Richard Rathe, MD

I’ve annotated this NY Times graphic, rounding the estimates for COVID19 (aka the 2019 novel coronavirus) to 2% mortality and 2 people infected for every one who has the disease. What this suggests is the infection rate will be similar to a bad cold season, but unlike the common cold, a significant number of people will die of the disease.

This is not good news, but it is also not catastrophic.

By comparison the 1918 “Spanish” Flu had a similar infection rate but killed 1 in 10! The 1918 Flu infected approximately one third of ALL the people on earth… and that was before air travel!

Based on these numbers alone, up to 110 million people in the US could be infected. 80% of these cases should be mild, 20% will need some form of special care in hospitals or at home, and 2% (2.2 million) could die.

One area where the 1918 Flu differs from COVID19 is the latter does not seem to kill children, which is a small bright spot. However, children may be a source of infection for vulnerable adults!

Note that the currently reported mortality rate is over 3% but here in the US we have done almost NO testing, so the total number of infected individuals is unknown, but likely to be much higher. This means the reported mortality rate is probably inflated due to lack of good data. I’ll stick with 2% as a conservative figure for now.

It is also important to remember one of the lessons of the 1918 epidemic, things may look better when warmer weather returns but this should not be reassuring. In the spring of 1918 it looked like the disease was weakening, but it returned with a vengeance in the fall and winter!

Social Isolation and Hand Washing

For the general public, not living with or caring for a sick person, wearing masks really doesn’t help. Social isolation (eg, staying home) and effective hand washing remain the best things you can do to prevent getting most viral respiratory diseases (the common cold, influenza, and the new coronavirus). Just being in the same room with a sick person is not that risky, but touching surfaces and objects in that room and then touching your face is a good way to catch one of these viruses. Note that hand washing studies suggest we generally neglect our fingertips (including the nails) and wrists.

Wash Your Finger Tips!

The best information available comes from CDC and WHO (which I consider the best source since it is not caught up in US political shenanigans). For up-to-date worldwide information download the most recent WHO Situation Report.

I was taught in medical school that pain scores were only useful for tracking progression/remission for individual patients. In my opinion it was a fundamental mistake to apply this tool as an objective “vital sign” across all patients. It simply doesn’t work that way. Scores remove context!

During my training we used an implicit 3 point scale: mild, moderate, severe. In later years I pondered why I still preferred this over the newer 10pt version. I decided it was because a 3pt scale is easy to tie to function: mild=annoying, mod=disruptive, severe=disabling.

I think the average patient views their condition in similar terms, rather than “my pain a 6 or a 7 out of 10.”

If you take the 3pt functional symptom scale and couple it with a 3pt frequency scale: intermittent, frequent, constant—then you have a 3×3 grid that might be more meaningful than a one dimensional 10pt scale.

It may be appropriate to record more than one score, for example: the patient has constant mild dizziness (3/1) with occasional disabling episodes (1/3).

I have worked with home-grown and commercial Electronic Medical Records for over thirty years. The use of “dot commands” (a period followed by a trigger phrase) is about as old as personal computing. (I first encountered dot commands in the WordStar word processing program during the 1980s!) These commands generally fall into three categories: a) links to retrieve data (name, age, lab values, etc.), b) links to specialized services or applications and c) macros that expand into “canned” text. I’ve always found the later somewhat cumbersome and an impediment to fluent typing.

While working on a new set of macros for my Rational HPI Project, it struck me that all those periods were getting in my way. (I can touch type but special characters generally slow me down.) This is particularly true when using many short fragments. I decided to experiment with a standardized clinical shorthand based on these criteria:

1. Include common abbreviations when they exist.
2. Expand common but potentially dangerous abbreviations into plain text (qd, qod, etc. are often banned in written notes).
3. Make the triggers (aka “shorts”) as mnemonic as possible.
4. Prevent transformation of non-shorthand text. (A paragraph of plain text should generally pass through unchanged.)
5. Limit punctuation and the need to use modifier keys (i.e., shift, alt, control, etc.). This is especially important for handheld devices such as tablets and smartphones.

My first attempt resulted in approximately 130 Shorts (listed below). Note that I’ve repurposed the article ‘a’ and the pronouns ‘I’ and ‘us’, since they have very little value in problem-focused notes. The goal is to generate functional documentation with the fewest keystrokes—not flowing paragraphs.

Typing all lower-case is assumed but not required. Capitalization is only enforced for canonical abbreviations (HPI, ROS, RUQ, etc.) and certain keywords that are frequently missed (NO, NOT, LEFT, RIGHT).

Punctuation is optional and should be kept to a minimum. How much punctuation to add automatically is an unsettled issue. At this time I include colons when a list is expected and commas for certain qualifying phrases. This is subject to change.

I have an expanding number of disease and symptom shorts that are not listed here. Ultimately these may need to be specialty specific.

Finally, certain phrases are included to support the needs of billing and compliance.

Clinical Shorthand v1.0

‘99′ indicates any number and ‘zz’ any phrase. Time unit ‘i’ indicates m(i)nute. Other time units are what you’d expect.

I’ve also created a proof of concept web app to experiment with documentation in real time…

Thanks to My Students & Colleagues!

Patients who are actively engaged with their health have better outcomes. Good communication is key. I made this Patient Instructions Card about six months ago and have been very pleased with it. I’ve turned it into a generic PDF with custom name, phone and tobacco resource fields. Fill in your particulars and have it printed at 50% on card stock. Enjoy!

The front has lots of blank space to write on…

The back has advice on common problems…

This is the first release of quickHPI for general use by students, residents and clinicians. It is based on version 1.1 with persistent client-side data storage added. Once installed, it can be used offline when network connectivity is unavailable or undesired.

The purpose of quickHPI is threefold: 1) Provide a practical tool for clinicians; 2) Create a functional expression of the ideas presented here; and 3) Offer a self-directed learning tool for students and residents.

QuickHPI is both a tutorial and a practical tool for recording the History of Present Illness. As such it has wide applicability in medical education and patient care. The program instantiates my research into best practices for outpatient documentation aka The Rational History of Present Illness.

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 

During April 2014 I gave a talk at an EMR meeting concerning the changes manifest in the everyday clinic note. Most are familiar with various approaches to generating notes that meet all requirements for billing, compliance, and liability. Patient care and physician efficiency frequently suffer. I am particularly concerned about “note bloat” and the tendency of automated systems to add noise and imprecision to medical documentation. I’ve been working with our EMR vendor to create a better approach to the History of Present Illness.

This week I launched the The Rational History of Present Illness project to develop and promote better documentation tools.

This is sample output from a template driven tool that does NOT attempt to generate english prose. For this hypothetical patient, the structured approach yields a complete HPI in 48 words and 290 characters (60 words if you count the labels). Also note the judicious use of directly typed input (in blue). One of the biggest problems I see with EMR templates is they over specify. Why check a box for a phrase that is sort of what you’re thinking when in a few words you can say precisely what you mean?! Now contemplate how easy it is to visually scan and assimilate the information. As Dr. Tufte would say, “there is very little Pfuff!” Least Ink in action!!

Cowboys vs Pit Crews

Key facts from his recent TED talk…

• Modern doctors have 4000 procedures and 6000 medications at their disposal
• In 1970 it took 2 FTEs to care for a patient in the hospital
• In 2001 it took 15 FTEs to care for the same patient
• Checklists and mandatory pauses help decrease mistakes

“We are all specialists now, even primary care physicians.”

Atul Gawande “How do we heal medicine?” (video)

To sum up his talk in two words, both the Intensity and Complexity of care have increased in the past forty years. This is not surprising when you think about it. What aspect of life in the US has not become more intense and complex? He focuses on how the education of physicians lags behind the realities of team-based care. This is a valid point as far as it goes. He does not speculate on why “even primary care physicians” are specialists and no one is charged with understanding the entire spectrum of care for individual patients? (So I will…) In a word, Economics! During the 1990s Medicare and other payers decided to stop paying for coordination of care by primary care physicians. This was exacerbated by division of outpatient and inpatient care in the past ten years (again primarily for economic reasons). The new focus on high-volume, episodic care leaves little time for the big picture. This is not a flaw in medical education and not a choice made by primary care physicians! Merely the predictable consequence of decisions made by insurers and the federal government.

The US Supreme Court is about to hear arguments for and against the recent healthcare insurance reform law enacted by Congress. (aka The Affordable Care Act, aka Obamacare) At issue is the constitutionality of the individual mandate to buy health insurance. It is unclear whether they will set new precedent or rule on a much narrower basis. Their decision could have profound effects on the current law and future legislation. Partisans on both sides of this issue should welcome a broad decision and be ready to accept the consequences!

If the mandate is found to be constitutional, this strengthens the notion that private insurance has a significant role to play in our mixed healthcare system. It bolsters popular components of the law that limit insurance companies’ ability to deny coverage for pre-existing conditions, place lifetime caps on benefits, and allow young adult children to remain on their parents’ insurance. The real challenge—reining in the exploding costs of healthcare—is still ahead of us.

If the mandate is found to be unconstitutional, this is the best argument yet for a single payer system in the US. You cannot have a viable healthcare insurance system if a significant number of citizens have the ability to opt out when they’re young and healthy. Beware the law of unintended consequences! If it turns out to be unconstitutional to force broad participation in private insurance, can a public option be far behind? Careful what you wish for!

Update: Nice Summary of What’s at Stake on ProPublica!

July Update: Justice Roberts decided the mandate was actually a tax. So the law stands, but apparently this presages other limits to the Commerce Clause.

I recently updated my lecture on cough, given to third-year medical students and residents. It presents an algorithm I developed based on the omnibus supplement published in the journal Chest and other sources. Here is a quick summary of key points that are often missed  by primary care physicians…

1. Acute cough is largely due to viral infections and therefore antibiotics are NOT indicated.
2. There is growing evidence that first generation antihistamines are the drugs of choice for undifferentiated acute cough.
3. Post-infectious cough is the most common etiology for cough lasting between 3-8 weeks.
4. Secondary causes of cough (reactive airways, GERD, post-nasal drip) should be considered based on symptoms and time course.
5. In areas of high TB prevalence consider testing for active disease in any patient with a cough lasting more than two weeks (WHO recommendation).

I concise version of the lecture is available as a podcast.

The lecture handout is available in PDF format.