E-Prescribing Is Becoming a Standard of Care

AJPB® Translating Evidence-Based Research Into Value-Based Decisions®July/August 2015
Volume 7
Issue 4

E-prescribing adoption has increased dramatically over the past decade. This study provides empirical evidence that utilization among adopters continues to increase, suggesting application “stickiness.”

Providers with electronic prescribing (e-prescribing) systems—whether as stand-alone applications or as part of an electronic health record (EHR)—have the ability to electronically request and receive information about a patient’s medication history and plan eligibility, and to submit new prescriptions electronically to the patient’s preferred pharmacy.1 Provider use of e-prescribing has been demonstrated to improve use of on-formulary and generic prescribing, to decrease adverse drug events, and to be associated with increases in first-fill medication adherence.2-7

Growth in provider adoption of e-prescribing and EHR systems has risen significantly over the past several years, in large part because of federal incentive programs such as the Medicare Improvement for Patients and Providers Act and the Health Information Technology for Economic and Clinical Health Act (HITECH).8-11 Today, almost three-fourths of office-based providers and 95% of community pharmacies now actively use e-prescribing systems.12 Furthermore, and contrary to reported disparities in adoption of EHRs among urban and rural hospitals,13 the adoption of e-prescribing has been uniform across rural and urban care settings and practice sizes.14,15

There is mounting interest in and research around provider adoption of EHRs. However, substantially fewer data are publicly available, and fewer studies have been published regarding provider utilization of EHR functions once they have been adopted. Previous survey- and interview-based studies have shown that despite the benefits, prescribers continue to experience barriers that can inhibit use, including among others: 1) managing separate work flows for e-prescribing for controlled substances, 2) inconsistent pharmacy use of e-renewals, 3) variability between e-prescribing systems, and 4) application usability issues.16-19


The purpose of this study is to understand provider utilization of the e-prescribing functionality within an EHR through analysis of empirical transaction data provided by Surescripts, a national health information network. In particular, we sought to understand whether providers utilized or discontinued use of the e-prescribing functionality after the point of adoption, whether there were any trends or consistent patterns over time, whether providers using different e-prescribing and EHR systems had different patterns of use, and whether there were disparities or trends across specialties and geographic regions.

At a time when the majority of providers have adopted an EHR but are not yet using the majority of available functions, we believe the study of available utilization data can be instructive for EHR software vendors, clinic information technology leaders, consultants, and policy makers.


A longitudinal study was conducted using transactional data from the Surescripts network. E-prescriptions (eRxs) were evaluated for a 6-year period from January 2007 to December 2012. Data from all 50 states and the District of Columbia were included for this study. More than 450,000 office-based providers—including physicians, nurse practitioners, physician assistants, and other medical professionals with prescriptive authority—were evaluated. Surescripts estimates that this number represents roughly two-thirds of all office-based physicians and providers. The prescriber population varies over time. Providers’ length of time using e-prescribing systems on the Surescripts network were categorized into 4 subgroups based on length of time on the Surescripts network. Intervals of time were as follows: less than 1 year, 1 to 2 years, 2 to 3 years, and more than 3 years.

Prescription transactions included new prescription and renewal authorization responses sent electronically by providers as the key metric used to calculate provider utilization levels. Utilization was observed following the providers’ point of adoption and defined by the count of prescription messages sent electronically over the Surescripts network for the 6-year period. Providers who sent at least 1 eRx message across the Surescripts network were considered active prescribers and were included in the study. Monthly prescriptions were gathered for each provider on the Surescripts network who sent at least 1 eRx message during a specific month and were aggregated by time on the network. Data anomalies, such as prescribers with no associated specialty group, region, or EHR system, were removed from the analysis. A total of 24 providers exhibited extreme prescribing behavior, with more than 25 million total prescriptions over the study period, and were subsequently removed from the population. An outlier analysis found that these 24 prescribers accounted for more than 1% of the total number of prescriptions in the study. Although the prescriptions are authentic, the number of prescribers associated with these transactions is drastically underrepresented.

This analysis assessed 3 provider characteristics to describe utilization trends over time: e-prescribing system used when sending eRxs, geographic location of prescribers, and prescriber specialty. (System names were removed for the purpose of this analysis to preserve confidentiality.) Data were summarized and aggregated by month and year, over each time period. The outcome measure, provider utilization rate, was defined as:

[(sum (new prescriptions) + sum (refill responses)) / number of prescribers] (1)

in order to calculate the number of prescriptions sent electronically per provider. The primary explanatory variable of interest was time on the network, followed by the interaction of time and geographic region, provider specialty, and prescriber technology.

Data were stratified by provider specialty and ranked by overall percentage of active prescribers. The top 10 specialty groups were identified in the ranking step and analyzed by time on the Surescripts network; these 10 specialty groups were developed using 9 specialties and aggregating all other specialty groups into “other specialties.” A similar method was applied to evaluate utilization over time by e-prescribing technology, to include both EHR and stand-alone applications. Prescriber technologies were masked for confidentiality and presented here as System 1-System 10. The final measure assessed was provider geographic region. Utilization patterns were tracked over the study period and compared using the categories: west, northeast, south, and midwest.

Descriptive statistics were generated for each variable of interest. Univariate analysis indicated non-normal distributions of data. Kruskal-Wallis nonparametric analysis of variance (ANOVA) was performed to determine whether there were statistical differences in prescriber utilization over time by region, specialty, and EHR system technology. Post hoc multiple comparison methods were generated by applying Wilcoxon tests to pairwise data for all statistically significant results. (Pairwise comparisons were conducted using the Dwass-Steel-Critchlow-Fligner method [Hollander and Wolfe (1999)] or Bonferroni/Dunn’s correction.) Statistical significance was accepted at P <.05. Data analysis was performed using SAS version 9.3 (SAS Institute Inc, Cary, North Carolina).


Overall utilization increased as prescribers’ use of e-prescribing on the Surescripts network grew over time. Specifically, provider utilization increased (over the time of the study) from an overall monthly utilization rate of 93 eRxs per provider in December 2007 to 156 eRxs per provider by the end of the study in December 2012. Figure 1 displays median utilization rates per provider over the 4 time intervals. Active providers on the network for less than 1 year exhibited median utilization rates of 72 eRxs per provider. Rates increased to 112 eRxs per provider for the 1- to 2-year time period, 137 eRxs per provider for 2 to 3 years, and 178 eRxs per provider for more than 3 years. Differences between utilization rates across each time interval were statistically significant (P <.0001).

Provider utilization patterns were further illustrated by analyzing associations between median eRx-utilization rates over time by prescriber geographic location, provider specialty groups, and e-prescribing technology. All results indicated overall significance between utilization rates over time and geographic region, provider specialty, and technology. Each application displayed a similar pattern of increased utilization over time. Among the top 10 systems, system 1 and system 2 both experienced the greatest increase for more than 3 years. Figure 2 shows median utilization rates per provider by prescriber technology over time. Increases in utilization rates over time and across systems were statistically significant (P <.001).

Figure 3

displays provider utilization rate by 10 specialty groups over time on the network. Patterns of increased rates persisted over time across provider specialty. Results were statistically significant (P <.0001). Family practitioners and internists accounted for nearly 35% of all specialists in the data. Increases in utilization rates over the 4 time periods for family practitioners and internists were greater than differences between all other indicated groups.

Provider geographic location was the final measure of analysis. Provider locations were divided into 4 regions: midwest, northeast, south, and west, as shown in Figure 4. Each region displayed consistent increases over the 4 time periods (P <.0001). The south had the largest number of active prescribers on the Surescripts network with the highest volume of prescriptions, followed by the midwest, northeast, and west. (See the


, available at www.ajmc.com, for descriptive statistics.)


Our results suggest that provider utilization of e-prescribing increases following the point of adoption and that utilization continues to increase over time. We observed this behavior consistently across choice of EHR software vendor, across practice settings and provider specialty, and across geographic regions. The results strongly suggest that, once adopted, the use of EHR systems to e-prescribe is “sticky”—meaning that once prescribers enact the requisite change management process to adjust their clinical work flow, they are likely to continue to use and increase their use of e-prescribing. This is true despite the fact that, for the majority of the study period, e-prescribing for controlled substances was not legal; therefore, separate prescribing work flows were required for a portion of prescriptions.

Surescripts’ National Progress Report suggests that in 2013 a total of 1.04 billion prescriptions were routed electronically; based on available data from IMS Health regarding total prescriptions written, this suggests that more than 50% of eligible prescriptions written in 2013 were routed electronically.12,20 (Surescripts’ analysis of IMS Health data assumes that 11% of all dispensed prescriptions are for controlled substances and therefore were not legal in many states at the time of the study.) Combined with information about current levels of adoption, this information does provide a guideline as to how long we should expect average utilization to continue to increase.


Our findings have relevance to policy makers and the administrators of the HITECH program. The Office of the National Coordinator for Health Information Technology is responsible for setting thresholds for core measures of Meaningful Use. Historically, it has relied on input from provider organizations, hospital organizations, the vendor community, and others regarding how to set those measures, including the thresholds for e-prescribing. Our findings have several implications: 1) overall utilization of e-prescribing is quite high, with a simple calculation suggesting that the average e-prescriber routes roughly two-thirds of eligible prescriptions electronically today; 2) e-prescribing utilization behavior is dynamic and increases over time, suggesting that average utilization at any one time is not an appropriate reference point for setting Meaningful Use thresholds; and 3) utilization is likely to continue beyond HITECH and the Meaningful Use program given the high overall levels of utilization and the trend of increased use over time.


There are several limitations to this study. First, Surescripts does not have visibility to all eRxs. For example, eRxs routed between prescribers and pharmacies within a fully integrated delivery network will not show up on the Surescripts network. Despite this, we believe that Surescripts’ connectivity to more than 600 e-prescribing applications and 62,000 community and mail pharmacies allows it to serve as a proxy for utilization analysis at the national level. Second, utilization data were aggregated by year and month, provider specialty, region, and technology, therefore not allowing for analyses of individual prescribers and potential variability. Third, to evaluate utilization for active prescribers, only those providers who had activated e-prescribing on the Surescripts network and who had remained active over the period of the study were included. Future research on prescribers who activate but never use e-prescribing, or those who discontinue use, may be warranted. Fourth, controlled substances—estimated to account for 11% of all prescriptions&mdash;are significantly understated in this analysis because barriers to adoption and use of e-prescribing of controlled substances have minimized overall e-prescribing utilization rates. As regulatory and technological constraints to e-prescribing of controlled substances diminish over time, e-prescribing adoption and utilization trends are expected to rise.


We believe that e-prescribing can be viewed as a gateway technology that may lead to scaled interoperability. Specifically, e-prescribing is the first form of health information being exchanged ubiquitously, across geographic regions, practice settings, and prescriber specialties. We believe this is because the e-prescribing ecosystem is mature, with: 1) clear transaction standards, 2) an established governance forum for development of those standards in the form of the National Council for Prescription Drug Plans, and 3) early adoption among pharmacies, which allowed for positive network effects as prescribers adopted.

Our findings are promising and suggest that providers are making fundamental changes in the way they practice medicine and that those changes are sticking. Furthermore, the consistency of our findings across EHR systems, geographic regions, and practice settings suggests that providers are finding value in e-prescribing, irrespective of participation in or support from Regional Extension Centers or other programs. The findings are an encouraging sign of provider willingness to use EHRs to send and receive electronic clinical messages. Questions remain as to whether these findings can be extrapolated to suggest how other EHR functionality will be used over time, and if so, to what extent. We believe additional research is warranted to explore these questions and suggest that observation of transactional data from a health information exchange, rather than traditional survey-based methods, can be helpful in addressing these questions.


The authors would like to acknowledge the support provided by Jill M. Mulligan and the executive sponsorship provided by Mary Ann Chaffee.

Related Videos
Practice Pearl #1 Active Surveillance vs Treatment in Patients with NETs
© 2024 MJH Life Sciences

All rights reserved.