The Rational History of Present Illness

Efficiency, Readability, Validity

Richard Rathe, MD (2017)

Associate Professor of Family Medicine (ret.) and Medical Informatician

This page is home to my study of the changes to Clinical Documentation in the age of Electronic Medical Records. Many of these changes are not positive!

The History of Present Illness (HPI) has not changed appreciably since the introduction of the Problem-Oriented Medical Record in 1968. Until recently most medical documentation was created via dictation or handwritten notes. Narrative was the natural format for the HPI. This is no longer true in the era of online forms and computer-generated text.

The work of physicians has changed profoundly since the late 1990s. At that time the changes made by CMS (then HCFA) were coming into full effect and the Electronic Medical Record (EMR) revolution was gathering momentum. We are now approaching an era where most medical data will be created and accessed electronically. At the same time resources are being withdrawn from patient care to deal with the onslaught of regulation and the complexities of reimbursement. This has put pressure on physicians to become more efficient and do more with less.

Modern EMR systems provide a number of semi-automated tools and techniques that may increase physician productivity when used properly. But there is a dark side to EMR automation that often leads to bloated, inaccurate medical documentation.

By tradition the HPI is a narrative of what the patient experienced and why they sought medical care. This is the natural format for dictation or writing longhand. A well-written HPI is concise, accurate and easy to assimilate. With the new dominance of EMRs however, the process has changed in significant ways. The typical note is now based on a template, which in turn is a composite of several auto-copy links and point-and-click forms. The modern HPI is more constructed than written.

Many systems attempt to generate grammatically correct sentences that mimic a narrative HPI and this is where the trouble begins. Key historical points become obscure, buried within paragraphs of mostly filler text. The time course is often ambiguous and the narrative structure is lost. Reading becomes monotonous and every case sounds the same. Together these factors lead to increased noise, decreased signal, and LESS efficient communication.

Equally important are the subtle ways each tool determines what is actually recorded. To cite just two examples: 1) auto-linking a symptom when it is not pertinent to the current episode, and 2) being forced to select from a list when there are no options that match the situation at hand. In both scenarios the EMR creates pseudo-findings that were not intended or information is simply discarded because there is no place to put it.

A rational approach to the History of Present Illness abandons the pretense of narrative and records just the facts with as little filler as possible. This is an example of Tufte's Least Ink Principle (see below) applied to medical documentation. Clinicians should continue to use narrative in their notes when appropriate, but constructed pseudo-narrative notes should be replaced by more useful formats.

I have been working on tools within the Epic Systems EMR for several years under the rubric Rational History of Present Illness (RHPI). In 2015 I created an open-source tool called quickHPI (which you will see referenced below). Recently I began work on a Clinical Shorthand (aka ShortNote) and a web app to interpret it.

Below I’ve listed the major issues that must be addressed by a more rational History of Present Illness…

1. Striking a Balance Between Clicking and Typing – Many HPI forms have too much specificity for use in primary care. They tend to have fields of buttons where a few typed words would be faster and more precise. quickHPI avoids this trap by providing text fields where buttons just get in the way.
2. Platform Independence and Best Practices – With the widespread adoption of EMRs, the HPI has gone underground. Each proprietary system has its own set of templates that are not generally accessible for open review and discussion. This inhibits collaboration, quality improvement and research. It is no wonder my initial literature review yielded very little useful guidance for this project. Quick HPI is open source and available to anyone with a web browser.
3. Support for Problem-Oriented Charting – EMR history forms often fall down when documenting multiple complex problems. The resulting documentation suffers from duplication and lack of context.
4. Drawing a Bright Line Between Associated Symptoms and the Review of Systems – Healthcare in the USA is all about rules. This includes documentation required for billing. From a physician’s perspective these rules are rather arcane. One casualty has been the outpatient note. In order for coders and auditors to efficiently review our notes, symptoms must be put into categories and certain key phrases must be present. Otherwise we run the risk of not being paid for our effort. A sad situation, but there it is. To a coder (and an EMR) a symptom is a symptom is a symptom, regardless of its clinical significance. This has led to the bad practice of including non-relevant symptoms in the HPI. Associated Symptoms (positive or negative) should be pertinent to the problem under discussion. The Review of Systems (ROS) is just a series of questions without context.
5. Format Notes to Facilitate Visual Scanning and Rapid Assimilation of Information – The phenomenon of Note Bloat is becoming a significant problem in healthcare. EMRs make it easy to build lengthy notes with a paucity of useful information (low signal to noise). Synthesized Text may actually obscure the key findings! Even worse are the errors introduced when information is repeatedly copied (garbage in, garbage out). QuickHPI generates notes in a structured format that minimizes the amount of non-informative text. This is referred to as the Least Ink Principle…

The best clinical documentation is that which gives to the reader the greatest amount of information in the shortest time with the fewest pixels.

Paraphrased from Edward R. Tufte
The Visual Display of Quantitative Information