CLINICAL RESEARCH STUDY
Perceptions of Electronic Health Record Implementation: A Statewide Survey of Physicians in Rhode Island Matthew C. Wylie, MD,a Rosa R. Baier, MPH,a,b Rebekah L. Gardner, MDb,c a c
Brown University School of Public Health, Providence, RI; bHealthcentric Advisors, Providence, RI; Warren Alpert Medical School of Brown University, Providence, RI.
ABSTRACT OBJECTIVES: Although electronic health record use improves healthcare delivery, adoption into clinical practice is incomplete. We sought to identify the extent of adoption in Rhode Island and the characteristics of physicians and electronic health records associated with positive experience. METHODS: We performed a cross-sectional study of data collected by the Rhode Island Department of Health for the Health Information Technology Survey 2009 to 2013. Survey questions included provider and practice demographics, health record information, and Likert-type scaled questions regarding how electronic health record use affected clinical practice. RESULTS: The survey response rate ranged from 50% to 65%, with 62% in 2013. Increasing numbers of physicians in Rhode Island use an electronic health record. In 2013, 81% of physicians used one, and adoption varied by clinical subspecialty. Most providers think that electronic health record use improves billing and quality improvement but has not improved job satisfaction. Physicians with longer and more sophisticated electronic health record use report positive effects of introduction on all aspects of practice examined (P < .001). Older physician age is associated with worse opinion of electronic health record introduction (P < .001). Of the 18 electronic health record vendors most frequently used in Rhode Island, 5 were associated with improved job satisfaction. CONCLUSIONS: We report the largest statewide study of electronic health record adoption to date. We found increasing physician use in Rhode Island, and the extent of adoption varies by subspecialty. Although older physicians are less likely to be positive about electronic health record adoption, longer and more sophisticated use are associated with more positive opinions, suggesting acceptance will grow over time. Ó 2014 Elsevier Inc. All rights reserved. The American Journal of Medicine (2014) 127, 1010.e21-1010.e27 KEYWORDS: Electronic health record; Electronic medical record; Physician experience
As part of the American Recovery and Reinvestment Act of 2009, the Health Information Technology for Economic and Clinical Health Act allocated $29 billion over 10 years for supporting the adoption and “meaningful use” of Funding: MCW work is supported by an educational stipend from the University Emergency Medical Foundation. Conflict of Interest: None. Authorship: All authors had access to the data and played a role in writing this manuscript. Requests for reprints should be addressed to Matthew C. Wylie, MD, Brown University School of Public Health, 125 South Main St, Providence, RI 02903. E-mail address: [email protected] 0002-9343/$ -see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjmed.2014.06.011
electronic health records.1 Electronic health records have great potential to improve healthcare by facilitating fast accurate patient data transmission, standardizing medical processes, enabling decision support, and allowing real-time medical error prevention, to name the major benefits. Although the potential gains from adoption of such technology make intuitive sense, research has just started to demonstrate the efficacy of electronic health record use in the United States. Natural experiments with electronic health record introduction in a large integrated hospital system have shown improvement in adherence to evidencebased practice, overall reduction of markers of disease, decreases in hospital admissions, and emergency
Wylie et al
Electronic Health Record Implementation in Rhode Island
1010.e22
department visits.2,3 Other studies have shown decreases did not complete the survey were excluded from these in total healthcare cost, ambulatory cost growth over analyses. time,4 improved process of care,5 and improved performance in disease screening.6 Study Measurements Despite this great potential and the growing evidence Survey questions included provider and practice defor efficacy of the electronic health record, physician mographics, information regarding electronic health record implementation of electronic system, and Likert-type scale health records is far from univerquestions regarding how elecCLINICAL SIGNIFICANCE sal. As of 2011, only 54% of tronic health records have affected physicians have adopted a com Some 81% of physicians in Rhode Island medical practice. Only closedplete electronic health record ended questions were used for use an electronic health record. system or have some mix of this study. Younger age, more patients with electronic and paper records. Medicaid insurance, hospital practice, Some of these systems reported as Data Analysis “mixed” can be rudimentary too, practice in larger groups, primary care because programs like FileMaker Survey response was calculated practice, and nonsurgical specialization Pro (FileMaker Inc, Santa Clara, using the American Association are associated with increased electronic Calif) can qualify as an electronic for Public Opinion Research calhealth record use. health record. As the first and culator V3.1.8 Responses were Most respondents think that electronic only state (to our knowledge) to downloaded into Microsoft Excel health record adoption has decreased systematically assess and publi(Microsoft Corp, Redmond, WA) job satisfaction. Experience varied cally report electronic health reand analyzed in STATA 13.0 cord adoption for every licensed (StataCorp LP, College Station, significantly by electronic health record physician providing patient care, Tex). Missing, “not applicable,” vendor. Rhode Island provides an opporand other responses were excluded Longer and more sophisticated use were tunity to identify characteristics from analysis, and “don’t know” associated with a more favorable opinion of electronic health records and responses were reported only of electronic health records. physicians that might enable betwhen pertinent. ter electronic health record inteSummary statistics were used gration in the future. to describe respondents to the In this study of all physicians currently practicing in 2013 survey. Respondents were asked how electronic Rhode Island, we sought to accomplish 3 goals: (1) describe health record adoption had affected 6 elements of medical adoption of electronic health records by Rhode Island practice: job satisfaction, billing, internal communication, physicians; (2) describe how electronic health record patient care, quality improvement, and workflow. Responses implementation has changed elements of physician practice, were extent of agreement with a statement of the pattern such as workflow, patient care, and job satisfaction; and (3) “electronic health record adoption improved Job Satisfacfind characteristics of physicians and their electronic health tion,” for example, selecting one of “strongly disagree,” record system associated with positive experiences. We “disagree,” “agree,” or “strongly agree.” hypothesized that older age, late electronic health record A composite measure of how advanced a physician’s adoption, and partial electronic health record adoption electronic health record was developed by assigning a would be associated with more negative physician point for every function that was used and summing them experiences. to yield a score out of 12: (1) drug interaction warnings at point of prescription; (2) prompts for patient-specific care at point of care; (3) reminder regarding indicated or overdue care; (4) patient demographics; (5) electronic list of MATERIALS AND METHODS medications; (6) electronic progress notes; (7) electronic Study Design problem lists; (8) clinical summaries for transfer of care or This study has a cross-sectional survey design. referral; (9) report of clinical quality measures; (10) transfer of laboratory test results electronically; (11) transfer of radiologic test results electronically; and (12) Study Setting and Population electronic prescribing. There were 122 electronic health record vendors in use by Rhode Island physicians in 2013, The Rhode Island Department of Health administered the so a variable was created including only electronic health Health Information Technology Survey electronically record used by 10 or more physicians, thereby allowing a over the course of 3 weeks in January and February of more manageable analysis. 2009 to 2013.7 The department used its licensure database Student t test was used for continuous variables, and to send letters and e-mails (when available) to all licensed chi-square test was used for categoric variables to physicians providing direct patient care. Participants who
1010.e23
The American Journal of Medicine, Vol 127, No 10, October 2014 training are less likely to have an electronic health record than other types of doctors (Figure 1). Over time, increasing numbers of physicians in Rhode Island have reported use of an electronic health record in their primary practice, and in 2013, 81% of physicians used an electronic health record (Figure 2). Most providers think that electronic health record introduction has not improved their job satisfaction (Figure 3). However, they also report that electronic health record introduction improved billing and quality improvement. Physicians are evenly divided on whether electronic health record introduction improved workflow, and slim majorities agree that electronic health record introduction improved internal communication and patient care. Physicians with electronic health record in place for longer are significantly more likely to report a positive effect of electronic health record introduction on all aspects of clinical practice examined (P < .001). Physicians who report more sophisticated electronic health record use (ie, use of electronic health record for more aspects of their practice, documentation, and prescription-writing) also were more likely to report a positive impact of electronic health record adoption on all aspects of clinical practice (P < .001). Older age is associated with a worse opinion of the effect of electronic health record introduction, because those who disagreed and strongly disagreed with statements
demonstrate differences between physicians with and without an electronic health record. Histograms and boxplots were generated to show physician opinion of the effect of electronic health record adoption. Analysis of variance and Kruskal-Wallis tests were used to test for differences between provider opinions of electronic health record adoption effect and the following characteristics: brand of electronic health record, age of provider, years of electronic health record use, and level of electronic health record technology. Statistical significance was taken to be P < .05.
RESULTS The survey response rate was 62% in 2013, with 2365 of 3799 physicians completing surveys. Most doctors work more than 20 hours per week, have Doctor of Medicine degrees, have their primary practice in an office, are primary care providers, and work in groups of more than 10 providers (Table 1). There are differences between respondents with and without an electronic health record. Respondents with an electronic health record are likely to have larger populations of patients with Medicaid insurance, work in hospitals, work in larger practices, and practice primary care. Physicians with surgical subspecialty and psychiatric
Table 1
Characteristics of Respondents Respondents Without EHS
Total Age (y) Hours of patient care per week: 20 h License type DO MD Percent of patients using Medicaid 0% 60% Primary practice is: Hospital Office Size of practice 10 clinicians Primary care physician Yes No
All Respondents
% (n) or Mean (SD)
Respondents With EHS
2365 50.9 (0.2)
19.2 (454) 56.6 (0.6)
80.8 (1911) 49.5 (0.2)
5.9 (136) 13.3 (308) 80.9 (1877)
5.5 (24) 14.1 (62) 80.5 (354)
6.0 (112) 13.1 (246) 81.0 (1521)
5.9 (139) 94.1 (2228)
5.1 (23) 94.9 (431)
6.1 (116) 93.9 (1795)
14.2 66.9 14.0 4.8
22.8 69.5 5.6 2.1
10.9 65.9 17.3 5.9
P Value
Perceptions of Electronic Health Record Implementation: A Statewide Survey of Physicians in Rhode Island Matthew C. Wylie, MD,a Rosa R. Baier, MPH,a,b Rebekah L. Gardner, MDb,c a c
Brown University School of Public Health, Providence, RI; bHealthcentric Advisors, Providence, RI; Warren Alpert Medical School of Brown University, Providence, RI.
ABSTRACT OBJECTIVES: Although electronic health record use improves healthcare delivery, adoption into clinical practice is incomplete. We sought to identify the extent of adoption in Rhode Island and the characteristics of physicians and electronic health records associated with positive experience. METHODS: We performed a cross-sectional study of data collected by the Rhode Island Department of Health for the Health Information Technology Survey 2009 to 2013. Survey questions included provider and practice demographics, health record information, and Likert-type scaled questions regarding how electronic health record use affected clinical practice. RESULTS: The survey response rate ranged from 50% to 65%, with 62% in 2013. Increasing numbers of physicians in Rhode Island use an electronic health record. In 2013, 81% of physicians used one, and adoption varied by clinical subspecialty. Most providers think that electronic health record use improves billing and quality improvement but has not improved job satisfaction. Physicians with longer and more sophisticated electronic health record use report positive effects of introduction on all aspects of practice examined (P < .001). Older physician age is associated with worse opinion of electronic health record introduction (P < .001). Of the 18 electronic health record vendors most frequently used in Rhode Island, 5 were associated with improved job satisfaction. CONCLUSIONS: We report the largest statewide study of electronic health record adoption to date. We found increasing physician use in Rhode Island, and the extent of adoption varies by subspecialty. Although older physicians are less likely to be positive about electronic health record adoption, longer and more sophisticated use are associated with more positive opinions, suggesting acceptance will grow over time. Ó 2014 Elsevier Inc. All rights reserved. The American Journal of Medicine (2014) 127, 1010.e21-1010.e27 KEYWORDS: Electronic health record; Electronic medical record; Physician experience
As part of the American Recovery and Reinvestment Act of 2009, the Health Information Technology for Economic and Clinical Health Act allocated $29 billion over 10 years for supporting the adoption and “meaningful use” of Funding: MCW work is supported by an educational stipend from the University Emergency Medical Foundation. Conflict of Interest: None. Authorship: All authors had access to the data and played a role in writing this manuscript. Requests for reprints should be addressed to Matthew C. Wylie, MD, Brown University School of Public Health, 125 South Main St, Providence, RI 02903. E-mail address: [email protected] 0002-9343/$ -see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjmed.2014.06.011
electronic health records.1 Electronic health records have great potential to improve healthcare by facilitating fast accurate patient data transmission, standardizing medical processes, enabling decision support, and allowing real-time medical error prevention, to name the major benefits. Although the potential gains from adoption of such technology make intuitive sense, research has just started to demonstrate the efficacy of electronic health record use in the United States. Natural experiments with electronic health record introduction in a large integrated hospital system have shown improvement in adherence to evidencebased practice, overall reduction of markers of disease, decreases in hospital admissions, and emergency
Wylie et al
Electronic Health Record Implementation in Rhode Island
1010.e22
department visits.2,3 Other studies have shown decreases did not complete the survey were excluded from these in total healthcare cost, ambulatory cost growth over analyses. time,4 improved process of care,5 and improved performance in disease screening.6 Study Measurements Despite this great potential and the growing evidence Survey questions included provider and practice defor efficacy of the electronic health record, physician mographics, information regarding electronic health record implementation of electronic system, and Likert-type scale health records is far from univerquestions regarding how elecCLINICAL SIGNIFICANCE sal. As of 2011, only 54% of tronic health records have affected physicians have adopted a com Some 81% of physicians in Rhode Island medical practice. Only closedplete electronic health record ended questions were used for use an electronic health record. system or have some mix of this study. Younger age, more patients with electronic and paper records. Medicaid insurance, hospital practice, Some of these systems reported as Data Analysis “mixed” can be rudimentary too, practice in larger groups, primary care because programs like FileMaker Survey response was calculated practice, and nonsurgical specialization Pro (FileMaker Inc, Santa Clara, using the American Association are associated with increased electronic Calif) can qualify as an electronic for Public Opinion Research calhealth record use. health record. As the first and culator V3.1.8 Responses were Most respondents think that electronic only state (to our knowledge) to downloaded into Microsoft Excel health record adoption has decreased systematically assess and publi(Microsoft Corp, Redmond, WA) job satisfaction. Experience varied cally report electronic health reand analyzed in STATA 13.0 cord adoption for every licensed (StataCorp LP, College Station, significantly by electronic health record physician providing patient care, Tex). Missing, “not applicable,” vendor. Rhode Island provides an opporand other responses were excluded Longer and more sophisticated use were tunity to identify characteristics from analysis, and “don’t know” associated with a more favorable opinion of electronic health records and responses were reported only of electronic health records. physicians that might enable betwhen pertinent. ter electronic health record inteSummary statistics were used gration in the future. to describe respondents to the In this study of all physicians currently practicing in 2013 survey. Respondents were asked how electronic Rhode Island, we sought to accomplish 3 goals: (1) describe health record adoption had affected 6 elements of medical adoption of electronic health records by Rhode Island practice: job satisfaction, billing, internal communication, physicians; (2) describe how electronic health record patient care, quality improvement, and workflow. Responses implementation has changed elements of physician practice, were extent of agreement with a statement of the pattern such as workflow, patient care, and job satisfaction; and (3) “electronic health record adoption improved Job Satisfacfind characteristics of physicians and their electronic health tion,” for example, selecting one of “strongly disagree,” record system associated with positive experiences. We “disagree,” “agree,” or “strongly agree.” hypothesized that older age, late electronic health record A composite measure of how advanced a physician’s adoption, and partial electronic health record adoption electronic health record was developed by assigning a would be associated with more negative physician point for every function that was used and summing them experiences. to yield a score out of 12: (1) drug interaction warnings at point of prescription; (2) prompts for patient-specific care at point of care; (3) reminder regarding indicated or overdue care; (4) patient demographics; (5) electronic list of MATERIALS AND METHODS medications; (6) electronic progress notes; (7) electronic Study Design problem lists; (8) clinical summaries for transfer of care or This study has a cross-sectional survey design. referral; (9) report of clinical quality measures; (10) transfer of laboratory test results electronically; (11) transfer of radiologic test results electronically; and (12) Study Setting and Population electronic prescribing. There were 122 electronic health record vendors in use by Rhode Island physicians in 2013, The Rhode Island Department of Health administered the so a variable was created including only electronic health Health Information Technology Survey electronically record used by 10 or more physicians, thereby allowing a over the course of 3 weeks in January and February of more manageable analysis. 2009 to 2013.7 The department used its licensure database Student t test was used for continuous variables, and to send letters and e-mails (when available) to all licensed chi-square test was used for categoric variables to physicians providing direct patient care. Participants who
1010.e23
The American Journal of Medicine, Vol 127, No 10, October 2014 training are less likely to have an electronic health record than other types of doctors (Figure 1). Over time, increasing numbers of physicians in Rhode Island have reported use of an electronic health record in their primary practice, and in 2013, 81% of physicians used an electronic health record (Figure 2). Most providers think that electronic health record introduction has not improved their job satisfaction (Figure 3). However, they also report that electronic health record introduction improved billing and quality improvement. Physicians are evenly divided on whether electronic health record introduction improved workflow, and slim majorities agree that electronic health record introduction improved internal communication and patient care. Physicians with electronic health record in place for longer are significantly more likely to report a positive effect of electronic health record introduction on all aspects of clinical practice examined (P < .001). Physicians who report more sophisticated electronic health record use (ie, use of electronic health record for more aspects of their practice, documentation, and prescription-writing) also were more likely to report a positive impact of electronic health record adoption on all aspects of clinical practice (P < .001). Older age is associated with a worse opinion of the effect of electronic health record introduction, because those who disagreed and strongly disagreed with statements
demonstrate differences between physicians with and without an electronic health record. Histograms and boxplots were generated to show physician opinion of the effect of electronic health record adoption. Analysis of variance and Kruskal-Wallis tests were used to test for differences between provider opinions of electronic health record adoption effect and the following characteristics: brand of electronic health record, age of provider, years of electronic health record use, and level of electronic health record technology. Statistical significance was taken to be P < .05.
RESULTS The survey response rate was 62% in 2013, with 2365 of 3799 physicians completing surveys. Most doctors work more than 20 hours per week, have Doctor of Medicine degrees, have their primary practice in an office, are primary care providers, and work in groups of more than 10 providers (Table 1). There are differences between respondents with and without an electronic health record. Respondents with an electronic health record are likely to have larger populations of patients with Medicaid insurance, work in hospitals, work in larger practices, and practice primary care. Physicians with surgical subspecialty and psychiatric
Table 1
Characteristics of Respondents Respondents Without EHS
Total Age (y) Hours of patient care per week: 20 h License type DO MD Percent of patients using Medicaid 0% 60% Primary practice is: Hospital Office Size of practice 10 clinicians Primary care physician Yes No
All Respondents
% (n) or Mean (SD)
Respondents With EHS
2365 50.9 (0.2)
19.2 (454) 56.6 (0.6)
80.8 (1911) 49.5 (0.2)
5.9 (136) 13.3 (308) 80.9 (1877)
5.5 (24) 14.1 (62) 80.5 (354)
6.0 (112) 13.1 (246) 81.0 (1521)
5.9 (139) 94.1 (2228)
5.1 (23) 94.9 (431)
6.1 (116) 93.9 (1795)
14.2 66.9 14.0 4.8
22.8 69.5 5.6 2.1
10.9 65.9 17.3 5.9
P Value
