Monthly Archives: January 2013

Journal Club: Information for decision making from imperfect national data

A nice connection from last week’s journal club paper to this week’s: the errors in health information system data. Last week was about correcting the bias from missing individuals. This week is about correcting the bias from missing facilities.

From the key figure, it looks like missing individuals bias things more:

Gething et al. BMC Medicine 2007 5:37   doi:10.1186/1741-7015-5-37

1 Comment

Filed under global health

k-NN in SPARQL?

Is a SPARQL query capable of finding the k nearest neighbors for several vectors simultaneously? I don’t think so, but I’ve been wrong before. Tell me on Stack Overflow.

Comments Off

Filed under machine learning

Journal Club: Sampling-Based Approach to Determining Outcomes of Patients Lost to Follow-Up in Antiretroviral Therapy Scale-Up Programs in Africa

This week’s paper for journal club is short and has a nice figure:

Geng et al, Sampling-Based Approach to Determining Outcomes of Patients Lost to Follow-Up in Antiretroviral Therapy Scale-Up Programs in Africa
naive_and_corrected

It looks like this corrected estimate is quite different than the uncorrected version!

I think the mathematics involved have an extended treatment in this work referenced by Geng et al: Addressing an idiosyncrasy in estimating survival curves using double-sampling in the presence of self-selected right censoring

2 Comments

Filed under global health

MLK’s address to APA

Today is Martin Luther King Day in the US, a civil rights holiday. I heard a recording of this address by King to the 1967 meeting of the American Psychological Association, and then couldn’t find a copy of the text… until now, when I searched the web with two critical typos in the search terms.

The Role of the Behavioral Scientist in the Civil Rights Movement
By Martin Luther King Jr.


There are certain technical words in every academic discipline which soon become stereotypes and even clichés. Every academic discipline has its technical nomenclature. You who are in the field of psychology have given us a great word. It is the word maladjusted. This word is probably used more than any other word in psychology. It is a good word; certainly it is good that in dealing with what the word implies you are declaring that destructive maladjustment should be destroyed. You are saying that all must seek the well-adjusted life in order to avoid neurotic and schizophrenic personalities.

But on the other hand, I am sure that we will recognize that there are some things in our society, some things in our world, to which we should never be adjusted. There are some things concerning which we must always be maladjusted if we are to be people of good will. We must never adjust ourselves to racial discrimination and racial segregation. We must never adjust ourselves to religious bigotry. We must never adjust ourselves to economic conditions that take necessities from the many to give luxuries to the few. We must never adjust ourselves to the madness of militarism, and the self-defeating effects of physical violence.

(all)

Comments Off

Filed under general

ipython autoreloading

This will be useful.

Comments Off

Filed under Uncategorized

Power of SPARQL?

Is a SPARQL query computationally powerful enough to test (s,t)-connectivity? I don’t think so, but I don’t understand the mysterious PropertyPath, and even without it, I’m not sure. Tell me on Stack Overflow.

1 Comment

Filed under combinatorial optimization

Journal Club is Back

Or actually, I am back. Back to facilitating the post-graduate fellowship (PGF) journal club. Here is what we are reading this week, Impact assessment of malaria vector control using routine surveillance data in Zambia: implications for monitoring and evaluation, which is a highly accessed article according to the Malaria Journal website. Is it also highly accessible? We shall see. Any wisdom on this that I can pass on to the fellows is welcome.

Figure 1. Estimated operational ITN distributions by district in Zambia from 2003–2010, representing percentage of district households receiving 3 ITNs per household (HH) in overlapping 3-year intervals (MoH, 2010).

Figure 1. Estimated operational ITN distributions by district in Zambia from 2003–2010, representing percentage of district households receiving 3 ITNs per household (HH) in overlapping 3-year intervals (MoH, 2010).

1 Comment

Filed under global health