How to hack your health

February 8, 2012

I thought I would have to live with an unusual medical condition for life. However, using the method I describe in this post, I found the one surgeon in the world who could help me.

For a little over four years, I have had a paralyzed left hemi-diaphragm. The result is reduced lung capacity (down by about 1/3), particularly on exertion. Until about six months ago, I thought I would live with this for the rest of my life. Many people live with much worse, of course, but this condition made it hard to do the very things I was trying to do to live a more healthy life, such as bicycling and weight-lifting.

Some medical conditions are more unusual than others

Being both an inquisitive geek and a bit of a hypochondriac, I dug into this problem as I would any other. What I learned about the current state of health care for unusual conditions was eye-opening, and I want to share those lessons here.

If you have a somewhat unusual medical condition, I am going to tell you exactly how to get the best health care you can from the current system. I think you’ll agree that what I discovered isn’t exactly intuitive. I hope you will find that some or all of what I describe is useful to you as well.

The first part of this post gives some background as to my personal story and how I came to need the information in this post. If you’d like to skip directly to the “how to,” click here.

No bad guys

It's better to have a common problem

I am not out to bash the health care system or any particular part of it. Heck, I’m married to a doctor, and I have many friends in health care. (So do you—they represent about 11.5% of the workforce.) It’s my experience that these are wonderful professionals doing the best—often extraordinary—jobs they can do under difficult circumstances.

It’s simply the fact of the matter that most health care professionals spend most of their time dealing with relatively routine issues. Talk to them about high blood pressure or diabetes, and you’ll be getting the benefit of 100s or 1000s of cases that they’ve treated in their careers.

However, ask them about your paralyzed left hemi-diaphragm, and…

The diagnosis

I “presented” with reduced breath sounds in my lower left lung and severe pain in my left upper back and shoulder. After shrugging it off for a few days, I saw my general practitioner, a family practice doc at a very large university medical center.

He diagnosed pneumonia, prescribed an antibiotic, and sent me for a chest x-ray.

Radiology looked at my chest x-ray and said they wanted another film with me lying down.

Given the new x-rays, my GP now diagnosed me with a pleural effusion (fluid outside one’s lungs, where pneumonia is fluid inside). Sent me home to essentially ride it out.

After 7-10 days, not much had changed, so GP sends me to a pulmonologist. The pulmonologist takes my history, then speculates that I have a paralyzed diaphragm. He sends me downstairs to radiology, where I have a simple “sniff test,” which confirms that, indeed, the left half of my diaphragm is not moving.

Each phrenic nerve innervates one-half of the diaphragm.

What to do? “I have no idea,” says the pulmonologist. He mentions a procedure called a plication, and sends me on my way.

The consultations

Over the next several months, I discuss my condition with a number of doctors, with the following results:

General practitioner: there’s nothing to be done.

Pulmonologist at university medical center: looks up “paralyzed diaphragm” in Wikipedia while I’m in exam room. No joke. (I did not pay for this visit.)

Friend who is top-notch pediatric neurosurgeon: nothing to be done but plication, see this general thoracic surgeon, best around.

Sometimes one opinion is enough.

Friend who is top-notch adult neurosurgeon: nothing to be done, but see this pulmonologist, best around.

Best general thoracic surgeon: we can plicate your left hemi-diaphragm. It won’t fix anything, but will stop the diaphragm from impinging upwards into your left lung.

(In the middle of all this, they find a cyst on my pancreas, which leads me to oncology, but that’s another story–with a happy ending.)

Best pulmonologist: nothing to be done, but do these breathing exercises, they may help strengthen your “accessory” breathing muscles to make up for the lack of diaphragm action on the left side. Oh, and don’t do the plication. (This turns out to be good advice.)

My conclusions

After hearing lots of advice from lots of doctors, I independently come to the following conclusions:

  • the problem is with my nerve, I should not go messing with my diaphragm (no plication)
  • surgeons repair nerves all the time; in fact, I had a nerve repaired when I cut my head open once; surely there is a way to actually fix my phrenic nerve

Note that these conclusions were almost the exact opposite of what I was told by every doctor I consulted. It is not easy to go against all that expert advice, though it helps to be somewhat smug and pig-headed (or so says my wife).

I essentially decided to keep researching until I found a method or a person that could actually repair my broken/damaged phrenic nerve.

But how?

Here’s the “how to” part of this post, finally.

How to

Step 1: discover all the possible names and descriptions for your problem

In my case, the first phrase I heard was “paralyzed left hemi-diaphragm.” But there were others. My full list included:

  • paralyzed hemi-diaphragm
  • diaphragmatic paralysis
  • diaphragmatic palsy
  • phrenic nerve palsy
  • Parsonage-Turner Syndrome (refers more to the pain in my arm/chest/shoulder, but this term was used by one of the pulmonologists I saw)

To build this list (and it’s critical), Googling will help, but be sure to ask every doctor you see for his or her description of the problem. Read your doctor’s visit notes. There are idiosyncratic usages, and you want to discover them all.

Make a comprehensive list of all these terms. Include any related concepts, if any. For example, the phrenic nerve is often accidentally cut during heart surgery, so I might have included the keywords, “phrenic nerve heart surgery”. Vet the list with a doctor friend if you can, but err on the side of too broad a search, rather than too narrow.

Step Two – search the medical literature

Fortunately for us, your tax dollars really are at work here. The National Center for Biotechnology Information (NCBI) (part of the National Institutes of Health’s National Library of Medicine) has an online database,, that is extensive, easy to search, and up-to-date.

PubMed is easy to search, but this will be the most time-consuming part of your research. You want to see if anything has already been published about your condition. (In the next step, you will set things up to continually monitor future publications.)

You are likely to find hundreds or thousands of articles about your condition, and there is no easy way to filter them–you’ll have to examine each one individually. You will probably only have to scan the article title, but this will still take some considerable time.

To get started, browse to, enter the first of your search terms in the search box at the top of the page, and settle in for some reading. Your results will look something like this:

Repeat this process for each of the search terms you listed in Step One. If you are a search guru and want to use some of PubMed’s more advanced features, such as ORing your search terms together, it may save you some scanning time.

As you perform your searches, be sure to look at PubMed’s “Related searches” section on the right side of the results pages. You may find some search terms that are relevant to your search that you hadn’t identified by other means in Step One. Add any such new  terms to your list.

(Note: if you happen to find some interesting results here, you can at least temporarily skip right to Step Four and follow up. If this doesn’t lead to a definitive solution for you, come back here and continue.)

At first, the (often) arcane, (usually) verbose language of science will be challenging, but you will soon develop a heuristic filter that will eliminate the need to dig into most of the articles presented to you.

For example: since PubMed indexes both clinical and “bench science” articles, you will probably see a mix of basic research and more practical (for your purposes) clinical information. You will often be able to easily weed out the bench science articles because they will contain clue words in the article title, such as the names of genes (usually presented in all caps). E.g., if you see, “Role played by P2X and P2Y receptors in evoking the muscle chemoreflex,” you can tell by the names of the genes P2X and P2Y (they are usually capitalized) that this is basic research, and unlikely to be of any immediate relevance to you.

Abstracts will tell you what you need 95% of the time. You will get used to quickly scanning article abstracts for clues to the relevance of the article to your condition. In particular, skip to the “results” or “conclusions” section(s) of the abstract, if one or both of these exists. These summaries will often tell you all you need to know.

For example, here is the title and the Results and Conclusions sections of the article that led me to finding a possible solution for my problem:

Reinnervation of the paralyzed diaphragm: application of nerve surgery techniques following unilateral phrenic nerve injury.

“Results: Measures of postoperative improvement included pulmonary function testing, fluoroscopic sniff testing, and a standardized quality-of-life survey, from which it was determined that eight of nine patients who could be completely evaluated experienced improvements in diaphragmatic function.

“Conclusions: Based on the favorable results in this small series, we suggest expanding nerve reconstruction techniques to phrenic nerve injury treatment and propose an algorithm for treatment of unilateral phrenic nerve injury that may expand the current limitations in therapy.”

(If you’ve read this entire post, you can probably imagine my excitement upon seeing this article, several years into my research.)

Step Three – set up and monitor an on-going search of the medical literature

If you are unsuccessful in finding any already-published literature that addresses your condition in a way meaningful for you, you can set up an on-going search to monitor PubMed for you and send you the results on a regular schedule. This is how I found my solution, several years after my initial PubMed search.

PubCrawler is a tool for setting up this ongoing search. It was was developed and is hosted by Ken Wolfe’s lab in the Genetics Department, Trinity College Dublin, Ireland. The web site (as opposed to the search tool behind the scenes) was developed by, and is currently maintained by Karsten Hokamp, Ph.D.

Find PubCrawler at

Since I started my research, PubMed itself has added a tool that functions similarly to PubCrawler. It is accessed via PubMed’s “saved searches” functionality, available to you once you have created an account.

You may find PubMed’s functionality slightly easier to use than PubCrawler. However, since I used PubCrawler, I want to give them props here.

Create an account and set up your query(ies). Setting up queries in PubCrawler is a little different than you may be used to. Instead of simply typing keywords into a search box, you enter your search terms individually, and must connect those search terms with the logical keywords AND, OR, or NOT. In addition, each of these logical keywords can include parentheses.

It may take a bit of fiddling to get what you want, and there is not much in the way of help. To get you started, here’s how I created one of my searches. I wanted to wind up with the following logical statement:

(“phrenic nerve” OR diaphragm) AND (paralysis OR palsy)

The PubCrawler way to achieve this is:

PubCrawler uses a form-based search.

The “search field” can have many values, which map to search fields in PubMed. For our purposes, Title/Abstract is really the only one you need to use, at least to get started.

At the bottom of the search form, you must choose one of PubMed, PubMed Neighbours, Nucleotide, or Necleotide Neighbours. Choose PubMed.

If you need more search term rows, change the number in the “Terms” field and click Change! (There is no need to reduce the number of terms–unused rows will be ignored.)

After you save your query(ies), you will get the “dashboard” page of PubCrawler. From the dashboard, you can change a number of options, including how often your receive the results emails, how far back PubCrawler searches, etc.

One option that you may want to change is found in the “Parameters” section. The Relentrezdate parameter controls how far back PubCrawler searches for new “hits.” Since you just looked at all of PubMed to date in Step 2, you can set this parameter to 10 days. (Don’t set it to 5, otherwise, your weekly searches will only go back 5 days.)

That is all you need to do to get PubCrawler working for you. The current default is that queries are run some time on each Sunday, so you can expect to see a results email each Monday morning. Results emails will look something like this:

PubMed results, delivered to your inbox.

Each article in the results email is hyperlinked to its PubMed page, so one click will bring you to the same type of article page on PubMed that you are used to from Step 2. Use the same sort of scanning process you used in Step 2 to identify articles of interest. You will have a far smaller number of articles to scan each week than you did when you searched all of PubMed, because you are only getting results from the last week.

Step Four – follow up on promising leads

Once you find an article of interest, there are a couple of things you may want to do with the article:

  • read the full text of the article, for your own education
  • bring the article to the attention of your doctor
  • contact the author(s) of the article for more information, information on clinical trials, information about getting the treatment outlined, etc.

To read the full text of the article, or to get the full text of the article to bring to your doctor, you will have to find that full text. It may or may not be available online. PubMed has a great, short video on how to find the full text article.

You may also want to contact the author(s) of the article. This is generally quite easy, as the authors’ affiliations are almost always listed on the article page in PubMed. Sometimes, the email address of the appropriate author is listed there as well. This will usually put you in touch easily.

If no contact information is in PubMed, then you will just need to do a little Googling, using the name of the author and their affiliated institution. Or browse to the web site of the affiliated university, hospital, or clinic and search there. Or just call the affiliated institution. This step will likely be easy.

One small note: these articles will usually have multiple authors. It is usually the case that you’ll want to speak to the first author listed. If that leads to a dead end (if, for example, that person has left the affiliated institution, since articles are often published months or years after the fact), try the last listed author next. This will often be the senior person in the group, and that person is most likely to still be at the institution. If that doesn’t work either, just work your way through all the authors.

It’s all up to you now

From here, it’s up to you, your new contacts, and your doctors.

In my case, I spoke with Dr. Matthew Kaufman, who had innovated this new procedure, then discussed it with my doctors at home. Dr. Kaufman recently performed the surgery on me, which went very well. I am now in a waiting period to see if my nerve will regrow as we hope it will.

I was the 42nd person in the world to have this surgery. I would have preferred to have been the 420th person, but there was a downside to waiting. Had I not followed the procedure described here, I would probably not have found out about this surgeon and this procedure for years, if at all, and it might have been too late.

I hope this can help you as well.

Stability or creative chaos?

January 4, 2012

Last night I joined Karthik Hariharan, Curtis Summers, and Greg Vaughn on the very low-tech bar stools at the front of the Dallas Ruby Brigade‘s monthly meeting. We, the panel, were there to discuss our versions of “The Path to Ruby”–topic courtesy of super-organizer Mark McSpadden.

Among the many interesting topics to emerge was the differing circumstances in which each of us were launched into our careers, post-university.

To over-simplify, I found myself contrasting my work-force entry roughly 20 years before my co-panelists.


Stable but deadly?

  • Small-ish number of large employers
  • High demand for programmers: have Computer Science degree? Welcome aboard!
  • Nascent start-up/venture capital ecosystem (so as to be pretty much invisible to new grads)
  • Seemingly clear traditional career path
  • Well-defined, and narrowly defined, roles for CS grads

Early 2000s (ignoring for the moment the post-dot-com, post-9/11 meltdown):

Burning out of control?

  • Large number of small- and medium-sized employers
  • High demand for “rock star” programmers
  • Start-up/VC culture in full bloom
  • Many more career path options
  • “Programmers” often expected to double/dabble in stuff you didn’t learn in school, like visual design

Couple this with the rise of open source software, the ubiquity of the blog-twit-IM-chat-osphere, the trend toward shorter job tenures and multiple careers, and my younger co-panelists have spent their early careers in an environment radically different than the one I found post-university.

Which is better?


We’re all different. From some of my colleagues last night, I heard a little yearning for a more stable work life. Understandable, particularly when you’re young and perhaps starting a family.

Call me a glass-half-full guy, but as much as I appreciated being in demand post-graduation, I also very much like today’s more dynamic, more creative environment. Never before in history have “hackers” been able to be so creative, to bring so much value into the world, as they are right now.

How about you? Pick your mix. What blend of stability vs creative chaos would you pick? 50/50? 75% stability, with a dash of chaos? Vice versa?

IE8 doesn’t like my HTML

August 3, 2011

The problem

jQuery('div#space-for-' + institutionId).html(data);

This seemingly simple line of code was working fine in Chrome and FireFox, but did not result in placing the “data” HTML content in the selected div. Nothing at all changed on the page when in IE8.

I tested it in jQuery-1.4.2 and jQuery-1.6.2 with identical results.

The cure

Finally, I tried inserting a simple HTML string directly in the html() method:

jQuery('div#space-for-' + institutionId).html('<h1>Testing</h1>');

And it worked. Hmmm.

Mind your closing tags

After some careful parsing of my HTML, I discovered a missing </div> tag. That was causing IE8 to ignore the (fairly length) inserted fragment altogether. Add in the missing tag and…voilà…all is well again.