|Home | About | Journals | Submit | Contact Us | Français|
Little is known about physicians' perception of the ease or difficulty of implementing electronic health records (EHR). This study identified factors related to the perceived difficulty of implementing EHR. 163 physicians completed surveys before and after the implementation of EHR in an externally funded pilot program in three Massachusetts communities. Ordinal hierarchical logistic regression was used to identify baseline factors that correlated with physicians' report of difficulty with EHR implementation. Compared with physicians with ownership stake in their practices, physician employees were less likely to describe EHR implementation as difficult (adjusted OR 0.5, 95% CI 0.3 to 1.0). Physicians who perceived their staff to be innovative were also less likely to view EHR implementation as difficult (adjusted OR 0.4, 95% CI 0.2 to 0.8). Physicians who own their practice may need more external support for EHR implementation than those who do not. Innovative clinical support staff may ease the EHR implementation process and contribute to its success.
Physicians who have not demonstrated ‘meaningful use’ of health information technology (HIT) by 2015 may face reduced future Medicare reimbursement.1 Failure to implement electronic health records (EHR) may prove costly to individual physicians and may make it hard for them to engage in care redesign that will be needed under healthcare reform.2 3 Awards up to US$44 000 through Medicare or even US$63 750 through Medicaid are currently available to physicians in office practice who demonstrate meaningful use of their EHR over 5 years.1 To receive the maximum reimbursement providers must begin participating by 2012.
Through the federal government, health information technology regional extension centers (REC) are positioned across the country to offer technical support and best practices to at least 100 000 providers over 2 years.4 In addition, approximately US$500 million has been awarded in federal support to states and territories to increase health information exchange.5 6
Earlier studies have identified factors that predispose practices to successful EHR adoption, namely larger practices, those affiliated with large hospitals, or those involved with teaching.7–10 Physicians face numerous barriers to participating in the meaningful use program, including a bewildering number of choices when selecting an EHR.11 It is unclear how fast providers will adopt. Ford et al,12 using EHR adoption data from six previous surveys, found under their most optimistic scenario a predicted EHR adoption rate of 61% by 2014, far short of ‘widespread adoption’.
In contrast to adoption, comparatively little is known about the factors that predict successful implementation, the installation of EHR and their incorporation into the usual operations of a functioning practice. Few data are available regarding what factors influence physicians' perception of the ease or difficulty of EHR implementation. This perception of ease or difficulty is important. It may influence the actual successfulness of EHR implementation or even the decision to complete an EHR implementation that is underway. Therefore, we undertook the present study, using data from practices in three communities participating in the Massachusetts eHealth Collaborative (MAeHC) EHR implementation project. Our objective was to identify factors present at baseline that were subsequently related to the perceived difficulty of implementing EHR.
We carried out a pre–post evaluation of the MAeHC three-community intervention to promote physicians' adoption and implementation of EHR. We collected physician attitudes, demographic and practice characteristics and used post-intervention surveys to measure physicians' perceived difficulty of EHR implementation. The Partners HealthCare Human Research Committee approved the study protocol.
The details of the MAeHC program have been described previously.13–15 In summary, between 2006 and 2008, MAeHC installed robust EHR in physician practices in three Massachusetts communities. MAeHC provided extensive on-site consultation, at no cost to these practices, to facilitate workflow redesign and integration of the EHR into the practice. They also provided technical support related to hardware and software.
The Massachusetts communities of Brockton, Newburyport and North Adams were selected for the MAeHC program through a competitive process described in detail in Goroll et al.13 Briefly, MAeHC issued a request for applications to which 35 communities throughout Massachusetts submitted applications to participate. MAeHC selection committee members made site visits to six finalist communities and ultimately selected the three participant communities. Selection criteria emphasized local physician leadership and community support of the application.
A total of 167 physician practices, representing 91% of all eligible primary care and specialty practices, participated. We identified all physicians in each of the participating practices for pre-intervention (2005) and post-intervention (2009) surveys. In the pre-intervention survey we identified 500 physicians from 167 practices; 355 completed the survey (response rate 77%).16 In 2009, 468 physicians were eligible, of whom 319 completed the survey (response rate 68%). For this study we included only the 163 physicians from 134 practice sites who completed both the 2005 and 2009 survey questionnaires.
The design and content of the pre and post-intervention surveys were based on similar statewide surveys of physicians described previously.16 17 Briefly, the 2005 survey included items intended to measure physicians' attitudes toward the use of computers in healthcare, perceptions of the quality of care, physician demographics and practice characteristics. The post-intervention (2009) survey retained verbatim all of the items from the original survey and incorporated new items that would not have been relevant before the intervention. For example, new questions asked about health information exchange, the EHR implementation process and workflow analysis.
Survey administration and data collection methods in 2005 and 2009 were similar. In 2005, paper surveys were hand-delivered to physicians' offices by MAeHC practice consultants, who had been making routine visits to the practices in preparation for the anticipated implementation. Physicians had the option of returning the survey directly to practice consultants or by mail.16 In 2009, the Survey Lab of the University of Chicago, a third-party research and consulting service, administered the survey to MAeHC physicians; these physicians returned the completed survey by mail or via the web to the Survey Lab in Chicago.
Our main outcome was the response to the following question in the post-intervention survey: ‘Was the implementation process for your EHR… (a) very difficult; (b) somewhat difficult; (c) not difficult?’
The pre-intervention survey included physician characteristics (age, sex, race, years in practice, number of outpatient visits per week); practice characteristics (number of physicians in practice, specialty, ownership, and financial resources available for expansion); and information on individual physician stressors (isolation from colleagues, working long hours, personal/job stress, and demoralization). Several dimensions of organizational culture were assessed, including innovation, initiative, quality improvement, evaluation, and risk assessment (see table 1). We requested information on the availability and physicians' use of 10 key functions of the EHR post-intervention to calculate an EHR usage score.9 As in previous analyses, we calculated an EHR usage score as the number of functions reported as used most or all of the time, divided by the number of available functions. For example, physicians who reported using none of their available EHR functions would score 0, while those who reported using all of their available functions most or all of the time would score 1. In another example, physicians who reported using four of their eight available functions most or all of the time would score 0.5.
To identify potential confounders between our main outcome and additional covariates we calculated χ2 tests or Fisher's exact test among categorical values and Wilcoxon rank-sum tests for continuous variables. Variables noted to be significantly associated (p<0.05) with the main outcome measure were included in the multivariate model. For multivariate analysis, we performed ordinal hierarchical logistic regression to account for the clustering of physicians within practice; the ordinal logistic regression assumption was met. We decided a priori to include sex and categorical age in the multivariate model. We validated use of the ordinal model using the score test for proportional odds assumption.
Overall, 163 physicians completed both surveys in 2005 and 2009. Among these physicians, 156 physicians (96%) answered questions relating to our main outcome. The physicians were mostly men, and most had full or partial ownership stake in their practice (see table 1). Most participants (70%) had been in practice for at least 15 years, with a median of 22 years since graduation from medical school. A total of 22 physicians (15%) reported having a ‘solo’ practice; 99 physicians (63%) indicated that they practised in a primary care or single specialty group practice and 35 (22%) worked in multispecialty group practices. A total of 26 physicians (17%) indicated that their practice had at least moderate resources for expansion or improvements of any kind; only two physicians (1%) stated that their practice had extensive resources for expansion. The mean EHR usage score was 0.68 (SD=0.28).
Overall, 54 physicians (35%) reported that the implementation process for their EHR was very difficult. Eighty-four (54%) of the physicians found the process to be somewhat difficult; only 18 (12%) reported their implementation process was not difficult.
Factors associated with difficulty of EHR implementation in bivariate analyses are shown in table 2. Compared with physicians who were partial or full owners of their practice, those without any ownership stake in their practice were less likely to view the implementation process as difficult or very difficult (OR 0.5, 95% CI 0.2 to 0.9). In terms of absolute numbers, 26% of non-owners (14/54) indicated that implementation was very difficult, compared with 38% of owners (40/104).
Physicians who indicated that the office staff was innovative were less likely to view the implementation process as difficult or very difficult (OR 0.4, 95% CI 0.2 to 0.8), compared with physicians who did not indicate that their staff was innovative. Higher EHR usage scores were associated with lower likelihood of viewing the implementation process as difficult or very difficult (OR 0.3, 95% CI 0.1 to 0.9).
In multivariate analysis, however, physicians' EHR usage score was not associated with the perceived difficulty of EHR implementation. Physicians who were employed by their practice (ie, those who did not have partial or full ownership stake in the practice) were less likely to view EHR implementation as difficult (adjusted OR 0.5, 95% CI 0.3 to 1.0). Similarly, physicians who perceived their staff as innovative were less likely to view EHR implementation as difficult (adjusted OR 0.4, 95% CI 0.2 to 0.8). We did not find evidence of effect modification by practice size on owners' perception of difficulty (p=0.69).
We evaluated how physicians practising in three communities in Massachusetts perceived the EHR implementation process and factors that facilitated EHR implementation. In this setting, we found that physicians who were not owners of the practice and those who perceived their office staff to be innovative were less likely to view the implementation process as difficult or very difficult.
Earlier studies have demonstrated that physicians working in larger practices and those affiliated with hospitals are more likely to have adopted EHR and to use those systems meaningfully.7 9 10 However, relatively little attention has been paid to the ease or difficulty physicians perceive when they implement EHR in their practices.
We found that physicians who were owners of their practice perceived their EHR implementation to be more difficult, compared with physicians who did not have an ownership stake. Because ownership is generally associated with greater levels of responsibility for day-to-day practice operations and management, these physicians probably experienced more underlying challenges associated with EHR implementation and workflow transformation. Physician owners probably bear financial risk for failure of implementation out of proportion to payers or publicly funded health plans who benefit from patient safety and quality but not as directly from practice efficiency or revenue cycle management.18 These physician owners may need not only financial support but also training and expert consultation to bolster the implementation process.
The association between physicians' perception of having innovative staff and their report of less difficult EHR implementation deserves exploration. Having staff who are innovative may reflect an office culture that is committed to progress or potentially a marker for a physician who values the role of non-physician staff in ensuring the smooth operations of the office. Understanding the role of ‘innovative staff’ may uncover a potential facilitator for easing EHR adoption and implementation.
This study has several limitations. Although the sample includes a broad range of primary care physicians and specialists from three diverse communities across Massachusetts, the findings may not generalize to practices in other settings. It should be noted that the physicians in this study were part of a well-supported EHR implementation program. While the HITECH-mandated REC program is supporting practices, it is likely that most practices nationally will receive less support from their REC than the physician groups in our study. We also note the limitation as with any survey-based research, that our variables are self-reported perceptions, rather than objective measures, and subject to a variety of biases, such as social desirability. Nonetheless, these perceptions are important, because they can influence physicians' behavior.
Physicians who have ownership stake in their practices experience first-hand the challenges of EHR implementation, and they generally perceive EHR implementation as being more difficult than their colleagues who are employed by the practices in which they work. Physicians who report that their staff is innovative perceive less difficulty with EHR implementation, suggesting that office staff play a critical role in EHR implementation, as they do in any transformation effort. Efforts to expand EHR implementation should focus on the needs of physician owners, and future efforts should emphasize the important role that non-physician staff members play in the process of innovation and transformation.
The authors would like to thank the physicians who participated in the MAeHC for completing the surveys. The authors also thank the anonymous peer reviewer who provided helpful suggestions to improve the manuscript.
Funding: This study was funded in part by the Agency for Healthcare Research and Quality cooperative agreement no 1UC1HS015397 and the Massachusetts e-Health Collaborative. MF was supported by a Ruth L. Kirschstein National Research Service Award grant no T32 HP12706.
Competing interests: None.
Ethics approval: The Partners HealthCare Human Research Committee approved the study protocol.
Provenance and peer review: Not commissioned; externally peer reviewed.