The Physical Attractiveness of Electronic Physician Notes

Abstract

Though notes in electronic medical record systems (EMRs) have advantages, they are often criticized for their unattractive and unprofessional appearance. We sought to identify notes regarded by physicians as attractive and unattractive and the characteristics of both. We sent a sample of 10 notes representing a variety of common note types to a sample of 70 physicians who are either leaders in UW Medicine or randomly selected note authors and asked them to rank the notes by their physical attractivness. We found their rankings were not random, and notes clustered into those considered most, intermediate, and least attractive. Characteristics of each group are presented. We’ll use these results to format EMR notes in a more attractive form—a goal important to our physicians—while preserving important clinical, quality and compliance features that were our original goals in moving to electronic notes.

INTRODUCTION

Though notes in electronic medical record systems (EMRs) have many advantages such as improved legibility and ease of access, EMR notes have been criticized for containing inappropriately copied and pasted text,¹ poor readability² and also for their unprofessional general appearance. While several articles have addressed copying and pasting, less attention has been devoted to how notes appear when printed or viewed electronically by other clinicians or referring physicians. Many physicians regard this to be an important matter, since notes are often transmitted to referring physicians in hopes they have a professional, attractive appearance. In the years since we began our transition to electronic records, many physicians have complained to us that while electronic notes have advantages they have generally declined in their appearance. Though attractive appearance is not the major goal of writing an electronic note, it is regarded as important by many and perhaps most physicians. The purpose of this study is to identify notes physicians consider to be physically attractive and unattractive and the characteristics of both, so that we can improve the appearance of notes in our EMRs and when viewed by users of our EMRs and by physicans outside our institution.

BACKGROUND

Setting

This project was undertaken as part of a quality improvement initiative of the UW Medicine Health Information Management Committee. UW Medicine includes 3 hospitals, University of Washington Medical Center (UWMC), Harborview Medical Center (HMC) and the Seattle Cancer Care Alliance (SCCA). In 2007, UWMC had 390 beds, 72 ICU beds, and 18,120 admissions; HMC had 369 beds, 73 ICU beds and 18,777 admissions. The SCCA has one hospital for adult patients, consisting of 18 beds. All three hospitals serve as teaching hospitals for the University of Washington. There are 792 active hospital medical staff UWMC, 492 at HMC, and 297 at SCCA. Across all hospitals there are approximately 1,000 residents and fellows, 826 medical students 1,610 full-time and 1,182 part-time nurses. Outpatient care is provided in hospital associated clinics, 2 emergency rooms and 7 UW Neighborhood Clinics. Collectively these clinics and emergency rooms provide approximately one million outpatient visits annually.

How notes are written in our EMRs

Several EMR systems are used in UW Medicine. For inpatient care and most clinics, Powerchart (Cerner Corp.) is used for creation and storage of electronic notes.³ In some clinics, EpicCare (Epic Systems) is used. In operating rooms, Docusys is used for physician documentation. Physician notes created in Powerchart or EpicCare are created using a variety of techniques, including templates, direct typing of narrative text notes, voice recognition software, (predominantly Nuance Dragon Naturally Speaking), and by dictation with professional transcription. The method used is determined by physician preference and local policies. To reduce transcription costs, physicians have been encouraged to reduce use of dictation but in most care settings dictation is permitted. Dictation is used less often for inpatient care because of the need for rapid note availability. Templates used are created both by individuals and by information technology support personnel. All templates are reviewed by the professional fee compliance department except those created only for individual use. Template standards are proposed and loosely enforced by a variety of local committees; in practice, because of the rapid growth in electronic note use there is wide variation of template style. A proposal to standardized template overview in a high level committee is being reviewed by the Health Information Management Committee. Each day, roughly 1,500 inpatient electronic notes and 2,000 outpatient notes are created by resident and attending physicians.

METHODS

We asked 70 physicians to participate in this study. Forty are physician leaders in UW Medicine drawn from committees overseeing our medical record or with developing administrative and clinical policies for our hospitals and clinics, and 30 were randomly selected from a list of physicians who recently wrote notes in our EMR systems. All participants are clinically active in our academic medical center, and many maintain close professional relationships to referring physicians. We used 70 physicians for logistical rather than sample size considerations.

The authors selected 10 notes from among the most common note types used in our EMRs. These are clinic notes, inpatient progress notes, and consult notes. From these note types, we selected notes that in our opinion reflected a spectrum of physical attractiveness when viewed in printed form, understanding that our judgement may differ from that of others. These notes were created using different EMR systems, and within each EMR system using a variety of technicques. Each of the notes was manually reviewed and any text that could be used to identify a patient or provider was obscured by editing or using an opaque ink marker. This sample was representative of the notes contained in patient records created in UW Medicine.

We sent each physician these ten notes in random order in an envelope. Along with the collection of notes we sent a cover letter with the following instructions: “In the envelope with this letter you will find ten notes. Lay these on your desk, and sort them with the note you find to be the most attractive in appearance on the top, the note that is least attractive on the bottom, and with the other 8 notes sorted from most attractive to least attractive in between. For purposes of this study, do not consider the content of the note, but rather how attractive it appears when viewed by UW Medicine and referring physicians. When you are done, put the sorted stack in the enclosed envelope and mail it back to me via interdepartmental mail.” We included a card redeemable for five dollars of coffee as compensation for participating in this study. A separate sheet was included to track completion of the task. No identifier linked the returned stack of notes to an individual physician participant. We recorded the sequence of notes in the returned envelopes, and assigned rank of 1 to the first note, 10 to the last note, and 2–9 to the notes in between. We summed the rankings assigned by all physicians to each note. We tested the null hypothesis that physician ranking of attractiveness for the 10 notes could be explained by chance alone, using the Kruskal–Wallis one-way analysis of variance by ranks. The level of significance threshold was 0.05.

Note characteristics

Before ranking results were received, we created a list of characteristics that distinguished the notes from each other, shown in Table 1.

Table 1.

Characteristics of notes ranked most and least attractive.Notes ranked as most attractive had simpler formatting and more narrative text.

Note characteristics	Most attractive	Least attractive
No. of font types	1.7	6
No. of paragraphs/page	6	13.9
No. of heading types	1	4
No. of lines in note	91.3	143
% of lines that are checklists	0%	9%
Professionally transcribed	Yes	No
% of lines containing narrative text	50%	16%

The study was reviewed by the UW Human Subjects Division and deemed exempt from full review.

RESULTS

Of the 70 physician participants, 4 were unable to participate because they were no longer with UW Medicine or were on leave. We received stacks of sorted notes from 50 of the remaining 66 physicians invited to participate, a response rate of 76%.

Ranking of attractiveness by physicians

The sum of the rankings from all physicians is shown in Figure 1. We found that physician ranking of results was significantly different than that predicted by chance alone. Generally, the 10 notes fell into 3 groups according to the sum of the rankings: Highly ranked (106–115), middle ranked (213–259) and low ranked (331–396).

Figure 1.

Physician rankings of attractiveness of notes. Bars indicate sum of rankings. Lower sums indicate higher ranked notes. X axis are note numbers. Using Kruskal-Wallis test we rejected the null hypothesis that physician ranking can be explained by chance alone, p<0.0001.

Figure 2 shows the rank assigned to each note by physician participants. In this figure, the notes were sorted by the average attractiveness rank assigned by physicians. In general, there was greater agreement among physicians regarding the notes they felt were the most attractive and least attractive, and less agreement about the ranking of notes in the middle of the attractiveness ranking. Two notes (numbers 10 and 1) had a wide range of attractiveness rankings but these were the exception to the general pattern of polarized ranking of this sample of notes by physicians. The physicians selected to rank notes used a variety of methods to enter notes in practice.

Figure 2.

Ranks assigned to each note by physicans. Ranks of 1–10 are on left, note number is on top. Each dot indicates the rank assigned to the note by one physician. Dot plots are sorted by rank sum shown on bottom.

Association between physician-assigned attractiveness rank and note characteristics

We show a subset of the note characteristics that we used to describe each note. We counted use of bold type as a separate font from the same font in standard face. Narrative text means use of complete sentences.

Characteristics of notes ranked most and least attractive

The 3 notes in the group ranked most attractive had a simple structure—fewer paragraphs and heading types-- used only 2 fonts throughout, had bold, capitalized headings to the main sections, contained a large proportion of their content as narrative text with complete sentences, and were all professionally transcribed. They did not contain checklists created using boxes or brackets [].

The notes ranked least attractive had a large number of paragraphs, many heading styles (Figure 4). They were generated by templates, had a small proportion of their content was narrative text containing complete sentences, and may contain checklists.

Figure 4.

Excerpt from note ranked among least attractive. This notes has a lower proportion of narrative text, includes check boxes, uses many different fonts and many paragraphs in each page.

DISCUSSION

EMR notes are new to most of us. A recent survey found that only 12% of US hospitals have implemented electronic physician notes throughout their institution.⁴ The literature describes many problems with them, including inappropriate copying and pasting, their excessive length, difficulty quickly understanding their content, difficulty knowing what is new, and what is old within the note, and the fact that they aren’t as attractive as notes used to be. In our experience, physicians also complain about the long sections of information automatically inserted into notes, such as medication lists, laboratory results, health maintenance information, and their perception that this is done without relevance to the purpose of the individual note.

To our knowledge, this is the first study of the attractive appearance of physician notes. In national meetings attended by one of the authors, this is a topic on the minds of many physicians. Electronic notes can be entered in a variety of ways; our study included notes created in several possible ways. We realize that what we find is subjective and others may disagree with our subjects’ rankings.

It is true that attractiveness is only one characteristic of notes, and not the most important one. Other characteristics include legibility, clinical helpfulness, compliance, ability to support reimbursement, measurement of quality, conformance to Joint Commission requirements, and others. In the minds of some physicians, some of these other goals of notes have assumed importance at the cost of declining clinical utility and professional appearance, however.

Notes can be viewed in a variety of ways: Within the native EMR application used to create the note, printed for mailing or filing in the chart, or within a web browser. We’ve studied notes in printed form because in the US according to a recent NEJM article, only 12% of hospitals have entirely electronic physician notes, so the majority of notes are likely to be read on paper in the US.

We found that physicians generally agreed in their ranking of notes according to note attractiveness. As Figure 1 shows, some notes were regarded as more attractive and others less attractive. The characteristics that distinguish more attractive from less attractive are simplicity as reflected in fewer font types, fewer paragraphs per page, and a higher percentage of narrative text. The notes ranked as more attractive in our sample were professionally transcribed, though other entry methods might also achieve the visual appeal without transcription. This is important because the tools used to create and display note content can be changed by existing functionality provided by EMR vendors, and could potentially be changed by further functionality that is not yet available. Transcription costs are high in our organization, and the task of transcribing notes increases delay in note availability. Evidence such as that provided by this study, if confirmed by other investigations, can potentially aid in improving appearance of notes in electronic records.

Why is this important? Note appearance is important to physicians. Many participants thanked us for addressing note attractiveness, adding “Thanks for doing this!” “Our notes have become complete garbage.” There is also evidence from computer science and design literature that there is a close relationship between users’ initial perceptions of interface aesthetics, which can be measured,⁵ and their perceptions of the system’s usability⁶ and value.⁷ To the degree that this is true in the domain of physician notes, when faced with a large collection of notes to read, the attractiveness of notes might influence the choice to read a note. It may also affect the perception of professionalism of the note author and institution.

What can we do with results? After devoting effort to the transition from paper to electronic notes, we are now focusing attention on improving note quality, and improving appearnce is part of this effort. We can configure software so that notes are more attractive yet also achieve objectives sought when using EMRs, and permit direct entry, lower cost, higher compliance. Our study suggests that techniques to improve note appearance may be to simplify note formating and to increase narrative text; changing EMR configuration may help achieve this goal. These results might also be useful to EMR vendors when enhancing documentaton tools.

Our work may contribute to broader efforts to better understand how notes in EMRs can be of higher quality, clinical utility, and contribute to measument of clinical quality. Improving the attractiveness of notes can at least theorically be achieved while achieving characterstics described by Stetson.⁸ In that paper, the authors found that “clinicians believe high-quality clinical notes should be well-formed, comprehensible, accurate, and compact.” In our experience, the professional appearance of notes is also important to physicians, and may be another consequence of well-formed, compact notes.

LIMITATIONS

This study has several limitations. It is confined to a small sample of physicians in a single institution, using a small number of notes. The attributes we identified to distinguish highly ranked from lower ranked notes may not be the most important ones in the minds of physicians or in other institutions. We did not seek to balance the need for other note characteristics, such as compliance or inclusion of quality indicators in measurable form, with note attractiveness though this could be accomplished in further work.

However though document design principles have been described for other domains,⁹ to our knowledge this is the first attempt to analyze the attractiveness of electronic physician notes, which in our institution is considered an important issue by physicians.

CONCLUSIONS

There were clear and signicant differences between physician ranking of a the attractiveness of 10 notes sampled from our electronic medical record system. These notes were created using different techniques. The notes ranked most attractive by the 66 physicians used fewer fonts, simpler headings, more narrative text, and fewer sections, and were professionally transcribed. Notes ranked as intermediate in attractiveness were generated with locally-developed templates and included a greater number of different fonts, more paragraph headings, a lower proportion of narrative text, and were likely to contain a mixture of typed text and system generated material. The notes ranked as least attractive contained many different headings

Figure 3.

Note ranked as most attractive. It contains a high proportion of narrative text, only 2 fonts, clear headings, and few separate sections.