was our inability to center individual pre- and post HEDIS scores more precisely on the time at which individuals first used the PHR. Propensity score—matched groups for the HEDIS measures varied substantially in size, which almost certainly limited our ability to detect significant difference-in-differences in smaller groups. Baseline HEDIS scores were, in some cases, high enough that subsequent scores did not prove our hypothesis that equal PHR access would confer equal quality benefits. Finally, our project occurred in an integrated healthcare delivery system, which may limit the generalizability of our results.33

CONCLUSIONS

The quality benefits of PHR use accrue equitably across racial and ethnic groups with comparable access, even though registration rates vary substantially. The increasing availability of PHRs may have the potential to ameliorate some racial and ethnic differences in health outcomes and health status. However, if disparities in PHR access persist, health disparities may also increase.

Acknowledgments

We acknowledge the many KP providers enabling the integrated care experience through our patient portal, My Health Manager, that allow this potential to be true for so many members. We thank Jed Weissberg, MD, formerly of Kaiser Permanente, for his seminal question on access across different race/ethnic groups. Luther Scott’s expertise in geocoding at Kaiser Permanente was pivotal to an earlier version of this work. Jennifer Green, of Kaiser Permanente, provided editorial assistance.

We had several hypotheses:Author Affiliations: Health Information Technology Transformation & Analytics, Kaiser Permanente, Oakland, CA (DM, TG), Portland, OR (MT, JW); Southern California Permanente Medical Group (MK), Pasadena, CA; National Market Research, Kaiser Permanente (VS), Pleasanton, CA; Utility for Care Data Analysis, Kaiser Permanente (LS), Portland, OR.

Source of Funding: None.

Author Disclosures: The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (TG, MK, DM, VS, JW); acquisition of data (DM, LS, VS, JW); analysis and interpretation of data (MK, LS, JW); drafting of the manuscript (TG, MK); critical revision of the manuscript for important intellectual content (TG, MK, DM); statistical analysis is (DM, LS, JW); provision of study materials or patients (VS, JW); and supervision (TG, MK).

Address correspondence to: Di Meng, PhD, director of analytics & evaluation, Kaiser Permanente — HITTA, 1800 Harrison St, Oakland, CA 94612. E-mail: Di.X.Meng@kp.org.REFERENCES

1. Yamin CK, Emani S, Williams DH, et al. The digital divide in adoption and use of a personal health record. Arch Intern Med. 2011;171(6):568-574.

2. Consumers and health information technology: a national survey. California HealthCare Foundation website. http://www.chcf.org/publications/2010/04/consumers-and-health-information-technology-anational-survey. Published April 2010. Accessed February 4, 2015.

3. Woods S. Veterans Health Administration, Department of Veterans Affairs-patient engagement and the blue button. In: Oldenburg J, Chase D, Chistensen KT, Tritle B, eds. Engage! Transforming Healthcare through Digital Patient Engagement. Chicago, IL: HIMSS Books; 2012:222-224.

4. McCann E. Kaiser PHR sees 4 million sign on, most active portal to date. Healthcare IT News website. http://www.healthcareitnews.com/news/kaiser-phr-sees-4-million-sign-most-active-portal-date. Published August 6, 2012. Accessed February 4, 2015.

5. Kreps GL, Neuhauser L. New directions in eHealth communication: opportunities and challenges. Patient Educ Couns. 2010;78(3):329-336.

6. Dixon RF. Enhancing primary care through online communication. Health Aff (Millwood). 2010;29(7):1364-1369.

7. Zhou YY, Kanter MH, Wang JJ, Garrido T. Improved quality at Kaiser Permanente through e-mail between physicians and patients. Health Aff (Millwood). 2010;29(7):1370-1375.

8. HEDIS & performance measurement. National Committee for Quality Assurance website. http://www.ncqa.org/HEDISQualityMeasurement.aspx. Published 2014. Accessed February 4, 2015.

9. Osborn CY, Mayberry LS, Wallston KA, Johnson KB, Elasy TA. Understanding patient portal use: implications for medication management. J Med Internet Res. 2013;15(7):e133.

10. Wade-Vuturo AE, Mayberry LS, Osborn CY. Secure messaging and diabetes management: experiences and perspectives of patient portal users. J Am Med Inform Assoc. 2013;20(3):519-525.

11. Grant RW, Wald JS, Schnipper JL, et al. Practice-linked online personal health records for type 2 diabetes mellitus: a randomized controlled trial. Arch Intern Med. 2008;168(16):1776-1782.

12. Tenforde M, Nowacki A, Jain A, Hickner J. The association between personal health record use and diabetes quality measures. J Gen Intern Med. 2012;27(4):420-424.

13. Goel MS, Brown TL, Williams A, Hasnain-Wynia R, Thompson JA, Baker DW. Disparities in enrollment and use of an electronic patient portal. J Gen Intern Med. 2011;26(10):1112-1116.

14. Hsu J, Huang J, Kinsman J, et al. Use of e-Health services between 1999 and 2002: a growing digital divide. J Am Med Inform Assoc. 2005;12(2):164-171.

15. Miller H, Vandenbosch B, Ivanov D, Black P. Determinants of personal health record use: a large population study at Cleveland Clinic. J Healthc Inf Manag. 2007;21(3):44-48.

16. Lyles CR, Harris LT, Jordan L, et al. Patient race/ethnicity and shared medical record use among diabetes patients. Med Care. 2012;50(5):434-440.

17. Byczkowski TL, Munafo JK, Britto MT. Variation in use of Internet-based patient portals by parents of children with chronic disease. Arch Pediatr Adolesc Med. 2011;165(5):405-411.

18. Roblin DW, Houston TK II, Allison JJ, Joski PJ, Becker ER. Disparities in use of a personal health record in a managed care organization. J Am Med Inform Assoc. 2009;16(5):683-689.

19. Sarkar U, Karter AJ, Liu JY, et al. Social disparities in internet patient portal use in diabetes: evidence that the digital divide extends beyond access. J Am Med Inform Assoc. 2011;18(3):318-321.

20. Tsai J, Rosenheck RA. Use of the Internet and an online personal health record system by US veterans: comparison of Veterans Affairs mental health service users and other veterans nationally. J Am Med Inform Assoc. 2012;19(6):1089-1094.

21. Ancker JS, Barron Y, Rockoff ML, et al. Use of an electronic patient portal among disadvantaged populations. J Gen Intern Med. 2011;26(10):1117-1123.

22. Sequist TD. Health information technology and disparities in quality of care. J Gen Intern Med. 2011;26(10):1084-1085.

23. HEDIS 2010. National Committee for Quality Assurance website. http://www.ncqa.org/Portals/0/HEDISQM/HEDIS2010/2010_Measures.pdf. Published 2011. Accessed February 4, 2015.

24. DxCG Risk Analytics. Verisk Health website. http://www.veriskhealth.com/solutions/enterprise-analytics/dxcg-intelligence. Published 2014. Accessed February 4, 2015.

25. Parsons LS. Reducing bias in a propensity score matched-pair sample using greedy matching techniques. Paper presented at: The Twenty-Sixth Annual SAS (Users Group International Conference); April 22-25, 2001; Long Beach, California. http://www2.sas.com/proceedings/sugi26/p214-26.pdf. Accessed February 4, 2015.

26. Efron B, Tibshirani R. An Introduction to the Bootstrap. London: Chapman & Hall/CRC; 1993.

27. Ketterer T, West DW, Sanders VP, Hossain J, Kondo MC, Sharif I. Correlates of patient portal enrollment and activation in primary care pediatrics. Acad Pediatr. 2013;13(3):264-271.

28. Luque AE, van Keken A, Winters P, Keefer MC, Sanders M, Fiscella K. Barriers and facilitators of online patient portals to personal health records among persons living with HIV: formative research. JMIR Res Protoc. 2013;2(1):e8.

29. Sanders MR, Winters P, Fortuna RJ, et al. Internet access and patient portal readiness among patients in a group of inner-city safety-net practices. J Ambul Care Manage. 2013;36(3):251-259.

30. Lopez L, Grant RW. Closing the gap: eliminating health care disparities among Latinos with diabetes using health information technology tools and patient navigators. J Diabetes Sci Technol. 2012;6(1):169-176.

31. Goel MS, Brown TL, Williams A, Cooper AJ, Hasnain-Wynia R, Baker DW. Patient reported barriers to enrolling in a patient portal. J Am Med Inform Assoc. 2011;18(suppl 1):i8-i12.

32. Kanter MH, Abrams KM, Carrasco MR, Spiegel NH, Vogel RS, Coleman KJ. Patient-physician language concordance: a strategy for meeting the needs of Spanish-speaking patients in primary care. Perm J. 2009;13(4):79-84.

33. Berwick DM. Making good on ACOs’ promise—the final rule for the Medicare shared savings program. N Engl J Med. 2011;365(19):1753-1756.

Related Videos
dr jennifer green
dr jennifer green
dr ken cohen
dr ian neeland
Yael Mauer, MD, MPH
Pregnant Patient | image credit: pressmaster - stock.adobe.com
Diana Isaacs, PharmD
Beau Raymond, MD
Robert Zimmerman, MD
Beau Raymond, MD
Related Content
© 2024 MJH Life Sciences
AJMC®
All rights reserved.