Monday, November 3, 2014

Texas Democrats fight hard to neutralize mid-term malaise: What do EV results portend?

Fewer people voted during early voting period in 2014 compared to 2010:

At the end of 2014 early voting period, roughly 1.715 million votes have been cast in 2014 in the top 15 counties in the state of Texas, and this is roughly similar to 1.731 million votes cast during the early voting period in 2010 for the same counties. 

Although the total number of registered voters (in these 15 counties) in 2014 was higher in 2014 (8.978 million vs 8.339 million), a slightly lower percentage of people cast their ballot during EV period (19.11 % vs 20.76 %).

Therefore, voter malaise is real.

Does this mean Republicans have a big advantage?

No. Contrary to popular wisdom, the republican advantage due to slightly lower turn-out may not be high.  In most of the counties that democrats won in 2010, the EV percent is roughly similar in 2014, indicating that there was not a huge drop off in democratic voters.  There is also not a big uptick in the EV % in the counties won by republicans.

Early voting compared to 2010 in the top 15 counties in Texas.  The horizontal axis indicates the percentage point margin that democrats had over republican in each of the counties.  Positive numbers mean democrats won those counties, and negative numbers mean democrats lost these counties.  The vertical axis indicates the early vote differential in 2014 compared to 2010.  Positive numbers indicate more voted in 2014 than in 2010.  The size of the bubble indicates the size of the county.  Of the two largest counties, Tarrant county might turn blue this time, and Harris county might turn red. (see text)

Two large counties stand out.  First, Harris county which former Mayor Bill White - democrat, won narrowly in 2010, has seen a sizable drop in EV percentage (5 %), and most likely Harris will swing to republicans in 2014.   Second, EV in Tarrant county is up by 3% compared to 2010, and this may be because the democratic gubernatorial candidate Ms. Wendy Davis is from Tarrant county.  So, this district may swing democratic in 2014.  The swings in these large counties may offset each other.
Based on the analysis of EV in the top 15 Texas counties, it looks like that the democrats may have kept their head above the water in the early voting.

Get out the vote:

So, it boils down to who shows up at the voting booth on election night.  It appears that the Texas democrats are better organized than usual this time around, and there is no evidence of the 2010 tea-party wave fueled by anti-Obamacare sentiments roused by candidates such as Ted Cruz.  

In the absence of any wave, if each of the 30,000 volunteers that Ms. Wendy Davis’s campaign claims to have on the field, bring just 10 otherwise unlikely voters to the polling booth, this could be a very close election than most predictions.

Let us see tomorrow.

Tuesday, October 28, 2014

Epic right wing fear mongering about Ebola

Right wing hysteria about Ebola has reached a new high.  It is not just the usual, uninformed talk show blowhards who are whipping up the hysteria, but respected senators, and senior leaders of the republican party heartily take part in it.

Dr. Rand Paul – a physician,  proclaimed in the Glenn Beck show that, “…“I do think you have to be concerned. It’s an incredibly transmissible disease that everyone is downplaying, saying it’s hard to catch. Well, we have physicians and health workers who are catching it who are completely gloved down and taking every precaution and they’re still getting it.” 

Really, Dr. Paul? Physicians and health workers who are completely gloved down and taking every precaution are still getting it?  It is this monstrous disease that we know nothing about, and we are underplaying it?  The respected NEJM, published a paper on Oct 16th from the WHO team after analyzing 4000+ cases from the current epidemic, and the authors unequivocally state:

Although the current epidemic of EVD in West Africa is unprecedented in scale, the clinical course of infection and the transmissibility of the virus are similar to those in previous EVD outbreaks. The incubation period, duration of illness, case fatality rate, and reproduction ratio R 0 are all within the ranges reported for previous EVD epidemics.

Time between the exposure and onset of symptoms based on data from all three west African countries.  The gray bars indicate the frequency of candidates, and the thick red line is the model fit.  Symptoms start at 5 +/- 4 days after exposure.  (Courtesy; NEJM)

Thanks to the heroic work of physicians and scientists, we know a lot about the current Ebola epidemic,and this epidemic is biologically no different to other Ebola epidemics of the past.  What we need is effective control measures to bring the reproduction rate of Ebola to below 1.  That requires personal protection equipment (PPE), and other logistical support for the health-care workers in West Africa – in short aid.  Well, what is Senator Paul’s position on foreign aid?  He wants to eliminate it – gradually.  Yes, the senator who wants to eliminate foreign aid, is the one who is keen on whipping up panic based on unproven hypotheses.

Other prominent right wing panic mongers include republican senate hopefuls such as Scott Brown demanding a blanket travel ban , Louisiana Senate candidate Bill Cassidy accusing CDC of downplaying ebola risk, and Mike Rounds of South Dakota insisted on a travel ban.  The Republican party establishment from Boehner, Mitch McConnel, Chris Christie, etc. have all hyper-ventilated about how to respond to ebola:  they have insisted on a blanket travel ban or on quarantine. 
The prestigious NEJM strongly argued against calls by republican (and some democratic) calls for quarantine.  In plain language, NEJM stated the following about calls for quarantine, “… is not scientifically based, is unfair and unwise, and will impede essential efforts to stop these awful outbreaks of Ebola disease at their source, which is the only satisfactory goal

Obama deserves credit for keeping a cool head amidst all this hyperventilation. 

The facts are: (a) Ebola is difficult to spread.  It spreads via droplet transmission and is not airborne; (b) If the patient is not symptomatic, the subject cannot transmit the disease (Mr. Duncan’s family was not infected during the early stages of the development of the disease, whereas the nurses who took care of him when he had the full-blown symptoms did); (c) US has tremendous knowledge about Ebola and (d) US has the necessary healthcare infra-structure to handle it.

Both nurses who were infected at the Dallas hospital have now been cured of Ebola.  A couple of dozen people who came into contact with Mr. Duncan, who were monitored, are now beyond the window for developing the disease, and have been cleared.

Does this mean that we will not have any more Ebola outbreaks in the US?  No.  It is just that we know how to deal with it.  

Let us follow the scientists, and not give in to irrational fear whipped up for crass, short-term political advantage.  

Sunday, October 26, 2014

Some more thoughts in Texas Early voting

Six days into the early voting, it is clear that the percent of early voting in 2014 in the 15 most populated counties in the state of Texas is roughly comparable to 2010.

A total of about 43,000 additional people have voted after six days of early voting compared to 2010.

Data summarized from the Texas Secretary of State EV totals published on 10/25/2014

I think it is safe to make a few observations after six days of early voting, including a weekend day.

1) It is a particularly negative sign for the democratic party that the largest county in Texas, Harris county is lagging in the early voting so far.  In 2010, Rick Perry and Bill White essentially split the number of votes cast.  It may not be a big deal.  But, even if a couple of percent of the usual democratic voters do not show up at the polls in Harris county, the deficit may be difficult to overcome with increases in other counties.  So, if you have friends who live in Texas call them and ask them to cast their ballot.

2) While there is a net increase in the number of votes in the Hidalgo county (compared to 2010), there is a net drop in El Paso county.  Both counties broke strongly in favor of democrats in 2010. So, that is an area to watch.

3) Lastly, this may be the silver lining for the Texas democrats.  There is no evidence to suggest that the republican efforts to increase their turnout has paid any dividends.  Thus far, it looks like 2014 may be yet another Texas mid-term election where fewer than 50% of Texans will participate in.  In the past, such elections overwhelmingly broke for the republican party. In this cycle, it looks like that the democrats have upped their game in canvassing, and micro targeting of their voters.   It remains to be seen, if these efforts are sufficient to win in November.

Saturday, October 25, 2014

Texas : Slowly trending blue?

To be a liberal in Texas (especially in the suburbs) can be dispiriting.  There is not a single democrat elected for statewide office.  For example, in TX-22 congressional district (where I live), Republicans traditionally have enjoyed a 2-1 advantage in voting compared to democrats, and that sort of advantage is fairly common across Texas.  For many positions, there are not even qualified democratic candidates on the ballot.  Well, that was the recent past.

I volunteered today to block-walk canvassing for Wendy Davis, who is running for the governor of Texas.   It was an interesting experience.  A couple of things were worth noting.

1)      I walked in and a gentleman gave me instructions about canvassing.  I walked out with a detailed map of potential democratic voters, and when I returned three hours later, the data was tabulated, and uploaded to a server.  This operation was run by “Battleground Texas”, and the level of sophistication was something that I have never seen before in Texas.

2)      While the volunteers were quite enthusiastic, it remains to be seen, if these efforts translate to votes.  

Based on the early voting results so far (from the top 15 counties), it looks like 2014 early voting trends are very similar to the most recent mid-term election (2010).  Five days into early voting, four of the top five most populated counties, (Harris, Dallas, Bexar, Travis) all show an early voting trend that is very similar to 2010.  However, newspaper reports suggest that early voting is up in Tarrant, and Hidalgo counties, compared to four years ago.

Graph generated with data from Texas Secretary of State's website.

In 2010, Rick Perry got about 20% of the registered voters to beat the democratic challenger Bill White who got 15% of the registered voters.  Compared to four years ago, Wendy Davis has run a more visible campaign, and definitely has a better ground operation. 

Whether all these factors are sufficient to put Wendy Davis in the governor’s office remains to be seen.  In any event, the increasing organizational capacity of the democrats, and the changing demographics of Texas, should concern the Republicans.  

Less than 40% of the registered voters cast a ballot in 2010.  Even if a small percentage of the unlikely voters turn up to vote democratic in this election, it might make for an interesting night on November.   

Monday, October 13, 2014

Ebola – A warning for India

While there are some valid concerns regarding the health care given to the first US Ebola victim, Mr. Duncan1, none would argue the fact that the care received by Mr. Duncan was far superior to what most victims in the West African nations of Liberia, Guinea, and Sierra Leone.  After a plodding start, the Atlanta based Center for Disease Prevention and Control (CDC),  has disseminated clear, simple-to-understand, facts about Ebola to the US public (some would argue that going overboard), and updates its website daily2.   The potency of the Ebola virus is underscored by the infection of a healthcare worker in Dallas who cared for Mr. Duncan. 

Ebola infections/death in West African Nations (Graphic from NYTimes).  Note the ‘J’-curve with a rapid increase particularly in Libeia

Simply put, Ebola is a deadly disease, and kills nearly 40% of those who are infected by it.  The current rate of Ebola infections and deaths, in Liberia, and Guinea may just be the beginning.  The following graphic from NY Times article, highlights an accelerating Ebola epidemic in Liberia [3].  

The unforgiving mathematics behind Ebola epidemic:

 Many have difficulty appreciating exponential growth factors.  Think about this story for a second.   

There is a pond with a single lotus flower, and every day the number of lotus flowers doubles.  An eighth of the pond is filled with lotus flowers.  How many days will it take to have the whole pond covered with lotus flowers?

Mathematical models can predict if an epidemic can spread or not.  If each person who has been infected, on the average, spreads to more than one person the epidemic will accelerate.  This number is called the ‘reproduction number (Ro)’, and when this number falls below unity, the epidemic will die out.  On Sep 2nd 2014, based on the data available until Aug 2014, Swiss researcher, Christian Althaus, published a paper in PLOS, which estimated that the basic reproduction number for infections in Guinea, Sierra Leone, and Liberia were 1.51, 2.53, and 1.59 respectively.   Encouragingly, his models predicted that the control measures taken in Guinea and Sierra Leone, may have brought the ‘effective reproduction number (Re)’, to unity by end of May and July respectively, and Liberia needs to take more control measures.  The graphs shown in the NY Times article seem to support his predictions, Guinea and Sierra Leone, seem to have more of linear increase in infections from Sep-Oct, whereas during the same period the number of infections in Liberia appear to have accelerated.

It is essential to bring the number of infected to a manageable level quickly.  CDC estimates that if we can get 15000 people to treatment centers in West Africa, the epidemic will be contained.  If we wait for a week, that number will quadruple to 40000, and in a month will increase by six fold to 100,000.  There is simply no such capacity in any of these countries.  This is a race against time, in which, at present Ebola is winning. 

How well is India prepared for an epidemic (Ebola or not) ?

In 2013, T.Dikid et al. reviewed the emerging and re-emerging infections in India in the last two decades.  Five of the eight infections were viral in nature.

Figure 2: Recent epidemics in India from 1991-2011 (From T. Dikid et al. Ref 6. below)

The infections ranged from Cholera, Plague, Chikungunya, H1N1, to Diptheria.   In fact, Chikungunya infected almost the entire nation (22 states), and can cause debilitating joint pain for up to six months in those who are infected.

Almost all of India suffered from the debilitating spread of Chikengunia.  The actual incidence of the disease is likely underreported.

The Economist reports that, “Peter Biot - the Belgian microbiologist who discovered Ebola in 1976 in the Congolese rainforest, told the Observer newspaper this month that he is especially worried about the state of public health in India as follows:

"Doctors and nurses in India, too, often don't wear protective gloves. They would immediately become infected and spread the virus." 

Peter Biot, the epidemiologist who discovered Ebola in 1977 in Congo. (From BBC news)

Others have suggested, " it would be a problem that India has only two facilities capable of testing for the virus. Moreover the prevalence of malaria, dengue and other fever-inducing illnesses in India could make it especially difficult to isolate those who might show early onset of Ebola, which has similar symptoms”.

Indian government’s response to date on any of these epidemics is underwhelming at best.  The number and frequency of epidemics in India are probably vastly under reported.    Indian healthcare system is woefully inadequate to meet any epidemic.  It is estimated that 600 million people in India lack access to clean sanitation, or access to healthcare.  Increasingly, most Indians have to buy bottled water to get drinkable water.  An Ebola or for that matter any other epidemic, with just a few thousand infected people in a highly populated urban area, can easily overwhelm the healthcare infra-structure of most cities in India.  

In contrast to the CDC website, the website of the National Center for Disease control in India is uninspiring.  The method of collecting information about potential Ebola infections is through a paper form, completed by hand.  For a country that boasts millions of software engineers, there is not even an online registry.  The website has little information or guidance for private practitioners – the bulk of Indian medical doctors who care for the India's sick.  There is little information about training for Indian nurses in public as well as private hospitals. 

India should actively get involved in the current management of Ebola crisis in West Africa.  It should study the methods that are used to contain the disease, and build the necessary infra-structure – personnel, equipment, knowledge, management, and process controls.  The private sector, non-governmental organizations (NGOs) and activists should turn up the heat on the government and ask tough questions about preparedness and transparency.

Otherwise, India is in for a rude shock for an epidemic (Ebola or not) that is bound to strike.