Physician and practice characteristics associated with longitudinal increases in electronic health records adoption.
Subject: Medicaid (Forecasts and trends)
Medicare (Evaluation)
Electronic records (Usage)
Authors: Menachemi, Nir
Powers, Thomas L.
Brooks, Robert G.
Pub Date: 05/01/2011
Publication: Name: Journal of Healthcare Management Publisher: American College of Healthcare Executives Audience: Trade Format: Magazine/Journal Subject: Business; Health care industry Copyright: COPYRIGHT 2011 American College of Healthcare Executives ISSN: 1096-9012
Issue: Date: May-June, 2011 Source Volume: 56 Source Issue: 3
Topic: Event Code: 010 Forecasts, trends, outlooks Computer Subject: Market trend/market analysis
Product: Product Code: 9105213 Medicaid NAICS Code: 92312 Administration of Public Health Programs
Geographic: Geographic Scope: Florida Geographic Code: 1U5FL Florida
Accession Number: 271594420
Full Text: EXECUTIVE SUMMARY

This article identifies practice- and physician-related characteristics associated with the increased use of EHRs by physicians in outpatient practices. Two Florida surveys conducted in 2005 and 2008 on physician use of EHRs were examined to determine the practice and physician characteristics associated with increased EHR use over time. Based on multivariate analysis, several variables were found to influence increased EHR adoption. Practice variables included participation in a single-specialty practice and percentage of Medicare patients in the practice, but not percentage of Medicaid patients in the practice. Physician characteristics included younger physician age, but not specialty nor years practicing in the community. Factors associated with EHR adoption at any given point in time did not necessarily predict longitudinal increases in EHR adoption. These results are important for physicians to consider in their potential adoption of EHRs and should also be considered by policymakers interested in promoting increased use of EHRs by physicians.

INTRODUCTION

The use of electronic heath records (EHRs) has increased in recent years (Simon et al. 2007; Ford et al. 2009) and is expected to increase further with the focus and funding contained in the Health Information Technology for Economic and Clinical Health (HITECH) Act. The HITECH Act makes incentive payments available through Medicare and Medicaid to physicians and hospitals beginning in fiscal year 2011 for the adoption and meaningful use of EHRs. The physician incentive payments aim to increase EHR adoption, but the HITECH Act also includes penalties to practices that are not meaningfully using EHR technology beginning in 2015, with harsher penalties in 2016 and 2017 (Blumenthal 2009). The incentive program will likely have a positive effect on EHR adoption, but the extent of that effect is currently unknown.

Despite the increased focus on EHRs, adoption rates and utilization rates of EHR systems have historically remained low (DesRoches et al. 2008). Many authors have reported on the characteristics of physicians and their practices that are associated with EHR adoption. For example, practice size (Hing, Butt, and Woodwell 2007; DesRoches et al. 2008), practice payer mix (Menachemi et al. 2007; Abdolrasulnia et al. 2008), physician age (Menachemi and Brooks 2006), and physician specialty (DesRoches et al. 2008; Simon et al. 2008) have all been linked to EHR adoption. However, all these studies were cross-sectional; they were only able to report the characteristics of physicians and practices associated with EHR adoption at any given point in time. As time elapses, certain medical practice and physician characteristics may increase the likelihood of EHR adoption; however, no study to date has examined this issue from an individual and a policy perspective.

This article, for the first time, reports EHR adoption based on a comparable sample of physicians at two points in time. This research approach makes possible the examination of three dimensions of EHR adoption that have not been previously reported. First, a longitudinal examination of overall EHR adoption levels is made over time for a comparable group of physicians. Second, a longitudinal breakdown of EHR adoption levels is presented based on specific practice and physician characteristics. Third, an interaction analysis is conducted on the practice and physician characteristics that relate to the adoption of EHRs in conjunction with the passage of time. These three dimensions provide a detailed perspective on how EHR adoption has changed over time and show how adoption levels have changed and have been influenced by specific practice and physician characteristics. These results, in combination, provide a comprehensive perspective on the changing status of EHR adoption and may also suggest how these changes may occur in the future. The research makes a contribution to the literature by demonstrating that the factors associated with EHR adoption at a given point in time do not necessarily predict a longitudinal increase in EHR adoption. Rather, a select number of variables identified in this research are responsible for changes in EHR adoption over time.

BACKGROUND

EHR adoption is an increasingly important issue to understand and investigate. Understanding EHR adoption levels and the factors that relate to EHR adoption can aid in the further use of EHRs, which subsequently can improve quality of care and reduce costs. How these factors contribute to changes in EHR adoption over time is important given the time frames associated with the HITECH Act. The adoption level of EHRs has been previously reported in the literature (Gans et al. 2005; Shields et al. 2007; Swartz 2006) and indicates a relatively slow increase in EHR use to this point. A study of primary care physicians conducted in Florida in 2005 found that 21 percent of primary care physicians used EHR systems (Abdolrasulnia et al. 2008). A 2007 Massachusetts study found that 35 percent of practices surveyed had EHRs, and that this level had increased from 23 percent between the period 2005 to 2007 (Simon et al. 2009). Other estimates of EHR adoption also indicate that the adoption level for physician offices is at approximately 35 percent (Lewis 2010).

Research also exists related to the factors that are associated with EHR adoption (Kralewski et al. 2010). Physicians located in areas with higher physician concentration were more likely to adopt EHRs (Abdolrasulnia et al. 2008). Physicians with high proportions of Medicaid patients in their practices have been reported to be less likely to use an EHR system when compared with practices that have lower volumes of Medicaid patients (Menachemi et al. 2007). The same study found that physicians with higher levels of privately insured patients have greater EHR use than those with relatively low levels of privately insured patients. Early adopters of EHRs have graduated from medical school more recently, work in larger practices, have fewer outpatient visits per week, and are less likely to report financial pressures to adopt information technology (Simon et al. 2009).

A number of barriers have been reported for the adoption of EHRs. These barriers relate to practice and physician characteristics and include concerns that EHR systems will be difficult to implement. The difficulty of implementation may involve a reduction in practice revenue (Managed Care Outlook 2010). Patient privacy concerns have also been cited as a factor that may reduce EHR adoption (Angst and Agarwal 2009). Given the reported benefits and need for increased EHR adoption, understanding the levels of EHR adoption, how those levels relate to practice and physician characteristics, and how adoption is influenced by the passage of time is important.

CONCEPTUAL FRAMEWORK

Innovations are diffused in successive stages based on the timing of the adoption and the type of adopter (Bass 1969; Norton and Bass 1987; Rogers 1962). Innovations are adopted in phases categorized by the nature of the adopter: innovator, early adopter, early majority, and late majority (Rogers 1962). The adoption of innovation and the passage of time are interrelated and tied to the use of information, which can improve medical practice (Lehmann and Kim 2005). The process of adoption must be considered in relation to the influence of other individuals or organizations that may be considering a similar innovation. With the exception of innovators, adopters are influenced by imitation because of pressures of the social system and the number of earlier adopters (Bass 1969, 2004). The social system also may involve thresholds that are related to time of adoption (Valente 1996). These thresholds may exist at an individual physician level or can be related to a network of physicians that influence individual adoption decisions.

The diffusion of innovation involves groups of organizations that interact with each other (Midgley, Morrison, and Roberts 1992; Rogers 1976). Communications that influence adoption through third parties are as important as more direct links from adopters to potential adopters (Midgley, Morrison, and Roberts 1992). In the diffusion process, an innovation spreads via communication channels over time to members of a social system (Rogers 1983; Rogers et al. 2005; Ryan and Gross 1943). Communication of a new innovation is also influenced by hype, the upsurge of public attention, and rising expectations about the potential of an innovation (Ruef and Markard 2010).

The acceptance of a new innovation is also related to the cost advantage of that innovation. In the case of EHRs, the cost and practice advantages are key drivers of acceptance (Berner et al. 2006; Hillestad et al. 2005; Wang et al. 2003). The costs of a technical innovation decline as output increases, making the relationship between costs and adoption timing an inverse relationship (Bass 1980). Additional generations of the technology are related to additional levels and timing of adoption (Mahajan, Muller, and Bass 1995). Innovations do not exist in a vacuum, and other innovations may influence adoption (Mahajan, Muller, and Bass 1990). Innovation adoption also follows a geographic pattern: the number of adoptions is greater closest to the location of the innovation origin or initial adoption (Mahajan and Peterson 1978, 1979). For a medical practice innovation, geographic proximity may influence adoption.

For this study, innovation is represented by EHR adoption, which also represents the unit of analysis. The adopter is the physician, and the nature of the adopter relates to the characteristics of the physicians examined. The social system entails the influence of other similar individuals or organizations and is represented by practice characteristics influenced by the interactions and communications between organizations. The cost advantage of the innovation is the percentage of Medicare and Medicaid patients and the anticipated payment bonus associated with EHR usage. Finally, the timing of the adoption is the longitudinal nature of the research design and the time-based interaction effects identified.

RESEARCH OBJECTIVE AND METHODS

This project addresses three research questions:

1. What is the overall change in EHR adoption between 2005 and 2008?

2. What are the changes in EHR adoption between 2005 and 2008 for each category of practice and physician characteristic?

3. What practice and physician characteristics are associated with the largest relative gain in EHR adoption rates between 2005 and 2008?

The answers to these questions will bring insight into how various practice and physician characteristics influence EHR adoption over time.

To address the research questions, surveys were conducted that assessed physician adoption of health information technology (HIT) applications focusing on the use of EHRs. The surveys were conducted in 2005 and 2008 using licensure addresses on file with the Florida State Department of Health. Both surveys used identical methodologies and proportionately equal sampling strategies of physicians stratified by specialty type. In each year, surveys were mailed to potential participants with a cover letter explaining the purpose of the study. Nonrespondents were mailed an additional survey and cover letter further urging their participation after four weeks. The questionnaires targeted a large sample of physicians (for 2005, n = 14,921; for 2008, n = 7,003). In 2005, all primary care physicians and a 25 percent random stratified sample of clinical specialists in the state were targeted. In 2008, the same proportion of physicians was targeted, but with one-half the overall sample size (i.e., 50 percent of all primary care physicians and a 12.5 percent random stratified sample of clinical specialists). The study focused on physicians practicing in outpatient settings and thus excluded physicians practicing in hospitals (e.g., radiologists, pathologists, and emergency physicians). The overall response rate was 28.2 percent for the 2005 survey and 29.4 percent for the 2008 survey. A formal and previously published analysis failed to detect meaningful response bias in the sample (Menachemi et al. 2006). The two surveys were merged into a common data file for the subsequent analysis.

Standard descriptive statistics were used to screen the data for any anomalies and to address the first research question (change in EHR adoption from 2005 to 2008). To test for significant changes in the overall level of EHR adoption and changes by practice and physician characteristics between 2005 and 2008 (i.e., the second research question), chi-square tests were used. Variables examined included those suggested by the literature to be associated with EHR adoption among physicians (Audet et al. 2004; Menachemi and Brooks 2006; Simon, Rundall, and Shortell 2005). Finally, multivariate logistic regression was used to predict EHR adoption with each practice and physician characteristic variable and with year of adoption (2005 or 2008). The analysis was performed with the binary dependent variable (EHR adoption) regressed on practice size (solo practices, 2 to 9 physicians, 10 to 49 physicians, or 50 or greater physicians), practice specialty (single- and multi-specialty practices), physician age (measured using the number of years since graduating from medical school as a proxy), years practicing in the community, percent of Medicare and Medicaid patients, geographic location (rural or urban), and training type (primary care including family physicians, general internists, and general pediatricians vs. physician specialists). The model also included a year fixed-effect and an interaction term between the year variable (2005 or 2008) and each of the independent variables. This approach addressed research question 3 about the practice and physician characteristics associated with the largest relative gain in EHR adoption rates and allowed for the identification of variable categories that were associated with a significant increase (relative to other categories of that variable) in EHR adoption between 2005 and 2008 while controlling for other factors in the model. As an example, the model as specified could determine if primary care physicians relative to specialists had greater increases in EHR adoption (between 2005 and 2008) while controlling for all other factors in the model.

RESULTS

In 2005, of the 14,921 surveys distributed, 4,203 were returned; in 2008, of the 7,003 surveys distributed, 2,057 were returned. These results represented a combined response rate of 28.6 percent. A description of the 2005 and 2008 samples appears in Exhibit 1 and the demographic makeup of respondents differed somewhat between the 2005 and 2008 samples. For example, the 2008 respondents were more likely to be 61 years of age or greater (21.3 percent vs. 16.0 percent; p < 0.001). Moreover, primary care physicians made up a larger percentage of respondents in 2008 compared to 2005 (65.3 percent vs. 51.8 percent; p < 0.001). Respondents in 2008 were also more likely to be in solo practice (34.9 percent vs. 30.9 percent) and less likely to be in practices with 2 to 9 physicians (51.2 percent vs. 54.2 percent; p = 0.008). Despite these differences, no variation between 2008 and 2005 respondents was detected for mean years practicing in the community.

EHR Adoption Levels

The results for overall EHR adoption level and adoption level by practice and physician characteristics are seen in Exhibit 2. In 2005, the EHR adoption rate was 23.7 percent; in 2008, this rate increased to 35.1 percent (a 48.1 percent increase).

Adoption level by practice characteristics

A number of significant findings are revealed by examining practice characteristics and how they related to changes in EHR adoption. For example, physicians in solo practices had a significant increase of 77.5 percent in EHR adoption between 2005 (13.8 percent) and 2008 (24.5 percent; p < 0.001). Likewise, physicians in small practices (defined as 2 to 9 physicians), and those in medium practices (defined as 10 to 49 physicians) experienced significant increases in EHR adoption of 66.2 percent (p < 0.001) and 27.7 percent (p = 0.005), respectively. Physicians in large practices (defined as having 50 or more physicians) did not experience a significant increase in EHR adoption between 2005 (72.8 percent) and 2008 (82.9 percent; p = 0.093).

When considering practice type, single-specialty practices went from 17.8 percent adoption in 2005 to 30.3 percent adoption in 2008 (a 70.2 percent increase; p < 0.001). Physicians in multispecialty practices did not experience a significant increase in EHR adoption during the same time period (40.5 percent in 2005 vs. 42.3 percent in 2008; p = 0.595). Based on Medicare patient caseloads broken into quartiles for the analysis, different EHR adoption trends were noticed over time. Physicians with smaller Medicare patient caseloads were more likely to have larger gains in EHR adoption rates. Specifically, physicians with 0 to 10 percent Medicare patients had a 74.3 percent increase in EHR adoption (p < 0.001); those with 10.1 to 30 percent Medicare patients had an increase of 72.4 percent (p < 0.001); and those with 30.1 to 55 percent had a somewhat smaller increase of 49.0 percent (p < 0.001). Physicians with the highest proportion of their practice made up of Medicare patients (e.g., 55 to 100 percent Medicare patients) did not experience a significant increase in EHR adoption from 2005 to 2008.

With respect to Medicaid patient caseload broken into quartiles for analysis, different EHR adoption trends were noted. While those in the bottom quartile (e.g., 0 percent Medicaid patients) had a decrease in EHR adoption rate between 2005 and 2008 of 21.1 percent (p = 0.022); those in the top quartile (greater than 20 percent Medicaid patients) had an increase in EHR adoption (p = 0.013). Physicians in practices in the two middle quartiles did not show meaningful differences in their EHR adoption gains. Last, rural and urban practice locations had a significant increase in EHR adoption of 60.8 percent and 52.7 percent, respectively (Exhibit 2).

Adoption level by physician characteristics

Examining EHR adoption changes by physician characteristics also revealed several significant findings. Primary care physicians and physician specialists experienced similar, statistically significant increases in EHR adoption between 2005 and 2008. Primary care physicians went from 22.4 percent adoption in 2005 to 34.0 percent adoption in 2008 (a 51.8 percent increase). Similarly, specialty physicians went from 25.2 percent in 2005 to 37.3 percent in 2008 (a 48.0 percent increase). In terms of physician age, younger physicians usually, but not always, showed a greater increase in EHR adoption when compared to physicians in the older-age category. Physicians younger than 40 years old had an increased adoption from 27.6 percent to 47.9 percent (a 73.6 percent increase; p < 0.001). The percent increase was less for physicians in the 41 to 50 age group who moved from 25.9 percent in 2005 to 37.5 percent in 2008 (a 44.8 percent increase; p < 0.001). Physicians in the 51 to 60 age group increased EHR adoption from a rate of 21.2 percent in 2005 to a rate of 35.0 percent in 2008 (a 65.1 percent increase). Physicians in the oldest age group (61 years or older) moved from an EHR adoption rate of 16.5 percent in 2005 to 21.9 percent in 2008 (a 32.7 percent increase; p = 0.043).

When considering the length of time each physician had spent in their current community, the increase in EHR adoption was relatively uniform. Depending on quartile, categories of "years in the community" ranged from a 45.7 percent increase to a 58.2 percent increase (see Exhibit 2).

Multivariate Analysis

In the multivariate analysis that included interactions between year of EHR adoption and each of the practice and physician characteristics, several significant trends were observed (see Exhibit 3).

Practice characteristics results

Although each category of practice size experienced an increase in EHR adoption from 2005 to 2008 (Exhibit 2), no category of practice size (relative to any other practice size category) had a significantly greater increase in EHR adoption after controlling for all other factors in the model (Exhibit 3). However, the increase in EHR adoption between 2005 and 2008 among physicians in a single-specialty practice was significantly greater than those in a multi-specialty practice (adjusted odds ratio [AOR] = 1.612, p = 0.042). Between 2005 and 2008, an increase in Medicare patients as a proportion of one's practice was associated with a decreased odds of EHR adoption (AOR = 0.98, p = 0.026). However, an increase in Medicaid patients as a proportion of one's practice was not associated with a significant change in EHR adoption (AOR=0.99, p = 0.699). Relative to physician location, the adoption rate increase for rural physicians between 2005 and 2008 was not significantly different for physicians located in an urban environment (OR=0.821, p = 0.532). This suggests that during the period of study, rural and urban physicians adopted EHR systems at similar levels.

Physician characteristics results

The multivariate analysis indicated that the EHR adoption level change for primary care physicians between 2005 and 2008 was not significantly different than for specialist physicians (AOR = 0.990, p = 0.552). Relative to physicians in the oldest age category, those in all other age categories had significantly higher increases in EHR adoption after controlling for all other factors in the model. Specifically, compared to doctors 61 years or older, those younger than 40 (AOR = 2.38; p = 0.034) or 51 to 60 years old (AOR = 2.08; p = 0.018) had significantly larger increases in EHR adoption between 2005 and 2008. The results of the multivariate analysis also showed that between 2005 and 2008, each additional year that a physician spent in the community was not associated with changes in the odds of EHR adoption (AOR = 1.01, p = 0.242).

DISCUSSION

While previous research has reported the characteristics of physicians and their practices that are associated with EHR adoption in a cross-sectional framework, this research incorporated the passage of time into the relationship between practice and physician practice variables and EHR adoption. Numerous findings warrant further discussion.

The results show significant increases in EHR adoption levels and wide variations in EHR adoption level based on practice characteristics. Many of the practice characteristics traditionally associated with increased EHR adoption at any given point in time were not significant predictors of EHR adoption longitudinally. This suggests that physicians in practices with lagging EHR adoption rates may be catching up. For example, large practices compared to smaller practices, and multi-specialty practices compared to single-specialty practices, had much higher levels of EHR adoption in both time periods. However, large practices had a much smaller increase in EHR adoption during the 2005 to 2008 period. If smaller practices adopt EHRs faster than large practices do, differences in adoption levels based on practice size may be relatively small in the future. Likewise, in terms of practice type, single-specialty practices had lower levels of EHR adoption compared to multi-specialty practices, but a significantly higher rate of adoption. Specialty practices could have EHR adoption levels comparable to multi-specialty practices in the future if present trends continue.

The results for percent Medicare patients show a similar pattern. Practices with lower levels of Medicare patients had a lower level of EHR adoption but a significantly higher rate of increase in adoption. Between 2005 and 2008, a higher proportion of Medicare patients in a given practice was associated with a lower level of EHR adoption. The same was not true for practices with differing levels of Medicaid patient caseloads; no differences were noted in EHR adoption gains by Medicaid volume quartile. Changes in Medicaid patients as a proportion of one's medical practice did not correlate with EHR adoption over time. The findings related to Medicare and Medicaid are particularly important given that the HITECH Act plans to encourage physicians to adopt EHR through these governmental programs. Future research should determine whether the current incentives are able to influence the trends identified in the current study.

The influence of practice location on EHR adoption levels and change in adoption level over time was found to be quite evident in our study. Physicians in rural locations had lower levels of EHR adoption than their urban counterparts; however, despite their late start at adopting EHRs, rural physicians' adoption rate during the three-year study period did not differ from the urban rate. Thus, the reported disparity in EHR utilization levels between rural and urban physicians (Paslidis, Schlesier, and Collier 2008; Shields et al. 2007) may be reversing.

With respect to physician characteristics, the empirical results from this study also provided interesting findings. Significant increases in EHR adoption levels occurred across all three physician characteristics examined, and two of the three categories are drivers of increases in EHR adoption. Primary care physicians, when compared to all other physician specialties, had a slightly lower level of EHR adoption in both time periods, and a similar and significant increase in EHR adoption between 2005 to 2008. Primary care physicians and other physician specialties had comparable EHR adoption rate increases between 2005 and 2008. With a comparable rate of increase in adoption, and without any evidence to the contrary, the various types of physician specialties may retain similar EHR adoption levels in the future. The multivariate analysis showed no significant time interaction effects; physician specialty is not a defining characteristic associated with increased levels of EHR adoption.

The results show major differences by physician age. Younger physicians have higher adoption levels than older physicians, and the increase in EHR adoption between 2005 and 2008 is significantly greater for physicians 60 years or younger. The level of increase in adoption is not linear, however, for the three younger age groups. Interestingly, the 51 to 60 age group had a higher rate of increase than the 41 to 50 age group. The cause for this is not evident in the analyses, but may be a result of an interaction with other variables outside the domain of the present research. Regardless, the results for physician age indicate a major disparity in EHR adoption level that is becoming greater with the passage of time. Future research should identify whether adoption levels and rates of increase in adoption levels for older age groups become greater as younger physicians move into that age category, or if variables that pertain to that age group (e.g., time remaining until retirement) dictate adoption levels. As adoption levels should not decrease from the base level of an age group as they move into another category, the focus should be on adoption rate. All physician age groups with the exception of the 41 to 50 group had a significant time interaction effect. Thus, physician age (younger than 60 years) proved to be a driver of increased EHR adoption in most cases.

The variable measuring years practicing in the community provided results that can be compared and contrasted with physician age. Similar to physician age, a greater number of years in the community was associated with lower adoption levels in both time periods. Not surprisingly, the adoption levels seen in Exhibit 2 are comparable between these two characteristics as age and years in the community tend to be parallel. Increase in adoption level, however, was not similar for these two characteristics in multivariate analysis. Whereas age was associated with increases in EHR adoption over time, the number of years practicing the community was not. Perhaps as a physician develops a more established presence in the community, the perceived need for EHR adoption increases, but this variable was no longer associated with EHR adoption over time when controlling for other factors in the model.

Several limitations are worth mentioning. First, this study is limited by the self-reported nature of survey research. Responding physicians may not have known or may have chosen to misreport EHR adoption level. Second, despite the large sample size, the response rate to both surveys was suboptimal. Third, the study was confined to a single state; thus, generalizing the findings outside of Florida must be done with caution. Last, despite the identically proportional sampling strategies in 2005 and 2008, the characteristics of the two samples were different in some ways. These differences may have been the result of true shifts in physician practice characteristics or an artifact of sampling. Either way, these differences may affect the analysis presented.

CONCLUSION

As policymakers continue to attempt to stimulate EHR adoption among physicians, longitudinal assessments of progress will become increasingly important. The current study demonstrated that the factors associated with EHR adoption at any given point in history did not necessarily predict longitudinal increases in EHR adoption. This suggests that decision makers in medical practices are finding ways to overcome the historic barriers associated with EHR adoption. For example, while large practice size historically made it easier from a financial and human resource perspective to adopt EHRs, smaller practices have found ways to begin closing the adoption gap. The same is true for rural physicians and those in single-specialty practices. Consistent with the diffusion of innovation framework, the passage of time seems to help medical practices overcome the cost and logistic barriers associated with EHR adoption. Future studies should monitor the effects of the HITECH Act on longitudinal EHR adoption rates among physicians in various practice settings.

REFERENCES

Abdolrasulnia, M., N. Menachemi, R. Shewchuk, P. Ginter, W. Duncan, and R. Brooks. 2008. "Market Effects on Electronic Health Record Adoption by Physicians." Health Care Management Review 33 (3): 243.

Angst, C., and R. Agarwal. 2009. "Adoption of Electronic Health Records in the Presence of Privacy Concerns: The Elaboration Likelihood Model and Individual Persuasion." Management Information Systems Quarterly 33 (2): 339.

Audet, A.-M., M. M. Dory, J. Peugh, J. Shamasdin, K. Zapert, and S. Schoenbaum. 2004. "Information Technologies: When Will They Make It Into Physician's Black Bags?" Medscape General Medicine 6 (4): 2.

Bass, F. M. 2004. "Comments on 'A New Product Growth for Model Consumer Durables.'" Management Science 50 (12): 1833-40.

--. 1980. "The Relationship Between Diffusion Rates, Experience Curves, and Demand Elasticities for Consumer Durable Technological Innovations." Journal of Business 53 (3).

--. 1969. "A New Product Growth for Model Consumer Durables." Management Science 15 (5): 215-27.

Berner, E. S., T. K. Houston, M. N. Ray, J. J. Allison, G. R. Heudebert, W. W. Chatham, J. I. Kennedy Jr., G. L. Glandon, P. A. Norton, M. A. Crawford, and R. S. Maisiak. 2006. "Improving Ambulatory Prescribing Safety with a Handheld Decision Support System: A Randomized Controlled Trial." Journal of the American Medical Informatics Association 13 (2): 171-79.

Blumenthal, D. 2009. "Stimulating the Adoption of Health Information Technology." The New England Journal of Medicine 360 (15): 1477-79.

DesRoches, C. M., E. G. Campbell, S. R. Rao, K. Donelan, T. G. Ferris, A. Jha, R. Kaushal, D. E. Levy, S. Rosenbaum, A. E. Shields, and D. Blumenthal. 2008. "Electronic Health Records in Ambulatory Care--A National Survey of Physicians." The New England Journal of Medicine 359 (1): 50.

Ford, E., N. Menachemi, L. Peterson, and T. Huerta. 2009. "Resistance Is Futile: But It Is Slowing the Pace of EHR Adoption Nonetheless." Journal of the American Medical Informatics Association 16 (3): 274-81.

Gans, D., J. Kralewski, T. Hammons, and B. Dowd. 2005. "Medical Groups' Adoption of Electronic Health Records and Information Systems." Health Affairs 24 (5): 1323-33.

Hillestad, R., J. Bigelow, A. Bower, F. Girosi, R. Meili, R. Scoville, and R. Taylor. 2005. "Can Electronic Medical Record Systems Transform Health Care? Potential Health Benefits, Savings, and Costs." Health Affairs 24 (5): 1103-17.

Hing, E. S., C. W. Burt, and D. A. Woodwell. 2007. "Electronic Medical Record Use by Office-Based Physicians and Their Practices: United States, 2006." Advanced Data from Vital and Health Statistics 393: 1-7.

Kralewski, J., B. Dowd, T. Zink, and D. Cans. 2010. "Preparing Your Practice for the Adoption and Implementation of Electronic Health Records." Physician Executive 36 (2): 30-33.

Lehmann, C., and G. Kim. 2005. "Prevention of Medication Errors." Clinics in Perinatology 32 (1): 107-23.

Lewis, M. 2010. "EHR Adoption Up 3 Percent Since 2009, Study Finds." Medical Economics 87 (6): 13.

Mahajan, V., E. Muller, and F. M. Bass. 1995. "Diffusion of New Products: Empirical Generalizations and Managerial Uses." Marketing Science 14 (3): 79-88.

--. 1990. "New Product Diffusion Models in Marketing: A Review and Directions for Research." The Journal of Marketing 54 (1): 1-26.

Mahajan, V., and R. A. Peterson. 1979. "Integrating Time and Space in Technological Substitution Models." Technological Forecasting and Social Change 14 (3): 231-41.

--. 1978. "Innovation Diffusion in a Dynamic Potential Adopter Population." Management Science 24 (15): 1589-97.

Managed Care Outlook. 2010. "Despite Lingering Barriers to Adoption, Physicians See Value in EHRs, Look for Ways to Adopt." Managed Care Outlook 23 (7): 1, 3, 5-6.

Menachemi, N., and R. G. Brooks. 2006. "EHR and Other IT Adoption Among Physicians: Results of a Large-Scale Statewide Analysis." Journal of Healthcare Information Management 20 (3): 79-87.

Menachemi, N., N. Hikmet, M. Stutzman, and R. Brooks. 2006. "Investigating Response Bias in a Health Information Technology Survey of Physicians." Journal of Medical Systems 30 (4): 277-82.

Menachemi, N., M. C. Matthews, E. W. Ford, and R. G. Brooks. 2007. "The Influence of Payer Mix on Electronic Health Record Adoption by Physicians." Health Care Management Review 32 (2): 111-18.

Midgley, D.F., P.D. Morrison, and J.H. Roberts. 1992. "The Effect of Network Structure in Industrial Diffusion Processes." Research Policy 21 (6): 533-52.

Norton, J. A., and F. M. Bass. 1987. "A Diffusion Theory Model of Adoption and Substitution for Successive Generations of High-Technology Products." Management Science 33 (9): 1069-86.

Paslidis, N., S. Schlesier, and S. Collier. 2008. "Introducing EHRs to a Rural Practice." Physician Executive 34 (3): 30-33.

Rogers, E. M. 1983. Diffusion of Innovations, third edition. New York: The Free Press.

--. 1976. "New Product Adoption and Diffusion." Journal of Consumer Research 2 (4): 290-301.

--. 1962. Diffusion of Innovations, first edition. New York: The Free Press.

Rogers, E. M., U. E. Medina, M. A. Rivera, and C. J. Wiley. 2005. "Complex Adaptive Systems and the Diffusion of Innovations." The Innovation Journal: The Public Sector Innovation Journal 10 (3): 1-26.

Ruef, A., and J. Markard. 2010. "What Happens After a Hype? How Changing Expectations Affected Innovation Activities in the Case of Stationary Fuel Cells." Technology Analysis & Strategic Management 22 (3): 317-38.

Ryan, B., and C. N. Gross. 1943. "The Diffusion of Hybrid Seed Corn in Two Iowa Communities." Rural Sociology 8 (1): 15-24.

Shields, A. E., P. Shin, M. G. Leu, D. E. Levy, R. M. Betancourt, D. Hawkins, and M. Proser. 2007. "MARKETWATCH: Adoption of Health Information Technology in Community Health Centers: Results of a National Survey." Health Affairs 26 (5): 1373-83.

Simon, S. R., R. Kaushal, P. D. Cleary, C. A. Jenter, L. A. Volk, E. J. Orav, E. Burdick, E. G. Poon, and D. W. Bates. 2007. "Physicians and Electronic Health Records: A Statewide Survey." Archives of Internal Medicine 167 (5): 507.

Simon, S. R., M. L. McCarthy, R. Kaushal, C. A. Jenter, L. A. Volk, E. G. Poon, K. C. Yee, E. J. Orav, D. H. Williams, and D. W. Bates. 2008. "Electronic Health Records: Which Practices Have Them, and How Are Clinicians Using Them?" Journal of Evaluation in Clinical Practice 14 (1): 43-47.

Simon, J. S., T. G. Rundall, and S. M. Shortell. 2005. "Drivers of Medical Records Adoption Among Medical Groups." Joint Commission of Quality and Patient Safety 31 (11): 631-39.

Simon, S. R., C. S. Soran, R. Kaushal, C. A. Jenter, L. A. Volk, E. Burdick, P. D. Cleary, E. J. Orav, E. G. Pooh, and D. W. Bates. 2009. "Physicians' Use of Key Functions in Electronic Health Records from 2005 to 2007: A Statewide Survey." British Medical Journal 16 (4): 465.

Swartz, N. 2006. "EHR Adoption Moving Slowly but Surely." Information Management Journal 40 (4): 19.

Valente, T. W. 1996. "Social Network Thresholds in the Diffusion of Innovations." Social Networks 18 (1): 69-89.

Wang, S. J., B. Middleton, L. A. Prosser, C. G. Bardon, C. D. Spurr, P. J. Carchidi, A. F. Kittler, R. C. Goldszer, D. G. Fairchild, A. J. Sussman, G. J. Kuperman, and D. G. Bates. 2003. "A Cost-Benefit Analysis of Electronic Medical Records in Primary Care." The American Journal of Medicine 14 (5): 397-403.

PRACTITIONER APPLICATION

Vance M. Chunn, MSHA, FACHE, chief executive officer/administrator, Cardiology Associates, Mobile, Alabama

There is no question that adoption of electronic health records (EHRs) has increased longitudinally. Menachemi, Powers, and Brooks do an excellent job of presenting this development and evaluating the variables that have affected it. Interestingly, many of the documented observations ran along predictable lines, while several variables that were analyzed produced surprising findings.

As the authors explain, in the past adoption of EHRs was driven largely by practice and physician characteristics. Anticipated cost benefits and quality improvements also played a part. The incentives associated with the adoption of EHRs were largely internal to the practitioner or organization.

Where does adoption and use of EHRs go from here? Certainly the incentives created by the Health Information Technology for Economics and Clinical Health (HITECH) Act are an important new driver for adoption. Many physicians and organizations are either evaluating or have made the decision to invest to take advantage of the financial incentives available through HITECH. Our organization is one of them. After implementing a medical document management system and moving to an almost completely paperless workflow, we were content to remain there for the foreseeable future. However, the HITECH Act and associated incentives were considerable factors that compelled us to change course and select and implement a complete EHR system.

Reduced costs to the overall healthcare system, better communication and coordination of care, enhanced quality, and increased efficiency are among the possible benefits associated with the adoption and use of EHRs. However, costs to practitioners and their organizations, the necessary infrastructure, the fear of failure, and the compulsory review and potential changes in work flow present barriers to adoption. Other barriers include organizational leadership, culture, and potential nonrecoverable loss of productivity for high-volume organizations. In addition, certain specialties may not qualify for incentives because of limited volumes of Medicare and Medicaid patients.

Our organization happened to fit nicely within the parameters of offered incentives by HITECH, and this was influential in our proceeding with adoption. We project that our investment in hardware, software, and implementation will be covered by available incentives. However, this does not take away the facts that the commission and implementation of EHRs can be challenging and the change can have a deep organization-wide impact.

The current movement toward adoption is generally viewed as positive. However, it is important to point out that for the system, individual organizations, practitioners, and patients to obtain the overall benefits, significant global adoption of EHRs must occur. Without a large number of adopters working and communicating together, improvements in efficiency and quality cannot be achieved.

Are the incentives created through HITECH enough? Have policymakers implemented the necessary incentives to cause significant levels of adoption? This will be determined over time. I encourage physicians and their organizations to look closely at HITECH and all other potential benefits and to evaluate the effect of future potential penalties as they consider the possible adoption of an EHR. It will also be important for the government to study the rates and patterns of adoption to evaluate the effectiveness of HITECH.

The movement toward adoption and use of EHRs longitudinally seems poised to grow. The benefits for healthcare organizations, physicians, and patients are certainly noteworthy. I hope that the healthcare industry will continue to move forward with the use of this technology and obtain its associated benefits. This topic certainly merits further study.

Nir Menachemi, PhD, MPH, associate professor, University of Alabama at Birmingham; Thomas L. Powers, PhD, MBA, professor, University of Alabama at Birmingham; and Robert G. Brooks, MD, MBA, MPH, professor and associate vice president for health care leadership, University of South Florida College of Medicine, Tampa
EXHIBIT 1
Characteristics of Samples

                                 2005            2008
                             (n = 4,203)     (n = 2,057)    p-value
Physician Characteristics

Physician age
  Less than 40                485 (16.0%)     350 (17.2%)   < 0.001
  41-50 years               1,130 (37.3%)     612 (30.1%)
  51-60 years                 930 (30.7%)     639 (31.4%)
  61 years or greater         486 (16.0%)     434 (21.3%)

Physician training
  Primary care              2,141 (51.8%)   1,334 (65.3%)   < 0.001
  Specialists               1,995 (48.2%)     709 (34.7%)

Practice Characteristics

Practice size
  Solo practice             1,228 (30.9%)     697 (34.9%)   0.008
  2-9 physicians            2,150 (54.2%)   1,022 (51.2%)
  10-49 physicians            385 (9.7%)      196 (9.8%)
  50+ physicians              206 (5.2%)       82 (4.1%)

Practice type
  Single specialty          2,713 (85.6%)   1,405 (74.7%)   < 0.001
  Multi-specialty             457 (14.4%)     475 (25.3%)

Geographic location
  Urban                     3,950 (94.2%)   1,623 (91.7%)   < 0.001
  Rural                       245 (5.8%)      146 (8.3%)

Percent Medicare patients    32.2 (27.7)     37.6 (27.0)    < 0.001
  (s.d.)
Percent Medicaid patients    13.9 (20.4)     16.8 (22.3)    < 0.001
  (s.d.)
Mean years in community      14.7 (9.8)      15.1 (10.5)      0.145
  (s.d.)

EXHIBIT 2
Comparison of EHR Adoption Levels

                                EHR Adoption
                                  Rates (%)
                                                         Percent
                                 2005   2008   p-value   increase

Overall EHR Adoption Rate        23.7   35.1   < 0.001     +48.1

Practice Characteristics

Practice size
  Solo                           13.8   24.5   < 0.001     +77.5
  Small (2-9 physicians)         20.4   33.9   < 0.001     +66.2
  Medium (10-49 physicians)      45.2   57.7     0.005     +27.7
  Large (50+ physicians)         72.8   82.9     0.093     +13.9

Practice type
  Single specialty               17.8   30.3   < 0.001     +70.2
  Multi-specialty                40.5   42.3     0.595      +4.4

Percent Medicare patients
  0-10% (first quartile)         17.9   31.2   < 0.001     +74.3
  10.1-30% (second quartile)     22.1   38.1   < 0.001     +72.4
  30.1-55% (third quartile)      25.5   38.0   < 0.001     +49.0
  55.1-100% (fourth quartile)    26.0   29.8     0.152     +14.6

Percent Medicaid patients
  0% (first quartile)            38.3   30.2     0.022     -21.1
  1-5% (second quartile)         21.5   22.2     0.890      +3.3
  5.1-20% (third quartile)       21.0   22.2     0.772      +5.7
  20.1-100% (fourth quartile)    19.1   25.3     0.013     +32.5

Geographic location
  Rural                          17.6   28.3   < 0.001     +60.8
  Urban                          24.1   36.8     0.015     +52.7

Physician Characteristics

Physician specialty
  Primary care                   22.4   34.0   < 0.001     +51.8
  Other                          25.2   37.3   < 0.001     +48.0

Physician age
  Less than 40 years             27.6   47.9   < 0.001     +73.6
  41-50 years                    25.9   37.5   < 0.001     +44.8
  51-60 years                    21.2   35.0   < 0.001     +65.1
  61 or greater years            16.5   21.9     0.043     +32.7

Years in the community
  0-6 (first quartile)           29.3   42.7   < 0.001     +45.7
  6.1-13 (second quartile)       26.0   38.4   < 0.001     +47.7
  13.1-21 (third quartile)       22.7   33.1   < 0.001     +45.8
  Greater than 21 (fourth        16.5   26.1   < 0.001     +58.2
    quartile)

EXHIBIT 3
Multivariate Analysis of Practice and Physician Characteristics
and Year on EHR Adoption

                                                           Adjusted
Variable/Year Interactions                 B (Std Err)     Odds Ratio

Practice Characteristics

Practice size 1 (solo practice)             Reference       1.00

Year x practice size 2 (2-9 physicians)   -0.044 (0.193)    0.889
Year x practice size 3 (10-49             -0.180 (0.346)    0.835
  physicians)
Year x practice size 4 (50+ physicians)   -0.771 (0.571)    0.477
Year x single specialty practice           0.478 (0.235)    1.612 **
(Reference is multi-specialty practice)
Year x percent Medicare patients in       -0.007 (0.003)    0.98
  practice
Year x percent Medicaid patients in       -0.002 (0.005)    0.99
  practice
Year x rural location                     -0.198 (0.341)    0.821
  (reference is urban location)

Physician Characteristics

Year x primary care physician             -0.100 (0.169)    0.90
  (reference is specialist physician)
Year x age category 1(< 40 years old)      0.868 (0.410)    2.38
Year x age category 2 (41-50 years old)    0.585 (0.342)    1.45
Year x age category 3 (51-60 years old)    0.734 (0.311)    2.08
Age category 4 (61 or greater years)        Reference       1.00
Year x years in the community              0.022 (0.013)    1.01

Note: Logistic regression model with constant includes main effects
(not shown) and interaction effects (shown). Adjusted odds ratio
control for all variables included in model.

* p < 0.10

** p < 0.05

*** p < 0.01
Gale Copyright: Copyright 2011 Gale, Cengage Learning. All rights reserved.