Erratum: Access to Electronic Personal Health Records Among Patients With Multiple Chronic Conditions: A Secondary Data Analysis (Journal of Medical Internet Research (2017) 19:6 (e188) DOI: 10.2196/jmir.7417)

The authors were notified of data errors in one of the Health Information National Trends Survey (HINTS) cycle datasets (HINTS 4, Cycle 4); the errors were in the weights provided for use in the analysis of these data [1]. Following the HINTS error notice [1], the authors reran analyses reported in Table 1 and Table 2. The originally published versions of these tables are in Multimedia Appendix 1 and Multimedia Appendix 2. In rerunning analyses for Table 1, only one difference was found that resulted in a change in conclusion. Namely, the chi-squared analysis of “Confidence that PHI is safe” versus “Number of Chronic Conditions” was not significant in the updated analysis (P=.11). In the original analysis, the results were significant, with “Very Confident” more likely with two or more chronic conditions, and “Not Confident” more likely with no chronic conditions. Additionally, for “Accessed EHRs at least once” and “Frequency of EHR Access,” the original table used column percentages instead of row percentages; this has been corrected in the updated table. In section “Associations Between Patient Factors and Number of Chronic Conditions” in the Results, the fourth sentence in the first paragraph originally read as follows: In general, having two or more chronic conditions was associated with being older, having health insurance, having a regular provider, being less confident in taking care of themselves, reporting fair to poor health, being less inclined to use the Internet or to use a mobile phone/tablet, and in feeling more confident that their PHI is safe and controllable (Table 1). It has been corrected as follows: In general, having two or more chronic conditions was associated with being older, having health insurance, having a regular provider, being less confident in taking care of themselves, reporting fair to poor health, and being less inclined to use the Internet or to use a mobile phone/tablet (Table 1). The following text from the Discussion was no longer accurate and has been removed from the corrected version of the article: Additionally, HINTS included items addressing concerns about safety and privacy of electronic health information, which revealed that those with MCC reported slightly higher frequencies of believing that they were “very confident” in having control of the privacy of their records and that their PHI was safe with their providers. This could provide one explanation for the increased use of ePHR among those with MCC. In rerunning analyses for Table 2, only one difference was found that resulted in a change in conclusion. Namely, for “Confidence that PHI is safe,” a significant association was found for “Very Confident” (OR 2.00, 95% CI 1.21-3.31; P=.01) and “Somewhat Confident” (OR 1.99, 95% CI 1.25-3.17) as compared to the reference of “Not Confident.” In the original analysis, neither was statistically significant. This change does not affect any of the text within the body of the manuscript. But it is a notable new conclusion, indicating that those who are more confident about the safety of their data are significantly more likely to use ePHR than those who are not confident about such security. The corrected versions of Table 1 and Table 2 are below: (Table Presented).