nexus one

In the Nexus One ad on google.com, the Google Maps demo plots a route starting from SFGH.

cpx gripe

The cpx is a simulated test where we work with patient actors to demonstrate that we can do a good history and physical. I do think it is a valuable experience, I like that we are video recorded and can watch ourselves and I like hearing the actors feedback. My gripe is with the rubric which has a long list of things we did or did not do, and sort of forces students to "go through the motions."

For example, one requirement is that students "personally confirm a reported BP measurement from a nurse." I find that we trust nurses to do all sorts of things that are both difficult to do andhave catastrophic effects if done improperly. I do not see why medical schools single out BP measurement as something requiring MD confirmation in particular.

health care bill

breakdown of how the bill affects various groups

http://online.wsj.com/article/SB125763756556136303.html

compensation

The RBRVS isn't really mentioned in our medical school curriculum. I think it's so important to keep a mental tab on how much any particular visit or procedure costs.

Article on healthcare spending waste

I had a couple gripes with the following article:

Reuters had a healthcare analytics group try to estimate wasted health care dollars. It estimated $700 billion with the following breakdown:
-Unnecessary care (antibiotics and lab tests) (37%)
-Fraud (fraudulent Medicare claims, kickbacks for referrals for unnecessary services) (22%)
-Administrative inefficiency / paperwork (18%)
-Medical mistakes (11%)
-Preventable conditions such as uncontrolled diabetes

So here's what I have to say:
#Unnecessary care:
Often, hospital admit order become "grandfathered" as a routine for daily labs. If you analyze it with a fine tooth comb, you can argue that patient so-and-so technically doesn't need lytes/CBC every day, or you need LFTs even less often. But I think it doesn't cost too much to run a CBC on blood that's already drawn, whereas it would take a lot of work to fine-tune each patient's routine labs for an active 10-20 patient census. That extra work that it would take would be a pretty big hit on hospital productivity I think. I would make the analogy that you waste more gas turning a car on and off continuously than just idling the tank.

#Preventable Conditions:
I think uncontrolled diabetes is indeed a preventable condition, but not one preventable by improving healthcare quality. The patients I saw last rotation who had uncontrolled DM seemed to have trouble following their medication regimen in spite of good continuous care by a patient fam doc.

i've always liked olives

random thought: is there an easier way to "get into AOA"

apparently there is:
http://australianolives.com.au/web/index.php

the Australian Olives Association

can you add likelihood ratios?

From random conversations I've had, I feel like there's been a big push to implement EBM and likelihood ratios among MS3's. The way it has been explained to me is that given various patient demographics and symptoms you can generate a pre-test probability for any sort of disease. Let's say 20% of older men with chest pain who present to the ED have a MI. You then do a test like EKG or serial trops q8h and based on intrinsic properties of the test (sensitivity, specificity) you can calculate a likelihood ratio. And based on the outcome, you can now get a post-test probability.

One question I've always had is what happens if you do two tests with differing outcomes. Do you add or subtract likelihood ratios, or do you use the LR table twice, or is it inconclusive. I've never had a EBM speaker give me a really good answer - often I get a range of responses from "no, you can only do 1 test," to "yes you can just repeat the LR more than one time and it'll intrinsically never exceed 100% likelihood."

This morning, I finally got a good answer that makes empirical sense to me. Basically the answer is it depends on whether the two tests are independent of each other. For example, nausea and vomiting have a lot of overlap, so if someone is being tested for gastroenteritis, it would be unfair to calculate LRs for both nausea and vomiting and add that to the pre-test probability in succession (e.g. 20% --> 40% --> 60%). It would make sense if +nausea makes someone go from 20% -> 40%, and +vomiting makes someone go from 20% -> 41% or so, and if both, maybe 20% -> 45%. But if you want to do genetic testing for someone for susceptibility to Alzheimer's and you have something like a chromosome 21 gene and a ApoE4 gene, these are totally independent markers, so it makes sense to calculate two likelihood ratio calculations in succession 20% -> 40% -> 60%.

So in real life, no one really knows offhand how independent any two tests are. How many people with +D-dimer get +CT angio, how many people with +RUQ pain also get +leukocytosis. So it's best to use algorithms when available (Well's, TIMI, APACHE, PORT score, CHADS2).

wolfram alpha

http://www.wolframalpha.com
is a new search engine for all things numbers-related. You can type in things like 2+2 and it'll give you 4. It also does conversions, and plots of weather and what not.

medically-related, here are search queries that work:
500bp upstream of gene FASTKD2
ldl 150 male smoker age 40
life expectancy male age 35 russia
ldl versus crp male smoker
inpatient surgeries Massachusetts General Hospital
icd-9 440
colon cancer
ibuprofen drug interactions

Notes from the ICU

a "wise guy" is someone who likes to say

neuro-wise, blah blah blah.
Cardiovascular-wise, blahdiddy blahddidy blah
Pulm-wise, blah blah blah
FEN/renally, we should blah-dah-dah-didity blah
GI-wise,,,
Heme/ID wise..
in terms of prophylaxis....

actual doctor's notes

found online:

These are doctors’ notes on patients’ charts: (Actual notes - unedited!)
1. Patient has chest pain if she lies on her left side for over a year.
2. On the 2nd day the knee was better and on the 3rd day it disappeared completely.
3. She has had no rigors or shaking chills, but her husband states she was very hot in bed last night.
4. The patient has been depressed ever since she began seeing me in 1993.
5. The patient is tearful and crying constantly. She also appears to be depressed.
6. Discharge status: Alive but without permission.
7. Healthy appearing decrepit 69 year-old male, mentally alert but forgetful.
8. The patient refused an autopsy.
9. The patient has no past history of suicides.
10. Patient has left his white blood cells at another hospital.
11. Patient’s past medical history has been remarkably insignificant with only a 40 pound weight gain in the past three days.
12. Patient had waffles for breakfast and anorexia for lunch.
13. Between you and me, we ought to be able to get this lady pregnant.
14. Since she can’t get pregnant with her husband, I thought you might like to work her up.
15. She is numb from her toes down.
16. While in the ER, she was examined, X-rated and sent home.
17. The skin was moist and dry.
18. Occasional, constant, infrequent headaches.
19. Patient was alert and unresponsive.
20. Rectal exam revealed a normal size thyroid. (ouch!)
21. She stated that she had been constipated for most of her adult life, until she got a divorce.
22. I saw your patient today, who is still under our car for physical therapy.
23. Both breasts are equal and reactive to light and accommodation.
24. Exam of genitalia reveals that he is circus sized.
25. The lab test indicated abnormal lover function.
26. The patient was to have a bowel resection. However, he took a job as a stockbroker instead.
27. Skin: Somewhat pale but present.
28. The pelvic examination will be done later on the floor.
29. Patient was seen in consultation by Dr. Blank, who felt we should sit on the abdomen and I agree.
30. Large brown stool ambulating in the hall.
31. Patient has two teenage children, but no other abnormalities.