Using Challenges Associated Meaningful-Use Criteria to Prioritize Needed Changes in Electronic Health Records

Transcription

1 2016 International Symposium on Human Factors and Ergonomics in Health Care: Improving the Outcomes 55 Using Challenges Associated Meaningful-Use Criteria to Prioritize Needed Changes in Electronic Health Records G. Talley Holman, PhD, MBA American Academy of Family Physicians Leawood, Kansas, USA A. Joy Rivera, PhD Children's Hospital of Wisconsin Milwaukee, Wisconsin, USA Steven E. Waldren, MD, MS American Academy of Family Physicians Leawood, Kansas, USA Lawrence D. Dardick, MD UCLA Health Santa Monica Bay Physicians Los Angles, California, USA To date, more than 30 billion dollars has been spent on the meaningful use (MU) program. While high adoption has been realized, there has been little improvement in usability of Electronic Health Records (EHRs), prompting continuous discussion. Further, very little scientific data has been provided regarding the type of challenges present for physicians when using an EHR. This study evaluates and prioritizes 6 of the most common challenges associated with EHR use by physicians during a patient encounter using meaningful use criteria as a basis. Data represents 430 physicians from across the US. Findings show all EHRs have significant challenges with organization and structure of information and the amount or method of documenting information. Analysis of the 31 MU criteria show only 1 criterion did not have any significant challenges. Alternately, 75% of criteria had 3 or more significant challenges with 9 criteria being significantly impacted by all reported challenges. Finally, individual groupings of physicians were found to have more challenges than others when considering physician s age, computer comfort level, and the organizational type. Overall, findings show significant challenges exist for physicians when using EHRs, especially related to routine/basic patient care MU criteria. Specifically, challenges exist regarding how information is organized and the amount and/or method of EHR documentation, which must be a priority for improvement. If healthcare is to move forward and progress, EHRs must effectively support physicians in their work. Copyright 2016 Human Factors and Ergonomics Society. All rights reserved. DOI / INTRODUCTION In the past 10 years, health information technology (HIT) has forced its way into everyday healthcare. Across the US, adoption of electronic health records (EHRs) has exceeded 75% due in part to programs such as meaningful use (MU) (CMS, 2016a). However, the potential benefits of these technology innovations have not been fully realized with many groups, citing a negative impact on patient care (American EHR Partners, 2014; Bailey, 2011; Beasley et al., 2011; Friedberg et al., 2013; Krist et al., 2014; McCrory, 2015). Consequently, many of these same groups identify the requirements set by the MU program as one of the main issues (American EHR Partners, 2014; Bailey, 2011; Beasley et al., 2011; McCrory, 2015). Meaningful use first started with the intention of taking what was already being documented on paper and incentivizing the same information to be entered into an EHR in order to further EHR adoption (CMS, 2015). Hence, program architects were tasked with identifying items which were beneficial to patient care during a visit and the associated documentation to be mandated; therefore, the program was named meaningful use. Unfortunately, just because a criterion was identified, mandated to be recorded, and incentivized, does not mean the task fits into the physician s workflow in its new form, nor did it mean that the stored EHR information was easy to find or functionally usable (Beasley et al., 2011; Rosenbaum, 2015). Hence, the objective of this study was to examine challenges associated with basic functionality or operability of EHRs by the physicians during a patient encounter by using the meaningful use criteria as a basis to identify issues and prioritize them. METHODS In the summer of 2015, the American Academy of Family Physicians (AAFP) performed a study to evaluate the 31 MU criteria from stages 1 and 2 to understand their impact on physicians during a patient visit. A national panel of 13 physicians, engineers, and experts was formed of individuals who were all either

2 2016 International Symposium on Human Factors and Ergonomics in Health Care: Improving the Outcomes 56 researching primary care and/or maintained an active practice. To increase the diversity of experiences relative to variation of patients and practices, the panel represented all 5 US regions. Over a 4 month period, the panel developed generalizable tasks related to each MU criterion and discussed known challenges and issues in clinics. Focus was placed on the physician-patient interaction (face-to-face time), which is arguably the most valuable segment of a patient encounter. From this work, six of the most common challenges discussed were defined and included in a national survey study of physicians to identify whether these challenges were also common to them by criteria. Additionally, physicians could also give their opinion whether each criterion should in fact be included in meaningful use. To provide common context and limit misinterpretation, specific definitions were written for commonly used terms, see Table 1. Table 1. Common terms defined Common Term Patient Encounter Useful Compliance Burden Definition Events or tasks dedicated to preparing, executing, or documenting a patient visit Likely to improve assessment, diagnosis, treatment, patient satisfaction, or safety The minimum action, thought, or task to complete the objective Not just relative to time, it is mental or physical effort including interruptions A summary of the study s progression is as follows. Phase 1 of the study recruited physicians and engineers to form didactic groups by each region. Drawing on their experience and research, each group met to think and develop separately from the others. Phase 2 then brought the groups together as a national panel to create a consensus so information would be generalizable for Table 2. Comparison of MU study and national demographics most primary care physicians. Information was used to develop a survey construct in phase 3, which was refined through 10 revisions by in-person interviews, onsite focus groups, and online testing to control for structure/layout, length, intent clarity, specificity, wording bias, etc. Execution of the study in phase 4 was performed using AAFP resources. Potential participants were identified using the list of family physicians who had attested to (e.g. been paid for) MU, stage 1 or stage 2. A randomized participant list was generated with the goal of collecting a population level dataset, so that results and findings would be generalizable to the US and not to a specific sample or sub-group. Hence, it was estimated that a minimum of 375 returns would be needed where the demographics matched the national distribution related to primary care physicians. To improve the number of responses, 3 reminders were used (2 , 1 postal mailer). The last phase of the study, phase 5, was to evaluate and interpret results. Analysis was done using non-parametric statistics (kuskal-wallis, mann-whitney, RxC contingency tables). Results were assessed by the panel for interpretation and implications to primary care practice and/or policy. RESULTS Data collection occurred in the summer of In total, 480 surveys were returned from a randomized national pool of family physicians, who had attested to MU, with 430 being usable for analysis. Excluded responses were for incomplete data or because a physician took the survey more than once. Assessment of demographics showed correlation to national averages for gender, age, and regional distribution with only minor variation with the exception of physicians under the age of 40, which was under-reported, see Table 2. Age Distribution (%) Regional Distribution (%) AMA s US Physician s Report AAFP s Annual Census MU Study Respondents Report period (year) Male / female (%) Mean Age (yro) <39 yro yro yro >60 yro Southwest Mountain West Coast / / / Midwest Southeast Northeast

3 2016 International Symposium on Human Factors and Ergonomics in Health Care: Improving the Outcomes 57 Hence, consideration was given for evaluations performed, including and not limited to additional analysis to understand or rule out potential effects relative to this age group. Lastly, representation for specialized characteristics was found to be appropriate with 23.3% of practices seeing underserved patients, 49.1% being rural practices, 31.5% of practices being self-owned, 31.5% being physicians groups, and 37.7% being hospital systems, government, or other. For data analysis, criteria were categorized based on information use to allow for better understanding of findings. Accordingly, 15 criteria were associated with routine/basic patient care; 11 were associated with advanced specific or managed patient care, and 5 were associated with population care. The first analysis of the challenges evaluated prevalence relative to criteria, which provided many initial impressions. Generally, overall prevalence of each challenge was greater than expected, being ranked as follows: 1. Organization or structure of information 2. Amount or method of information documentation 3. Does not fit into workflow 4. Changes how I practice 5. Location in EHR 6. Getting Information from the patient However, when examining the relationship of these challenges to specific MU criteria, interpretation or context was not clear for every challenge. Specifically, getting information from the patient is still unresolved. Hence, findings related to that challenge will be omitted from these results until they can be explained properly. Analysis of the 31 criteria for the remaining 5 challenges showed only 1 criterion did not have any significant challenges (recording and chart vitals). Alternately, 75% of criteria had 3 or more significant challenges with 9 criteria being significantly impacted by all 5 challenges, see Table 3. Further analysis showed that there were not only a high prevalence of challenges generally among physicians, but there were also differences among sub-groups significantly impacted, which is indicated by the (+) in Table 3 and listed by sub-group in Table 4. The sub-groups identified show differences in how a given challenge or criteria affected one group more than another beyond any general problems or challenges affecting all physicians. For this, demographic information was used to stratify the challenge data, which shows that 7 criteria were significantly associated with 7 sub-groups of physicians. Of the 7 sub-groups, 4 were of interest as they constituted 75% of findings, see Table Organization type Classified if the physician was self-employed or worked in a physicians group, health/hospital systems or other. 2. Physician s age Identified age on 4 levels of under 40 years old, 40-49, 50-59, 60 years old or greater 3. Computer comfort level Allowed the physician to self-report on a 5 point scale how comfortable they were using a computer for uncomfortable to very comfortable. 4. EHR Vendor Identified which EHR a physician based on the 8 major companies DISCUSSION Data shows significant challenges/issues associated with all but one MU criteria. Prevalence reported among all physicians for the listed challenges in Table 4 far exceeds the levels needed to be significant. The specific challenge, how information is organized or structured in EHR, was consistently reported as the most problematic for physicians. However, when how information is organized and the amount and/or method of EHR documentation were both significant, the cumulative effect revealed the most challenging MU criteria for physicians. Criteria identified as not fitting into the workflow were typically associated with work not attributed to routine/basic patient care, which has a strong correlation to whether or not physicians believed the criteria should be in MU. Specifically, more than 50% of physicians state that all but one of the 15 routine/basic care criteria should be MU with one exception, patient visit clinical summaries. Only 2 of the 16 criteria not related to routine/basic care had a 50% or greater approval to be MU criteria. Further, when assessing only population care criteria, the number one challenge was that the criterion does not fit into the workflow, which is not surprising given that the objective of population care is not specific to an individual patient, but global.

4 2016 International Symposium on Human Factors and Ergonomics in Health Care: Improving the Outcomes Table 3. Prevalence of common challenges for MU criteria 58

5 2016 International Symposium on Human Factors and Ergonomics in Health Care: Improving the Outcomes 59 Table 4. Significant sub-groups associated with common challenges for MU criteria Location in EHR was found to be a prevalent challenge for 60% of the criteria and relates strongly to the overall design of EHRs. Next, prevalence of changes how I practice had strong association with many criteria, but it should be noted that the change could be positive or negative to the practice of the physicians, which is yet to be determined. Finally, when examining all findings related to prevalence, data shows the majority of physicians believe most routine/basic criteria for seeing a typical patient should be included in MU, but there are still significant challenges that need to be addressed. Further implications beyond MU relate to the basic design and operability of EHRs as it relates to the support of the needs and/or requirements of physicians during a typical patient encounter. Findings for sub-groups having challenges showed 7 MU criteria relative to 7 physician sub-groups. Of which, 4 of the criteria were related to routine/basic care, which were analyzed further. Three were found to be of interest. The 4 th, which is EHR vendor used, was not considered as relevant due to the improvement required for all EHRs. In short, the minor differences between vendors paled in comparison to the overall prevalence of challenges before the industry. As for the remaining 3 groups, recommendations can be made for innovators, designers, or implementers of EHRs and other tools based on the findings of significant differences in these sub-groups. 1. There are differences in the organizations use and requirements of a self-employed practice, a physicians group, and a hospital/healthcare organization clinic that does affect patient care 2. Age of physicians differentiates groups from each other based on how each learned and has experienced medicine, which affect perspectives and ability to change methods easily 3. Comfort level with a computer can affect adoption and use of an EHR in rate of adoption, methods utilized, rate of change, comprehension & correlation of benefit, acceptance, etc. Data provided here is a limited snapshot of some of the significant challenges that physicians deal with daily when using an EHR as it relates the MU program, which

6 2016 International Symposium on Human Factors and Ergonomics in Health Care: Improving the Outcomes 60 governs documentation in patient encounters. These results should allow for individuals to gauge the scope of each challenge physicians face, further the cause for improvement and change, and provide a basis for prioritization. These data also show that EHRs have not adequately supported the most basic functions/needs of physicians when seeing a patient, which breaks the first rule of design in not supporting the foundation of the work. Hence, a good start to addressing EHR usability is with criteria that deal with routine/basic care. Findings show 5 of the 11 most challenging MU criteria are routine basic with the primary issue being how information is organized (e.g. access and usability) and the amount and/or method of EHR documentation (e.g. how and how much information is recorded), which provides a clear direction for improvement. If EHR design is expected to improve in the coming years, information such as this must be considered, so shortcomings can be addressed such that physicians work is supported, not hindered or overly burdened. CONCLUSION To date, more than 30 billion dollars has been spent on the meaningful use (MU) program (CMS, 2016b). EHR designers are using programs such as MU to guide their system s functionality, but this generally means the function is present, with little consideration of whether it is easy to use and useful to the physicians (Beasley et al., 2011; Friedberg et al., 2013; Rosenbaum, 2015; Krist et al., 2014). This prompts much discussion regarding the usability of these tools. However, very little scientific data has been provided on a national scale regarding the type of challenges present when using an EHR nor the corresponding prevalence of those challenges. In fact, one result of this work is realization that not much is known about what practicing physicians deal with on a daily basis, nor what specifically impacts them and how to alleviate it. Thus, the findings presented here for EHRs and MU is an opportunity to pause to understand the problems. More information is needed if healthcare is to move forward and progress. Alternately, to simply move forward with another stage of the MU program, or a successor program using these criteria, without better understanding and addressing the challenge, issues, and problems, is not progress. That being said, each MU criterion can provide a benefit and has a place in healthcare regardless of whether it is in the MU program. What needs to be determined is, what should physicians be responsible for based on their daily work and objectives. Acknowledgments Funding for this study was provided by the AAFP. Special thanks to John Beasley, MD, Chester Fox, MD, Jenna Marquard, PhD, Ryan Mullins, MD, Charles North, MD, Matt Rafalski, MD, Heidy Roberson- Cooper, Debora Cohen, PhD, and Tosha Wetterneck, MD, for serving on the MU panel and providing feedback for this proceeding. References American EHR Partners. (2014). Physicians Use of EHR Systems Retrieved October 20, 2015, from Website: Use-of-EHR-Systems-2014.aspx Bailey J.E. (2011). Does health information technology dehumanize healthcare? AMA J of Ethics, 13(3), Beasley, J.W., Wetterneck, T.B., Temte, J., Lapin, L.A., Smith, P., Rivera-Rodriguez, A.J., & Karsh, B-T. (2011). Information chaos in primary care: implications for physician performance and patient safety. J Am B Fam Med. 2011, 24(6), Friedberg, M.W., Chen, P.G., Aunon, F.M., van Busum, K.R., Pham, C., Caloyeras, J.P., Mattke, S., Quigley, D.D., & Brook, R.H. (2013). Factors Affecting Physician Professional Satisfaction and Their Implications for Patient Care, Health Systems, and Health Policy. Rand Corporation. Krist, A.H., Beasley, J.W., Crosson, J.C., Kibbe, D.C., Klinkman, M.S., Lehmann, C.U., Fox, C.H., Mitchell, J.M., Mold, J.W., Pace, W.D., Perterson, K.A., Phillips, R.L., Post, R., Puro, J., Raddock, M., Simkus, R., & Waldren, S.E. (2014). Electronic health record functionality needed to better support primary care. J Am Med Inform Assoc, 21(5), McCrory, K. (2015, April 11). Meaningful use doesn t benefit my patients. Physicians Practice. Retrieved November 4, 2015, from Rosenbaum, L. (2015). Transitional chaos or enduring harm? The EHR and the disruption of medicine. N Engl J Med, 373(17), US Centers for Medicare and Medicaid s (CMS). (2015). How to attain meaningful use. Retrieved November 4, 2015, from CMS Website: US Centers for Medicare and Medicaid s (CMS). (2016a, April). ONC Data brief: Adoption of electronic health records systems among US non-federal acute care hospitals. Retrieved March 14, 2016, from CMS Website: US Centers for Medicare and Medicaid s (CMS). (2016b, January). State breakdown of payments. Retrieved March 14, 2016, from CMS Website: Guidance/Legislation/EHRIncentivePrograms/DataAndReport s.html