Category Archives: Epidemiology

The Speed of Spread

I have opened my presentation slides on pandemics (nearly 400 slides) to find this small subset. This is a graphic representation of how quickly the Spanish Flu spread across the US in 1918 in just a few weeks. Remember, we didn’t have air travel at the time, so long distance travel was by train and recall that the Model T Ford had only come on the market in 1908, so travel by automobile was still relatively new. This is a good illustration of how quickly things can move through the US, but when you think about how much easier and quicker transportation is today, it’s likely things will spread much more rapidly.

Why Italy is a New Level of Concern

From February 26, 2020

I will try to explain why my views of the impact of COVID-19 dramatically changed due to the cluster of cases in Italy. Of course, I’ll use a graph to try to make the concepts clearer.

No photo description available.

One of the terms we use in epidemiology is called the incubation period. That is the time from which a person is first exposed to a communicable disease until they start showing symptoms of the disease. They also can start spreading the disease later during this time period. I’ve represented that with three horizontal bars with gradients somewhat indicating that they get more likely to spread infection as they get closer to the time they are identified as a patient. The incubation period for this disease is thought to be 14 days, which is the lengths of the bars I used.

The first two cases (purple) were a husband and wife who were tourists from China. It is thought that he caused her illness as well as well as 12 others. There is only one other case (orange) that shows up on 2/7. No more are identified until 2/21, when over the course of five days we have seen an additional 319 cases so far. That is one of the things that is so alarming to me. There are three possibilities that I can think of:

1. There were other cases that haven’t been identified that spread the disease. This could be good or bad, depending on how much of a proportion of the population are asymptomatic spreaders.

2. These three individuals combined are somehow responsible for the next 319 cases (so far), or worse, maybe just 1 or 2 of them are. If this is the case, that means at a minimum that one person was responsible for over 100 cases, and if it’s only 1 or 2 of them causing most of the spread, a number much higher.

3. In some individuals, the incubation period may be longer than 14 days.

People like this are called super spreaders. For comparison, during SARS it is thought that the majority of the disease in Singapore was spread by five of these types of individuals, the highest causing 76 cases.

This is a concept I covered earlier called the reproductive number, or R0 of a disease. It’s the average people infected by an individual that is infected. The WHO estimate had been 1.4-2.5, but a recent analysis (link below) of various literature indicates that the value is 1.4 to 6.49 and with an average of 3.28. For comparison, the Spanish Flu of 1918 was 1.2-3.0

The reproductive number of COVID-19 is higher compared to SARS coronavirus

Combining this with what I had described earlier of the differences between how I expected this to travel in the West versus the East is very alarming. We are looking at a disease that could easily rival 1918 in scope. The mortality numbers I have calculated assuming a 30% attack rate (the percentage of the world population that get infected), which is the estimated rate from 1918, are simply shocking. However, we do not know exactly how many people will be infected, so this could be far worse or far better.

I’m sorry this one was probably a little more technical and harder to grasp but some concepts but it’s late and I don’t have the energy to wordsmith. I will try to answer any questions though as I have time.

Disclaimer: This commentary is my own interpretation and does not represent the analysis by the government or my employer. The data is from the Johns Hopkins University’s Center for Systems Science and Engineering.

Not If, But When

From February 26, 2020

No photo description available.

I had been hopeful that COVID-19 could be contained because of some of the differences between eastern and western philosophies and social practices.. My update on Feb 24 had a change in tone. Now that we have almost 5 days of outbreak data in Italy, I am very concerned about the ramifications of this globally.

No photo description available.

This mirrors what was stated by Nancy Messonnier, the director of the CDC’s National Center for Immunization and Respiratory Diseases, “It’s not a question of if but rather a question of when and how many people will have severe illness.”

In short, start preparing. What transpires in Italy will be a good indicator of what will happen in other Western countries.

Please note that while the graphs of the diseases globally and in Italy use the same methodology, they have very different scales to represent the data. Don’t try to compare them to each other for that reason.

Disclaimer: This commentary is my own interpretation and does not represent the analysis by the government or my employer. The data is from the Johns Hopkins University’s Center for Systems Science and Engineering.

Italian Alarm

From February 24, 2020

File:Flag of Italy (1946–2003).png - Wikimedia Commons

While the global picture continues to improve, I do have some concerns about COVID-19 at national levels. My biggest concern is the 215 cases in Italy.

One advantage of a strong central government in an outbreak situation is the ability to quickly enact policies to mitigate the spread of the virus. The rate at which hospitals were built in China was truly amazing as well.

Another reason I was a little less concerned about the disease in Asia is that respiratory hygiene measures such as mask use are now a part of the culture. This had started in Japan in 1918 during the Spanish Flu for obvious reasons, was reinforced by the Great Kanto Earthquake which led to a massive inferno in the city which resulted in smoke and ash that remained in the air for weeks. The influenza pandemic of 1934 further made mask use a common practice.

This was also spread by eastern medicine and philosophy, where “qi” is considered an essential element of health, which is tied to concepts of air, atmosphere, odor, etc. The use of masks quickly spread across eastern Asia for these reasons.

While eastern countries embrace the common good, in the west individualism and libertarian ideas make dealing with disease spread much more difficult. For example, think about how the antivax movement is causing a resurgence of measles in the US, which has been completely eliminated in 2000 as a result of vaccination efforts. Last year, there were 1282 cases in the US, which is a direct result of the antivax movement.

What people don’t seem to understand is that globally, measles has a case fatality rate of 15%, and about 0.2% in the US. In addition, about 25% of those infected with measles develop neurological damage.…/pubs/pinkbook/downloads/meas.pdf

The other point I will add on this topic is that herd immunity is crucial in preventing disease spread. Among those vaccinated for measles, about 10% do not develop adequate antibody protection, and thus are susceptible. Herd immunity protects both that group of the population as well as those who have true medical contraindications to the vaccine.

Hence, the West is likely less prepared to deal with a large cluster of cases in some countries because of these philosophical differences and resistance to some basic public health interventions.

Disclaimer: This commentary is my own interpretation and does not represent the analysis by the government or my employer. The data is from the Johns Hopkins University’s Center for Systems Science and Engineering.

Don’t Use a Short Series of Data

From February 23, 2020

No photo description available.

New cases are back at a level yesterday that is more what I expected. It’s hard to say why there was the three days of considerably lower cased.

It’s also troubling that the case fatality rate has been slowly climbing.

This is a good example of why it’s never a good idea to rely on a very short series of data in a long stretch to identify trends.

Disclaimer: This commentary is my own interpretation and does not represent the analysis by the government or my employer. The data is from the Johns Hopkins University’s Center for Systems Science and Engineering.

COVID-19 Big Jump

from February 13, 2020

Don’t be alarmed at first glance.

No photo description available.

There appears to be a large jump in cases overnight. This is strictly due to the way that cases are being identified in Hubei province in China. Most testing has been through RNA tests, but results can take days. Hubei province decided to use CT scan to look for lung infections in order to start treatment earlier. This accounts for the large jump in cases. A similar thing occurred in the US in 1993 when the definition used for surveillance for AIDS changed, resulting in what looked like a massive jump that year.

A change in the testing methodology impacts two important variables: sensitivity and specificity. There are technical definitions for these but I’ll try to describe them in a way that makes them easier to understand.

In this case, switching to CT scan increased sensitivity. This essentially means that the test is more likely to capture cases. That’s why there is a large rise in the number of cases.

The other things that happens in this case is a decrease in specificity. A CT scan will identify any type of lung infection, not just from COVID-19. This is what is also known as a false positive test.

The takeaway is that this not something to get more concerned about. It makes sense to try to treat those at highest risk.

Disclaimer: This commentary is my own interpretation and does not represent the analysis by the government or my employer. The data is from the Johns Hopkins University’s Center for Systems Science and Engineering.

A Perfect Storm in Haiti and Surge Capacity Lessons for the U.S.

Haiti is still reeling from the massive earthquake with over 1 million people living in tent cities. The unfolding cholera outbreak is causing much more misery. Now, Hurricane Tomas is heading that direction. What is the synergy of destruction that is likely to come if Tomas hits Haiti as well?

As of October 27, there were 4,722 cases of cholera and 303 deaths according to the WHO. Realistically though, both the number of cases and deaths is likely much higher due to the difficulties of collecting this data in a country that has had its entire infrastructure so incredibly ravaged.

The health care system there now (at least what there is of it) is  overwhelmed by the number of patients with cholera. There are no beds remaining. Patients have to lie on the floor or outside on the ground. This actually should be a major warning shot across the bow of the US health care infrastructure.

Surge capacity is “The ability to obtain adequate staff, supplies and equipment, structures and systems to provide sufficient care to meet immediate needs of an influx of patients following a large-scale incident or disaster.” Essentially it is describing the ability to care for a large number of people.

Therein lies the problem in the US. There are a decreasing number of bed days available in hospitals. That is essentially a count of the number of available beds each day around the country. That is occurring at the same time that the baby boomer generation will be using MORE of those bed days as they age. (Healthc Financ Manage. 2007 Jun;61(6):114-5.)

Some people might wonder why that is a problem. Won’t the invisible hand of economics resolve the issue? Take a look at where the US stands in comparison to other countries on the number of hospital beds per 1,000 population. It should be noted that this represents the number of licensed, staffed beds. It isn’t looking at usage rates, which paints a far more ominous picture for the future of US health care. The US is far behind other industrialized countries on this measure.

During the winter (pneumonia and influenza season), hospitals can easily reach capacity, without the impact of a disaster. That is the most worrisome part of this problem. The US health care system keeps losing its capacity to care for patients with routine problems. If a major event were to unfold, the system would not likely be able to absorb the impact of mass casualties.

Most people don’t know what hospital care is like until they are older and need it for themselves. Labor and delivery isn’t a good comparison because hospitals fund and market these areas differently than the rest of the facility. Today, staff feel overworked and this can lead to mistakes or at least a less than desirable experience for the patient.

A good analogy of what is happening in health care is the airline industry. Think about what happens when airlines consolidate. Routes are reduced, smaller aircraft are used, and each plane is filled beyond capacity. Travelers now have to deal with minimal leg room, eight peanut meals, baggage fees, and increased chances of getting bumped from a flight. There will be a number of problems in the health care industry as well, although it will be hard to predict exactly what the parallels could be drawn between “riding the friendly skies” and “riding the friendly gurneys.” The equivalent of getting bumped in a hospital though could be tragic.

Is that realistic though? Could hospitals really become that overwhelmed in the US? One only has to look back at the 1918 Spanish Influenza pandemic for a model. Johns Hopkins University Hospital CLOSED to anyone but staff and students. Hospitals routinely turned away patients in Philadelphia. Two military base hospitals at the time paint a grim picture. Camp Devens (near Baltimore) had a hospital that was built to hold 1,200 patients but had OVER 6,000 during the worst of the outbreak. Camp Grant (near Rockford, Ill.) went from 610 to 4,102 patient in only six days. An event on that scale would cripple the health care system today. US hospital capacity is in critical condition.

Back to Haiti

The refugees in Haiti are facing something terrible. Fortunately, it doesn’t look at this time like Tomas will make landfall in the area. However, they have a 50% chance of having to deal with sustained tropical storm (>=39 mph) surface winds and rain. Imagine trying to live that way in a tent or shack. The minimal housing that they have could easily be destroyed. The government has already suggested evacuating these sites, but that will obviously not be possible for a number of people in these areas.

This whole scenario could easily increase the rates of cholera in the area. More of the environment could become tainted and  sanitation and clean water facilities could be compromised. This could give an entirely new meaning to the concept of “a perfect storm.”

Haiti Disaster Revisited

I can’t help but be saddened by the events unfolding in Haiti but receiving very little media attention. Months after the horrible devastation and loss of life as a result of the earthquake a few months ago, the survivors are about to face the scourge of a massive cholera outbreak.

The first major piece in the media that caught my attention was by Donald McNeil of the New York Times. The most disturbing part of that story reads “While normally less crowded than the cities, the Artibonite is now host to thousands of earthquake refugees. Many are crowding in with relatives and drinking from the local St. Marc River, into which raw sewage also flows. The area is prone to flooding in the rainy season, which is now in progress.”

I believe we are on the verge of a biological disaster of unparalleled proportion in recent history in North America in terms of the numbers of deaths.  There are still over 1 million Hatians living in refugee camps as a result of the earthquake. These camps are ideal breeding grounds for the causative agent of cholera, Vibrio cholerae. Untreated cholera can have mortality rates of 25-50%. These deaths are primarily due to the extensive dehydration that results from the massive vomiting and diarrhea caused by this organism. Rehydration is the most effective treatment and can reduce mortality to around 1%. The biggest challenge in most cholera outbreaks is that “delivery in remote areas remains difficult during epidemic periods.” (CDC)

Just one day after this story more troubling news emerged. Cases of cholera have been confirmed in Port-au-Prince, which is about 60 miles from where the first cases occurred and about a three hour drive due to the poor conditions of the roads. A cholera outbreak in this central hub is certain to lead to cases being spread to other refugee areas.

I truly fear that this outbreak could lead to thousands, if not tens of thousands of deaths. Worse yet, it may be too late to prevent much of this from unfolding without some incredible efforts at this point, but I don’t have much faith that it will be forthcoming. Haiti may have already crossed the point of too little, too late.

The Risks of Confounding

A study was just published titled “Short term effects of temperature on risk of myocardial infarction in England and Wales: time series regression analysis of the Myocardial Ischaemia National Audit Project (MINAP) registry” in the British Medical Journal. It’s kind of a mouthful but the authors are trying to draw a link between the risk of a heart attack with decreasing temperature.

There are probably a couple of factors in this that make this study somewhat flawed. The use of ambient temperature as an indicator for a heart attack in this case could be technically called a confounding variable. Here is a classic example to explain exactly what that is.

Assume a researcher wanted to study the causes of lung cancer but had no idea what factors could contribute to the development of disease. A case-control study would be the typical design used to look for the variables that contribute to disease. In essence, it takes a group of people and assigns them to two different groups – those who have disease and those who do not. The researcher would then look at variables that might explain the cause of the disease in the group that is ill.

Let’s say that our hypothetical researcher decides to study the presence of a lighter in a pocket or purse thinking that maybe some of the chemicals in the butane might cause lung cancer. Statistically, this variable is going to show up as a significant factor contributing to lung cancer. This is where we run into the problem of causality versus association. Just because a variable is associated with a particular outcome does not mean that it is the cause of the outcome. Today, we clearly know that carrying a lighter does not cause lung cancer. However, those who carry lighters are likely to smoke, which is a clear cause. Hence a lighter in this example is a confounding variable.

Back to the proposed link between temperature and heart attacks. There is a good chance that temperature is a confounding variable in this case. Colder temperatures also occur at the same times as snow and influenza. The risks of a heart attack increase after snowfall because of the extra exertion that is needed to shovel snow, especially when it is wet and heavy. Influenza causes inflammation, including in the coronary arteries, which increases the chance of a heart attack. It doesn’t look like the authors addressed these alternative hypotheses in their study.

So what are the two take-home messages? I’d argue that there is always more than meets the eye and you can’t believe everything you read, even if it comes from a reputable source.