Photo by Adam Fagen

I've just emerged from the worst two weeks of my life. My partner suffered a major stroke, ended up in intensive care, and went through brain surgery. Thankfully she's back at home now and on the road to a full recovery. We had no warning or time to prepare, so I had to learn a lot about how to be an effective caregiver on the fly. I discovered I had several friends who'd been in similar situations, so thanks to their advice and some improvisation here are some hacks that worked for me.

Be there

My partner was unconscious or suffering from memory problems most of the time, so I ended up having to make decisions on her behalf. Doctors, nurses and assistants all work to their own schedules, so sitting by the bedside for as long as you can manage is essential to getting information and making requests. Figure out when the most important times to be there are. At Stanford's intensive care unit the nurses work twelve hour shifts, and hand over to the next nurse at 7pm and 7am every day. Make sure you're there for that handover, it's where you'll hear the previous nurse's assessment of the patients condition, and be able to get support from the old nurse for any requests you have for the new one to carry out. The nurses don't often see patients more than once, so you need to provide the long-term context.

The next most important time is when the doctors do their rounds. This can be unpredictable, but for Stanford's neurosurgery team it was often around 5:30am. This may be your one chance to see senior physicians involved in your loved one's care that day. They will page a doctor for you at any time if you request it, but this will usually be the most junior member of the team who's unlikely to be much help beyond very simple questions. During rounds you can catch a full set of the doctors involved, hear what they're thinking about the case, give them information, and ask them questions.

Even if there are no memory problems, the patient is going to be groggy or drugged a lot of the time, and having someone else as an informed advocate is always going to be useful, and the only way to be informed is to put in as much time as you can.

Be nice to the nurses

Nurses are the absolute rulers of their ward. I should know this myself after growing up in a nursing family, but an old friend reminded me just after I'd entered the hospital how much they control access to your loved one. They also carry a lot of weight with the doctors, who know they see the patients for a lot longer than they do, and often have many years more experience. Behaving politely can actually be very hard when you're watching someone you love in intensive care, but it pays off. I was able to spend far more than the official visiting hours with my partner, mostly because the nurses knew I was no trouble, and would actually make their jobs easier by doing mundane things for her, and reassuring her when she had memory problems. This doesn't mean you should be a pushover though. If the nurses know that you have the time to very politely keep insisting that something your loved one needs happens, and will be there to track that it does, you'll be able to get what your partner needs.

Track the drugs

The most harrowing part of the experience was seeing my loved one in pain. Because neurosurgeons need to track their patients cognitive state closely to spot problems, they limit pain relief to a small number of drugs that don't cause too much drowsiness. I knew this was necessary, but it left my partner with very little margin before she was hit with attacks of severe pain. At first I trusted the staff to deal with it, but it quickly became clear that something was going wrong with her care. I discovered she'd had a ten hour overnight gap in the Vicodine medication that dealt with her pain, and she spent the subsequent morning dealing with traumatic pain that was hard to bring back under control. That got me digging into her drugs, and with the help of a friendly nurse, we discovered that she was being given individual Tylenols, and the Vicodine also contained Tylenol, so she would max out on the 4,000mg daily limit of its active ingredient acetaminophen and be unable to take anything until the twenty-four hour window had passed. This was crazy because the Tylenol did exactly nothing to help the pain, but it was preventing her from taking the Vicodine that did have an effect.

Once I knew what was going on I was able to get her switched to Norco, which contains the same strong pain-killer as Vicodine but with less Tylenol. There were other misadventures along the same lines, though none that caused so much pain, so I made sure I kept a notebook of all the drugs she was taking, the frequency, any daily limits, and the times she had taken them last, so I could manually track everything and spot any gaps before they happened. Computerization meant that the nurses no longer did this sort of manual tracking, which is generally great, but also meant they were always taken by surprise when she hit something like the Tylenol daily limit, since the computer rejection would be the first they knew of it.

Start a mailing list

When someone you love falls ill, all of their family and friends will be desperate for news. When I first came up for air, I phoned everyone I could think of to let them know, and then made sure I had their email address. I would be kicked out of the ward for a couple of hours each night, so I used that time to send out a mail containing a progress report. At first I used a manual CC list, but after a few days a friend set up a private Google Group that made managing the increasingly long list of recipients a lot easier. The process of writing a daily update helped me, because it forced me to collect my thoughts and try to make sense of what had happened that day, which was a big help in making decisions. It also allowed me to put out requests for assistance, for things like researching the pros and cons of an operation, looking after our pets, or finding accommodation for family and friends from out-of-town. My goal was to focus as much of my time as possible on looking after my partner. Having a simple way to reach a lot of people at once and being able to delegate easily saved me a lot of time, which helped me give better care.

Minimize baggage

A lot of well-meaning visitors would bring care packages, but these were a problem. During our eleven day stay, we moved wards eight times. Because my partner was in intensive care or close observation the whole time, there were only a couple of small drawers for storage, and very little space around the bed. I was sleeping in a chair by her bedside or in the waiting room, so I didn't have a hotel room to stash stuff. I was also recovering from knee surgery myself, so I couldn't carry very much!

I learned to explain the situation to visitors, and be pretty forthright in asking them to take items back. She didn't need clothing and the hospital supplied basic toiletries, so the key items were our phones, some British tea bags and honey, and one comforting blanket knitted by a friend's mother. Possessions are a hindrance in that sort of setting, the nurses hate having to weave around bags of stuff to take vital signs, there's little storage, and moving them all becomes a royal pain. Figure out what you'll actually use every day, and ask friends to take everything else away. You can always get them to bring something back if you really do need it, but cutting down our baggage was a big weight off my mind.

Sort out the decision-making

My partner was lucid enough early in the process to nominate me as her primary decision-maker when she was incapacitated, even though we're not married. As it happened, all of the treatment decisions were very black-and-white so I never really had to exercise much discretion, but if the worst had happened I would have been forced to guess at what she wanted. I knew the general outlines from the years we've spent together, but after this experience we'll both be filling out 'living wills' to make things a lot more explicit. We're under forty, so we didn't expect to be dealing with this so soon, but life is uncertain. The hospital recommended Five Wishes, which is $5 to complete online, and has legal force in most US states. Even if you don't fill out the form, just talking together about what you want is going to be incredibly useful.

Ask for help

I'm normally a pretty independent person, but my partner and I needed a large team behind us to help her get well. The staff at Stanford, her family, and her friends were all there for us, and gave us a tremendous amount of assistance. It wasn't easy to reach out and ask for simple help like deliveries of clothes and toiletries, but the people around you are looking for ways they can do something useful, it actually makes them feel better. Take advantage of their offers, it will help you focus on your caregiving.

Thanks again to everyone who helped through this process, especially the surprising number of friends who've been through something similar and whose advice helped me with the hacks above.

Photo by Glenda Sims

Over the last few months, I've been doing a lot more work with name analysis, and I've made some of the tools I use available as open-source software. Name analysis takes a list of names, and outputs guesses for the gender, age, and ethnicity of each person. This makes it incredibly useful for answering questions about the demographics of people in public data sets. Fundamentally though, the outputs are still guesses, and end-users need to understand how reliable the results are, so I want to talk about the strengths and weaknesses of this approach.

The short answer is that it can never work any better than a human looking at somebody else's name and guessing their age, gender, and race. If you saw Mildred Hermann on a list of names, I bet you'd picture an older white woman, whereas Juan Hernandez brings to mind an Hispanic man, with no obvious age. It should be obvious that this is not always reliable for individuals (I bet there are some young Mildreds out there) but as the sample size grows, the errors tend to cancel each other out.

The algorithms themselves work by looking at data that's been released by the US Census and the Social Security agency. These data sets list the popularity of 90,000 first names by gender and year of birth, and 150,000 family names by ethnicity. I then use these frequencies as the basis for all of the estimates. Crucially, all the guesses depend on how strong a correlation there is between a particular name and a person's characteristics, which varies for each property. I'll give some estimates of how strong these relationships are below, and I link to some papers with more rigorous quantitative evaluations below.

If you are going to use this approach in your own work, the first thing to watch out for is that any correlations are only relevant for people in the US. Names may be associated with very different traits in other countries, and our racial categories especially are social constructs and so don't map internationally.

Gender is the most reliable signal that we can gleam from names. There are some cross-over first names with a mixture of genders, like Francis, and some that are too unique to have data on, but overall the estimate of how many men and women are present in a list of names has proved highly accurate. It helps that there are some regular patterns to augment the sampled data, like names ending with an 'a' being associated with women.

Asian and Hispanic family names tend to be fairly unique to those communities, so an occurrence is a strong signal that the person is a member of that ethnicity. There are some confounding factors though, especially with Spanish-derived names in the Phillipines. There are certain names, especially those from Germany and Nordic countries, that strongly indicate that the owner is of European descent, but many surnames are multi-racial. There are some associations between African-Americans and certain names like Jackson or Smalls, but these are also shared by a lot of people from other ethnic groups. These ambiguities make non-Hispanic and non-Asian measures more indicators than strong metrics, and they won't tell you much until you get into the high hundreds for your sample size.

Age has the weakest correlation with names. There are actually some strong patterns by time of birth, with certain names widely recognized as old-fashioned or trendy, but those tend to be swamped by class and ethnicity-based differences in the popularity of names. I do calculate the most popular year for every name I know about, and compensate for life expectancy using actuarial tables, but it's hard to use that to derive a likely age for a population of people unless they're equally distributed geographically and socially. There tends to be a trickle-down effect where names first become popular amongst higher-income parents, and then spread throughout society over time. That means if have a group of higher-class people, their first names will have become most widely popular decades after they were born, and so they'll tend to appear a lot younger than they actually are. Similar problems exist with different ethnic groups, so overall treat the calculated age with a lot of caution, even with large sample sizes.

You should treat the results of name analysis cautiously – as provisional evidence, not as definitive proof. It's powerful because it helps in cases where no other information is available, but because those cases are often highly-charged and controversial, I'd urge everyone to see it as the start of the process of investigation not the end.

I've relied heavily on the existing academic work for my analysis, so I highly recommend checking out some of these papers if you do want to work with this technique. As an engineer, I'm also working without the benefit of peer review, so suggestions on improvements or corrections would be very welcome at pete@petewarden.com.

Use of Geocoding and Surname Analysis to Estimate Race and Ethnicity – A very readable survey of the use of surname analysis for ethnicity estimation in health statistics.

Estimating Age, Gender, and Identity using First Name Priors – A neat combination of image-processing techniques and first name data to improve the estimates of people's ages and genders in snapshots.

Are Emily and Greg More Employable than Lakisha and Jamal? – Worrying proof that humans rely on innate name analysis to discriminate against minorities.

First names and crime: Does unpopularity spell trouble? – An analysis that shows uncommon names are associated with lower-class parents, and so correlate juvenile delinquency and other ills connected to low socioeconomic status.

Surnames and a theory of social mobility – A recent classic of a paper that uses uncommon surnames to track the effects of social mobility across many generations, in many different societies and time periods.

OnoMap – A project by University College London to correlate surnames worldwide with ethnicities. Commercially-licensed, but it looks like you may be able to get good terms for academic usage.

Text2People – My open-source implementation of name analysis.

Photo by Julian Schroeder

I just spent quite a few hours tracking down a subtle problem with OpenCV's new Java bindings on gcc platforms, like my Ubuntu servers. The short story is that the default for linked symbols was recently changed to hidden on gcc systems, and the Java native interfaces weren't updated to override that default, so any Java programs using native OpenCV functions would mysteriously fail with an UnsatisfiedLinkError. Here's my workaround:

--- a/cmake/OpenCVCompilerOptions.cmake
+++ b/cmake/OpenCVCompilerOptions.cmake
@@ -252,8 +252,8 @@ set(OPENCV_EXTRA_EXE_LINKER_FLAGS_DEBUG   "${OPENCV_EXTRA_EXE_LINKER_FLAGS_DEBUG
 
 # set default visibility to hidden
 if(CMAKE_COMPILER_IS_GNUCXX AND CMAKE_OPENCV_GCC_VERSION_NUM GREATER 399)
-  add_extra_compiler_option(-fvisibility=hidden)
-  add_extra_compiler_option(-fvisibility-inlines-hidden)
+#  add_extra_compiler_option(-fvisibility=hidden)
+#  add_extra_compiler_option(-fvisibility-inlines-hidden)
 endif()
 

The tricky part of tracking this down was that nm didn't show the .hidden attribute, so the library symbols appeared fine, it was only when I switched to objdump after exhausting everything else I could think of that the problem became clear.

Anyway, I wanted to leave some Google breadcrumbs for anyone else who hits this! I've filed a bug with the OpenCV folks, hopefully it will be fixed soon.

Photo by Leo Reynolds

External framework problems in Go – Handling dependencies well is extremely hard, and can lead to insane yak-shaving expeditions like this when things go wrong. It's like an avalanche – changing versions on one library can impact several others, so you have to update or downgrade those too, and suddenly you're facing an ever-increasing amount of work just to get back to where you were!

D-wave comparison with classical computers – I don't know enough about quantum computing problems to comment on the details of the argument, but it's awesome to see such a deep technical dive as an instant blog post, rather than having to wait months for a paper.

Blogging is dead, but have we fixed anything? – "I find my blogging here to be too useful to me to stop doing it" sums up why I'm still working in a now-archaic medium!

What statistics should do about big data – "[Statisticians] want an elegant theory that we can then apply to specific problems if they happen to come up." That's been exactly my experience, and why I've never encountered statisticians as I've followed my curiosity to new problems data. The article this is in response to contains the assumption that 'funding agencies' have driven the CS takeover of data processing, but, despite a lot of the founders having roots in academia, almost all the innovations I've seen have been incubated in commercial environments.

The hidden sexism in CS departments – A portrait of managerial cluelessness when dealing with a nasty incident. Even if each occurrence is comparatively minor, it's the steady drip-drip of unwelcoming behavior that drives non-stereotypical geeks out of our world.