Physicians in ACOs Report Greater Documentation Burden

ABSTRACT

Objectives: First, to analyze the relationship between value-based payment (VBP) program participation and documentation burden among office-based physicians. Second, to analyze the relationship between specific VBP programs (eg, accountable care organizations [ACOs]) and documentation burden.

Study Design: Retrospective analyses of US office-based physicians in 2019 and 2021.

Methods: We used cross-sectional data from the National Electronic Health Records Survey to measure VBP program participation and our outcomes of reported electronic health record (EHR) documentation burden. We used ordinary least squares regression models adjusting for physician and practice characteristics to estimate the relationship between participation in any VBP program and EHR burden outcomes. We also estimated the relationship between participation in 6 distinct VBP programs and our outcomes to decompose the aggregate relationship into program-specific estimates.

Results: In adjusted analyses, participation in any VBP program was associated with 10.5% greater probability of reporting more than 1 hour per day of after-hours documentation time (P = .01), which corresponded to an estimated additional 11 minutes per day (P = .03). Program-specific estimates illustrated that ACO participation drove the aggregate relationship, with ACO participants reporting greater after-hours documentation time (18 additional minutes per day; P < .001), more difficulty documenting (30.6% more likely; P < .001), and more inappropriateness of time spent documenting (21.7% more likely; P < .001).

Conclusions: Office-based physicians participating in ACOs report greater documentation burden across several measures; the same is not true for other VBP programs. Although many ACOs relax documentation requirements for reimbursement, documentation for quality reporting and risk adjustment may lead to a net increase in burden, especially for physicians exposed to numerous programs and payers.

Am J Manag Care. 2024;30(Spec Issue No. 6):SP452-SP458. https://doi.org/10.37765/ajmc.2024.89552

Takeaway Points

We analyzed whether value-based payment programs such as accountable care organizations (ACOs) help reduce physicians’ documentation burden.

• Office-based physicians participating in ACOs reported greater difficulty documenting clinical care in the electronic health record, excessive time spent documenting, and greater after-hours documentation than nonparticipants.
• Analyses of value-based payment programs that combine programs into a singular “participation” variable may mask ACO-specific relationships with documentation burden outcomes.
• Given the multifaceted and complex relationship between payment models and documentation burden, more evidence is needed to tease apart the specific contributors to documentation burden in different payment environments.

Excessive physician documentation burden has been blamed in large part on documentation requirements of fee-for-service (FFS) billing models.¹ Value-based payment (VBP) models such as accountable care organizations (ACOs) may help decrease documentation burden via their fixed per-patient payments, reducing the need to document for reimbursement justification.² On the other hand, it is possible that the need to be responsive to a mix of reimbursement models exacerbates burden via switching costs as physicians navigate the requirements of multiple programs.^3,4 Furthermore, VBP models may introduce additional requirements to document information in structured ways for quality reporting.⁵ Reducing physician documentation burden remains a major national policy priority. Currently, parallel efforts include those led by the American Medical Association in coordination with CMS to reduce documentation requirements for evaluation and management (E/M) billing,⁶ the American Medical Informatics Association 25x5 Documentation Burden Reduction initiative to optimize electronic health records (EHRs),⁷ and the National Academy of Medicine’s National Plan for Health Workforce Well-Being, which includes goals aiming to reduce the time clinicians spend documenting.⁸

Despite the high priority put on reducing documentation burden writ large and frequent concerns over the role of FFS billing,¹ few empirical studies have focused on the specific role that different payment programs may play in alleviating or contributing to physician documentation burden. If it is indeed the case that onerous requirements for FFS reimbursement are at the root of excessive US physician documentation burden,^9,10 one would hypothesize that physicians who are relatively less exposed to pure FFS billing—including and especially those in concierge, direct-care, or “closed system” practices (eg, Kaiser Permanente)—may experience less burden. Although several studies have examined composite measures of participation in any VBP program or specific incentive programs such as the Merit-based Incentive Payment System (MIPS),^11,12 no studies to our knowledge have directly compared payment environments and specific VBP programs with respect to their independent relationship with physician documentation. This leaves little empirical evidence regarding the details of what programs and potential mechanisms from the numerous VBP systems in place relate directly to documentation burden. We aimed to expand this evidence base by examining 2 research questions. First, what is the relationship between participation in any VBP model and reported documentation burden? And second, how does this relationship vary across specific VBP models (eg, ACOs vs pay for performance)?

METHODS

Setting and Data Sources

We descriptively analyzed responses to the 2019 and 2021 National Electronic Health Records Survey (NEHRS), a nationally representative cross-sectional survey of US office-based physicians. The weighted response rate for the 2019 survey was 39.0% (n = 1372; 41.0% unweighted response rate), and it was 45.9% in 2021 (n = 1694; 46.6% unweighted response rate).

Measures: Reported EHR Burden

We analyzed 4 outcome measures capturing 3 constructs of documentation burden among physicians who reported using an EHR: appropriateness of time spent documenting, difficulty or ease in documenting, and time spent documenting after clinic hours. Appropriateness of documentation time was measured via responses to the prompt: “The amount of time I spend documenting clinical care is appropriate,” with response options on a 4-point Likert scale from strongly agree to strongly disagree. We constructed a binary measure for this with the 2 agree options coded as “agree” and the 2 disagree options coded as “disagree.” Difficulty or ease in documenting was measured similarly, via responses to the prompt: “How easy or difficult is it to document clinical care using your medical record system?” with response options of very easy, somewhat easy, somewhat difficult, and very difficult. This measure was also coded into a binary variable of “easy” or “difficult.” Finally, time spent documenting after clinic hours was assessed via the prompt: “On average, how many hours per day do you spend outside normal office hours documenting clinical care in your medical record system?” with response options of none, less than 1 hour, 1 to 2 hours, 2 to 4 hours, and more than 4 hours. We recoded this measure of documentation time 2 different ways: first, dichotomizing those responses into “1 hour or more” and “less than 1 hour,” and second, translating reported time into numeric minutes according to the methodology applied in Gaffney et al.¹¹ This approach converts survey responses using the midpoint of each option (0 minutes, 30 minutes, 90 minutes, and 180 minutes), excluding the highest value, which gets converted to 240 minutes (4 hours). There is evidence that time spent working in the EHR outside normal office hours correlates with provider burden,¹³ but this time can also align with some providers’ preferences for more flexibility in work practices.¹⁴ As a result, we interpret this variable simply as “documentation time outside clinic hours,” rather than assuming that all outside-of-clinic time is definitionally burdensome.

Measures: VBP Program Participation

The NEHRS 2019 and 2021 surveys asked respondents to indicate participation in the following 6 activities and VBP programs: patient-centered medical homes, ACOs, any “pay-for-performance arrangement,” the Medicaid EHR Incentive Program (eg, Meaningful Use or Promoting Interoperability), MIPS, and advanced alternative payment models. We preserved these program participation variables individually as well as computing a binary measure for any participation and a count of the total number of programs in which a respondent indicated participation. Program participation was not mutually exclusive.

Analysis

First, we computed descriptive statistics of our sample and outcomes stratified by both VBP program participation and practice ownership. To address our first research question, we ran multivariable ordinary least squares linear probability models for each of our 4 outcome measures, regressing each outcome on a binary indicator of any participation in VBP and covariates noted below. For our second research question, we exchanged the binary participation variable with a set of 6 binary indicator variables for each individual VBP program option to estimate the relationships between participation in a specific program and documentation burden. All regression models adjusted for physician practice ownership, practice size, whether the practice accepted Medicare and/or Medicaid, the practice EHR vendor, the presence or absence of staff support for documentation, and physician age, sex, and specialty. We also conducted a stratified analysis including only primary care providers. All analyses used survey weights (inflation factors) provided by NEHRS to derive nationally representative estimates,¹⁵ and we used heteroskedasticity-robust SEs clustered at the physician level and a P value less than .05 as the threshold for significance. All analyses were conducted in R statistical software 3.6.3 (R Foundation for Statistical Computing) using the tidyverse, survey, and fixest packages.

RESULTS

Our analytic sample of physicians who reported using EHRs represented 626,598 US office-based physicians (Table 1 [part A and part B]). Within this sample, 64.0% (weighted n = 400,847) reported participation in any of the 6 VBP programs in either survey year. Within this program participation group, the most common reported program was Meaningful Use (56.2%), followed by ACOs (47.9%) and MIPS (36.8%). Additional descriptive statistics of the sample are presented in Table 1.

Overall, 60.0% of office-based physicians reported that time spent documenting in the EHR for clinical care was inappropriate, and 36.0% reported that it was either very or somewhat difficult to document clinical care in the EHR. A total of 76.1% of physicians reported spending an average of more than 1 hour per day documenting in the EHR outside clinic hours, which equated to a mean (SD) of 110.8 (69.6) minutes spent documenting outside clinic hours on an average day.

VBP Program Participation and EHR Burden

Physicians participating in any VBP program indicated worse outcomes on 3 of 4 EHR burden measures compared with nonparticipants. A total of 63% of VBP participants (vs 55% of nonparticipants; P = .007) indicated that time spent documenting for clinical care was inappropriate, and 37% (vs 35%; P = .498) reported that documenting clinical care in the EHR was difficult. Similarly, 79% of VBP participants (vs 71% of nonparticipants; P = .002) reported spending more than 1 hour per day documenting outside clinic hours, translating to approximately 15 minutes more after-hours documentation (116 minutes per day vs 101 minutes per day; P < .001). Notably, VBP participants and nonparticipants did not differ in the provision of staff support for documentation (34.4% vs 31.8%; P = .379).

In descriptive analyses, ACO participants reported the highest rates of EHR burden across all 3 dichotomous outcome measures and office-based physicians reporting no program participation reported the lowest rates (Figure). Among ACO participants, 42% indicated that it was very or somewhat difficult to document clinical care in the EHR compared with 34% of respondents participating in no programs. A total of 84% of ACO participants reported spending more than 1 hour on documentation outside clinic hours compared with 70% of respondents participating in no programs.

In adjusted analyses (Table 2), participation in any VBP program was associated with a 10.5% greater likelihood of reporting more than 1 hour per day spent documenting outside clinic hours (β = 0.08; SE = 0.03; P = .01), and accordingly 11 more minutes per day spent documenting outside clinic hours on average (β = 11.02; SE = 4.95; P = .03). General VBP program participation was not associated with reporting inappropriate time spent on documentation (β = 0.06; SE = 0.04; P = .07) or difficulty of documentation (β = 0.01; SE = 0.03; P = .87).

In our models estimating the relationships between each program and our outcomes, ACO participation was the only program associated with worse outcomes on all measures. ACO participants were 21.7% more likely to report that time spent documenting was inappropriate (β = 0.13; SE = 0.03; P < .001) and 30.6% more likely to report that documenting clinical care was difficult (β = 0.11; SE = 0.03; P < .001). Accordingly, ACO participation was associated with an 11.8% higher likelihood of reporting more than 1 hour of after-hours documentation per day (β = 0.09; SE = 0.03; P = .001). This translated to an additional 18 minutes of daily after-hours documentation on average, a relative difference of 16.3% (β = 18.03; SE = 4.76; P < .001). No other VBP program demonstrated any significant relationship with EHR burden outcomes. In stratified analyses of primary care providers, we found analogous results with respect to the relationship between ACO participation and our outcomes. For primary care providers, participation in an advanced alternative payment model was associated with less time spent documenting outside clinic hours in both the binary and continuous formulations of the outcome (eAppendix Table 1 [eAppendix available at ajmc.com]).

DISCUSSION

In this nationally representative study of US office-based physicians’ NEHRS responses from 2019 and 2021, we found that physicians participating in VBP programs reported greater time spent documenting outside office hours. More specifically, our analysis illustrates that this relationship is driven entirely by office-based physicians participating in ACOs. Furthermore, relative to nonparticipants, ACO participants reported significant EHR burden in additional domains, namely inappropriateness of time spent documenting clinical care in the EHR and difficulty of documenting in the EHR. We expect these findings to be pertinent to commercial and public payers as well as federal policy makers aiming to reform payment policy to reduce burden.¹⁰

Clinical documentation chiefly supports clinical reasoning, care provision, and interprovider communication, but these are not the only ends to which documentation is a means.^16-18 The multiple roles of documentation influence how innovative delivery models intersect with documentation burden. In the case of ACOs under strict capitated payments, one might assume that the lack of reimbursement-related documentation needs would alleviate substantial portions of documentation burden.^5,19 Our results clearly illustrate that that is not the case for office-based physicians. We surmise that this relationship is due at least in part to the role that clinical documentation serves specifically in the ACO context to support both quality measure reporting and risk adjustment.¹⁹ Both of these aspects of ACO administration directly impact the revenue that a given practice receives under capitated payment models. Because clinical documentation is required for both, it is plausible that ACOs do not appreciably reduce documentation burden relative to FFS reimbursement models; it is even possible that exposure to a mix of payment models that includes ACOs (and other models) exacerbates physician burden. It is also possible that because many ACOs still utilize an FFS chassis,²⁰ physicians in those ACO arrangements get the worst of both worlds: increased documentation for ACO quality goals and no reprieve from reimbursement-based documentation requirements. Moreover, physicians exposed to a range of requirements across payment models likely default to the most comprehensive documentation standards to prioritize consistency and minimize cognitive switching costs, relinquishing opportunities to reduce documentation burden that may be available but incur high switching costs.

Practice and organizational leaders seeking to minimize the documentation burden imposed by VBP programs should consider first assessing their providers’ exposure to different VBP requirements. With this understanding of payment model mix, assessing the common documentation and administrative requirements that crosscut all models may inform strategic design of EHR tools, such as intake forms or shortcuts for templated note text, to simplify the documentation of those common data elements and minimize manual or duplicative documentation.²¹ Furthermore, our regression estimates suggest that the use of scribes for documentation support is associated with more appropriate—and less difficulty with—documentation (eAppendix Table 2), which is consistent with prior findings on the use of scribes in ambulatory care.²² Although scribes represent an additive solution rather than the subtractive solutions of simplifying the underlying billing requirements, adding scribes is one inexpensive and effective step that organizations can take to reduce documentation burden in the short to medium term.

The causal roots underlying US physician documentation burden remain complex and difficult to unearth. Reimbursement requirements are the most frequently identified culprit,^1,23 but policies directly aimed at reducing documentation requirements for reimbursement have met with tepid success.^6,24 The use of higher-reimbursement E/M codes has increased since the 2021 E/M guideline revision,²⁵ but clinical notes on average have not become shorter.²³ Our findings further complicate the documentation-reimbursement relationship, in that ACOs may increase documentation burden even though many ACO arrangements free physicians from the documentation-for-reimbursement model. Medical malpractice concerns are also frequently blamed, but there appears to be no relationship between documentation burden and state-level malpractice risk.²⁶ Furthermore, physicians who are sued tend to increase the intensity of care they provide following a malpractice case,²⁷ suggesting that some part of malpractice-derived documentation burden may be mechanically tied to the increased intensity of care provision. Given this multifaceted relationship, future research should seek to tease apart the contributors to documentation burden in specific settings. Qualitative research is needed to explore the factors within ACOs that may contribute to increased documentation burden, and other, more insulated settings such as insurer-owned practices and concierge medical practices should be explored similarly. Finally, quantitative analyses are needed to capture differences across practice settings in both the specific mechanisms driving documentation burden and physicians’ experience of documentation burden.

Limitations

Our study has several limitations to consider. First, our data are drawn from self-reported survey data, which can be biased due to nonresponse, recall errors among respondents, and other factors. Still, our measures of program participation are likely to be reliable, and our measures of documentation burden are consistent with those of past studies.^11,28 Second, all our analyses are correlational and we employ no causal study design, so we are unable to discount the possibility that our results are driven by some sort of unobserved omitted variable bias. Third, we are unable to directly attribute reported burden to specific aspects of documentation that support the VBP programs in NEHRS; as a result, EHR documentation burden as we measure it here is best understood as aggregate burden for a given respondent. Fourth, we do not observe in NEHRS the proportion of a given physician’s or practice’s payer mix that is under VBP contracting, so we cannot stratify by this variable or examine any nonlinearities that may exist in the relationship between relative exposure to VBP and documentation burden. Finally, due to the relatively small sample size of NEHRS, we are unable to exhaustively explore other factors (eg, staff support, practice ownership) that may moderate the relationship between VBP program participation and documentation burden.

CONCLUSIONS

Our study explored the relationship between VBP program participation and EHR documentation burden for US office-based physicians. Contrary to prevailing wisdom that VBP programs can reduce documentation burden, we find that ACOs specifically are associated with higher EHR documentation burden in terms of difficulty documenting care and after-hours documentation time. It is possible that the variety of functions that clinical documentation serves creates excess documentation burden, especially for physicians who are exposed to multiple models of payment. Future research is required to isolate the modifiable mechanisms that drive excess EHR documentation burden and to identify efficacious solutions.

Author Affiliations: MedStar Health Research Institute (NCA), Washington, DC; now with University of Maryland (NCA), College Park, MD; Office of the National Coordinator for Health Information Technology (VP, TLR), Washington, DC; University of California, San Francisco (AJH), San Francisco, CA.

Source of Funding: Office of the National Coordinator for Health Information Technology.

Author Disclosures: Dr Rolle is employed by the Office of the National Coordinator for Health Information Technology, which sponsors the National Electronic Health Records Survey. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (NCA, VP, TLR, AJH); acquisition of data (NCA, VP); analysis and interpretation of data (NCA, AJH); drafting of the manuscript (NCA, AJH); critical revision of the manuscript for important intellectual content (NCA, VP, TLR, AJH); statistical analysis (NCA); obtaining funding (TLR, VP); and supervision (AJH).

Address Correspondence to: Nate C. Apathy, PhD, University of Maryland, 4200 Valley Dr, College Park, MD 20742. Email: nca@umd.edu.

REFERENCES

1. Downing NL, Bates DW, Longhurst CA. Physician burnout in the electronic health record era: are we ignoring the real cause? Ann Intern Med. 2018;169(1):50-51. doi:10.7326/M18-0139

2. Aligning reimbursement with value. Harvard Business School Institute for Strategy and Competitiveness. Accessed November 17, 2023. https://www.isc.hbs.edu/health-care/value-based-health-care/key-concepts/Pages/aligning-reimbursement-with-value.aspx

3. Adler-Milstein J, Linden A, Bernstein S, Hollingsworth J, Ryan A. Longitudinal participation in delivery and payment reform programs among US primary care organizations. Health Serv Res. 2022;57(1):47-55. doi:10.1111/1475-6773.13646

4. Pham HH, Navathe AS. Why aren’t value-based payment models more successful? a failure to confront market dynamics. Health Affairs Forefront. January 27, 2022. Accessed November 17, 2023. https://www.healthaffairs.org/content/forefront/why-aren-t-value-based-payment-models-more-successful-failure-confront-market-dynamics

5. Addicott R, Shortell SM. How “accountable” are accountable care organizations? Health Care Manage Rev. 2014;39(4):270-278. doi:10.1097/HMR.0000000000000002

6. Apathy NC, Hare AJ, Fendrich S, Cross DA. Early changes in billing and notes after evaluation and management guideline change. Ann Intern Med. 2022;175(4):499-504.doi:10.7326/M21-4402

7. 25x5 Documentation Burden Reduction Toolkit. American Medical Informatics Association. Accessed May 3, 2023. https://amia.org/about-amia/amia-25x5/25x5-documentation-burden-reduction-toolkit

8. Stephenson J. National Academy of Medicine outlines plan to curb burnout, bolster health workforce well-being. JAMA Health Forum. 2022;3(10):e224549. doi:10.1001/jamahealthforum.2022.4549

9. Holmgren AJ, Downing NL, Bates DW, et al. Assessment of electronic health record use between US and non-US health systems. JAMA Intern Med. 2021;181(2):251-259. doi:10.1001/jamainternmed.2020.7071

10. Strategy on Reducing Regulatory and Administrative Burden Relating to the Use of Health IT and EHRs. Office of the National Coordinator for Health Information Technology; 2020. Accessed December 4, 2023. https://www.healthit.gov/sites/default/files/page/2020-02/BurdenReport_0.pdf

11. Gaffney A, Woolhandler S, Cai C, et al. Medical documentation burden among US office-based physicians in 2019: a national study. JAMA Intern Med. 2022;182(5):564-566. doi:10.1001/jamainternmed.2022.0372

12. Nguyen OT, Turner K, Parekh A, et al. Merit-based incentive payment system participation and after-hours documentation among US office-based physicians: findings from the 2021 National Electronic Health Records Survey. J Eval Clin Pract. 2023;29(2):397-402. doi:10.1111/jep.13796

13. Adler-Milstein J, Zhao W, Willard-Grace R, Knox M, Grumbach K. Electronic health records and burnout: time spent on the electronic health record after hours and message volume associated with exhaustion but not with cynicism among primary care clinicians. J Am Med Inform Assoc. 2020;27(4):531-538. doi:10.1093/jamia/ocz220

14. Attipoe S, Walker DM, Schweikhart SB, Hefner JL. A qualitative study of the dark and bright sides of physicians’ electronic health record work outside work hours. Health Care Manage Rev. 2023;48(2):140-149. doi:10.1097/HMR.0000000000000361

15. National Electronic Health Records Survey: what’s new. CDC. Updated February 1, 2024. Accessed November 15, 2023. https://www.cdc.gov/nchs/nehrs/new_nehrs.htm

16. Colicchio TK, Cimino JJ. Clinicians’ reasoning as reflected in electronic clinical note-entry and reading/retrieval: a systematic review and qualitative synthesis. J Am Med Inform Assoc. 2019;26(2):172-184. doi:10.1093/jamia/ocy155

17. Cimino JJ. Putting the “why” in “EHR”: capturing and coding clinical cognition. J Am Med Inform Assoc. 2019;26(11):1379-1384. doi:10.1093/jamia/ocz125

18. Apathy NC, Rotenstein L, Bates DW, Holmgren AJ. Documentation dynamics: note composition, burden, and physician efficiency. Health Serv Res. 2023;58(3):674-685. doi:10.1111/1475-6773.14097

19. Perloff J, Sobul S. Use of electronic health record systems in accountable care organizations. Am J Manag Care. 2022;28(1):e31-e34. doi:10.37765/ajmc.2022.88818

20. Bleser WK, McStay F, Muhlestein D, McClellan MB. Accountable care in 2023: evolving terminology, current state, and priorities. Health Affairs Forefront. February 24, 2023. Accessed August 2, 2023. https://www.healthaffairs.org/content/forefront/accountable-care-2023-evolving-terminology-current-state-and-priorities

21. Apathy NC, Hare AJ, Fendrich S, Cross DA. I had not time to make it shorter: an exploratory analysis of how physicians reduce note length and time in notes. J Am Med Inform Assoc. 2023;30(2):355-360. doi:10.1093/jamia/ocac211

22. Micek MA, Arndt B, Baltus JJ, et al. The effect of remote scribes on primary care physicians’ wellness, EHR satisfaction, and EHR use. Healthc (Amst). 2022;10(4):100663. doi:10.1016/j.hjdsi.2022.100663

23. Barry MJ, Tseng CW. Moving to more evidence-based primary care encounters: a farewell to the review of systems. JAMA. 2022;328(15):1495-1496. doi:10.1001/jama.2022.18346

24. Maisel N, Thombley R, Overhage JM, Blake K, Sinsky CA, Adler-Milstein J. Physician electronic health record use after changes in US Centers for Medicare & Medicaid Services documentation requirements. JAMA Health Forum. 2023;4(5):e230984. doi:10.1001/jamahealthforum.2023.0984

25. Neprash HT, Golberstein E, Ganguli I, Chernew ME. Association of evaluation and management payment policy changes with Medicare payment to physicians by specialty. JAMA. 2023;329(8):662-669. doi:10.1001/jama.2023.0879

26. Holmgren AJ, Rotenstein L, Downing NL, Bates DW, Schulman K. Association between state-level malpractice environment and clinician electronic health record (EHR) time. J Am Med Inform Assoc. 2022;29(6):1069-1077. doi:10.1093/jamia/ocac034

27. Carroll C, Cutler DM, Jena A. How do physicians respond to malpractice allegations? evidence from Florida emergency departments. National Bureau of Economic Research working paper 28330. January 2021. Accessed August 11, 2023. https://www.nber.org/papers/w28330

28. Rotenstein LS, Apathy N, Landon B, Bates DW. Assessment of satisfaction with the electronic health record among physicians in physician-owned vs non-physician-owned practices. JAMA Netw Open. 2022;5(4):e228301. doi:10.1001/jamanetworkopen.2022.8301