Do hospital investigations alter patients’ choices?

In the UK health care system, the Care Quality Commission can conduct high-profile investigations of hospitals. In a recent article published in Health Affairs, we investigated the impact of such investigations on patients’ behavior. For three investigations, we analyzed the trends of non-emergency patient admissions and other utilization indicators. Controlling for secular trends and using a control group, we employed a difference-in-difference approach. We found that the investigations only had an impact for one hospital with significant declines in inpatient admissions, outpatient surgeries, and in numbers of patients coming for their first appointment, but the effects disappeared six months after publication of the investigation report. Apparently, the publication and dissemination of the highly critical report does not have a sustainable impact on patients’ behavior like avoidance of the hospital. Assuming that hospitals implement improvements also without the threat of losing patients, public reporting has no adverse effects on the number of patients for the hospital.

A new contribution to embryonic stem cell research

Last year, we published an article in Nature (read my summary here) linking the so-called hmC DNA modification to an enzyme and its mutation in a subgroup of leukemia patients. At this stage, it was still not clear whether the DNA modification hmC indeed has a functional role at the molecular level or whether it is just an intermediate product in a reaction chain.

In our new article in Nature (here), we shed some new light on this hypothesis. Therefore, we developed two new biotechnological methods, which allow us to detect the positions of hmCs very accurately in the DNA with a resolution of a a few basepairs up to a few hundred basepairs. Based on a sophisticated statistical analysis, we are able to analyze the patterns of the hmC regions (around 100,000). The less randomly the regions are distributed, the more likely it is that hmC has a functional role. Indeed, we observe a strongly non-random pattern of hmC regions. More specifically, we discovered that hmCs are significantly more often co-located in the regulatory regions of genes with other known histone modifications. Therefore, we hypothesize that hmC plays a fundamental role in the regulation of developmental genes.

Gould’s "The Mismeasure of Man"

I just finished reading the exceptional book “The Mismeasure of Man” from Stephen J. Gould. He critically reviews the history of the intelligence quotient (IQ) from its predecessors like brain size to ‘modern’ definitions as the Spearman’s general intelligence (g).

Binet introduced the first test for intelligence in France. However, he repeatedly stated that the test can only be used to identify weak students in school to offer additional help to them. As the idea to measure intelligence migrated to the US, it was coupled with some further claims. Despite any scientific evidence, intelligence was considered to be mainly inherited and hardly influenced by the environment. Any correlation between intelligence and poor environment was explained by intelligence: ‘Stupid’ people are poor because they are ‘stupid’. Furthermore, intelligence was seen immutable meaning that it doesn’t change after adolescence. This was used as an argument against higher education for ‘stupid’ people. As Spearman introduced a measure for general intelligence based on factor analysis by using the first principal component, one could (and did) use this measure to rank people based on their intelligence.

Based on the ranking of people, classes and races, the US introduced laws to prevent immigration of lower classes and races, sterilized women with low IQs (one of the laws in Virginia was enforced until from 1924 to 1972) and cut social fare programs as well as education for ‘lower’ people. The book explains mainly the scientific errors for such a paradigm:

1. There is no single entity, which can be called intelligence. The derived intelligence by Spearman is just an average of all different mental tests. Certainly, these mental tests measured some ability of solving (specific) tasks. However, this doesn’t imply that the average of all mental tests can be seen as ‘general intelligence’. Furthermore, by changing the ensemble of mental tests, the definition of general intelligence changes as well. 
2. We cannot rank intelligence. Most psychologists would agree that there are different types of intelligence in terms of different abilities to solve cognitive problems. Although we always can summarize such multidimensional values to one value, nothing tells us that the derived value is meaningful and, more importantly, that this value can be ranked. 
3. There is no empirical evidence suggesting that environment has only minor (or no) influence on intelligence. As mentioned above, the correlation with environment was usually explained as an effect and not as a cause.
4. Inheritance of intelligence. One might attribute differences in intelligence to inheritance within a certain group (for example based on twin studies). Although this genetic component explains these differences within the group (e.g. white Americans), it is wrong to conclude that inheritance explains differences between different groups (white and black people). Why? An example of body size helps: Let’s say that we have figured out the body size is inherited in a group of white Americans. They have an average body size of 1.70m. Measuring a group of Indians in a poor village, I will certainly obtain a smaller average body size. Although inheritance can explain a certain fraction of variation in the white group, it doesn’t necessarily explain the difference between the two groups. Rather nutrition is the main contributor to the difference.
5. Variation occurs around the average value. For most genetic influences, the variation is much higher than the average difference between two groups. Therefore, group membership is a poor predictor for the individual value and there will be many members of the ‘lower’ group with higher values than many members of the ‘higher’ group. (In this case, using group membership as criteria (for example to be admitted to university) would be clear discrimination despite the differences in average.)

These issues are discussed in much detail in the book. Interestingly, an eugenic program requires all these assertions (intelligence as a single, innate, immutable and inherited value without high variation and environmental influences) to be true although all of them are empirically wrong. Now, one might still wonder whether one should implement an eugenic program if all these assertions would be true. Gould doesn’t discuss this issue but the answer is clearly no – and I will discuss in a forthcoming post (here).

In summary, the book is in-between a little repetitive and too talkative, but it’s definitely worth reading because of the following topics discussed in-depth:

• There is nothing like a general intelligence, which one can use to rank individuals.
• The re-use of these arguments every now and then (e.g. in The Bell Curve from 1994).
• The tragic impact of wrong science on society including the destruction of life.
• The subjectivity and blindness of scientists.

Especially due to the last point, I would almost consider this book as a must-read for any scientist. Gould shows how the different scientist drew wrong conclusions. Often, the arguments were patently circular. Although, these scientists often accused opponents of the very same logical fallacies, they were not able to see the circularity of their own argument. Moreover, they argued that they would have performed a purely scientific investigation and they claimed quantification as a criteria for objectivity. One cannot say that these scientists would have produced fraudulent results because they often published all the raw data such that the mistakes are obvious which a conscious cheater would never do. Hence, the scientists were subject to a personal bias based on a priori beliefs. I was surprised by the extent of this bias, which Gould shows with many quotations taken from the scientists’ publications. This shows the huge influence of cultural context on ‘objective’ studies, which every scientist should be aware of.

From nose picking to rejection and from cleanness to moral

I’ve been told about a lecture from the psychologist Daniel Gilbert about wrong decision making in everyday situations. One of the plentiful examples was about the first date of two persons. Picking one’s nose leads to an instantaneous rejection by a huge majority of people. Gilbert considers this unreasonable since picking one’s nose doesn’t tell anything about a person. Hence, using this to reject someone as possible partner doesn’t make sense.

I wouldn’t write about if I wouldn’t disagree with it. I think many decision making processes are based on limited information and constrained by time. Therefore, humans unconsciously use heuristics to derive decisions, which are not always the best but are likely to be advantageous and minimize certain risks. Without such heuristics nobody would survive in our complex environment. If we would approach any person as a blank slate and just use the explicit information provided, it would not only take ages to ‘know’ a person but many times we would be disappointed since we can only discover the ‘true’ character if corresponding situations occur. (I don’t even believe that this would help since already language is interpreted and thus subject to prejudices.)

Hence, here is my interpretation of the above date situation: The habit of picking one’s nose is correlated with other habits of a certain character. If one observes someone picking his nose, one automatically infers that there is a high probability that the person also comprises some of these other habits, which one dislikes. The probability that the person is a good match as a partner decreases. From this perspective, the decision making is reasonable and makes sense. Of course, sometimes this inference is wrong and one might reject someone, who would turn out to be a fantastic match. One might consider this to be unfair but there is obviously a trade-off between rejecting true matches vs. wasting time in ‘assessing’ false matches. Furthermore, I’m not claiming that this decision is optimal but it is reasonable.

In the book ‘The blank slate’ from Steven Pinker, the author draws wrong conclusions in the same way. In chapter 16 about politics, he discusses moral. One of his points is that people tend to assign moral values to non-moral subjects and the other way around. I certainly agree with this, however, some of his examples and the way he derives this conclusion could be more convincing. He cites a study based on showing different pictures to test persons, which showed that test persons assess other people, who observably value cleanness, as morally superior to people, who don’t care about cleanness. Strange? Not at all!

First of all, one tends to feel closer to people with shared values (cleanness, for example), which makes sense. Since most people do feel being moral, the step from ‘feeling close’ to assigning a certain moral value to these persons is not too big. Does this make sense? I think based on heuristics, this conclusion is reasonable (although maybe not optimal). Cleanness is probably related to a preference of long-term vs. short-term thinking (other people call this the root of civilization, by the way). Moral (as well as cooperation) only is advantageous in the long-term as, for example, game theory clearly indicates. Already based on this, it is not very far fetched to favor people who like cleanness since those people are likely to act based on a long-term perspective. This obviously does not imply that all ‘clean’ people are moral while all ‘non-clean’ people are not moral. It is just a tendency and not valid to blame individuals based on this, which would be one of the most common fallacies. Especially in the context of an experimental setup showing pictures to the test persons, I think the assignment of moral superiority to ‘clean’ persons is not surprising. Besides this point of long-term vs. short-term perspective, there might be even more ‘hidden’ attributes leading to favor cleanness in common with having a moral view than to like a dirty chaos and moral – for example to care for other people, who just might not feel so comfortable in the dirt of others.

So, I’m not declaring cleanness to be moral. I’m just stating a way to understand why one might make a heuristic connection between cleanness and moral. And I would further claim that the experiment rather tested a heuristic connection by implicitly measuring a subjective correlation than moral reasoning. Test persons would certainly refuse to assign a moral value to cleanness. So would I.

Naomi Klein: The Shock Doctrine

Accidentally, I got a glimpse at the book “The Shock Doctrine” from Naomi Klein. It looks like a big bashing against Milton Friedman – at least on the first pages. She argues that Friedman’s doctrine is to use a catastrophe like a natural disaster or a war to implement his ideas of turbo-capitalism. After a catastrophe, institutions are destroyed and one can start from a blank slate. She implies that this results in the implementation of ideas, which wouldn’t have a democratic majority for them.

I think these are two different problems. It is certainly true that most change happens after catastrophes. However, this is neither anything specific about Friedman’s doctrine nor anything new. The second problem is whether this way non-democratic decisions are made. I would still be optimistic to claim that voters can punish the ruling party in the next election; this is how democracy works. Especially if these decisions are important as Naomi Klein states. If this doesn’t happen than either society is ok with such decisions or seduced by media; anyhow this would be a different and very fundamental problem, which is also not specific to Friedman’s doctrine.

However, people might be repressed such that they cannot express their opinion. Here, Naomi Klein uses the sad history of South America. There is certainly a point for this but it doesn’t explain the capitalistic movements in other countries. Very puzzling is that Naomi Klein says that the Tiananmen massacre must be seen as another example: The turbo-capitalists in China used the massacre and the following threatening of the people to implement capitalism… This is pretty far-fetched. The free-market promoters were among the protestants and not within the government…

And a final paragraph on Naomi Klein’s elaborate use of statistics: She writes to claim that privatization is also implemented in the military sector:

To cite just three statistics that show the scope of the transformation, in 2003, the U.S. government handed out 3,512 contracts to companies to perform security functions; in the twenty-two-moth period ending in August 2006, the Department of Homeland Security had issued more than 115,000 such contracts. The global ‘homeland security industry’ – economically insignificant before 2001 – is now a $200 billion sector. In 2006, U.S. government spending on homeland security averaged $545 per household.

Just a few remarks: (i) Number of contracts is a poor comparison since the value of the contracts matters. (ii) If there is a war, it is obvious that security costs and number of contracts increase. Hence, a comparison should be made between non-private and private costs – otherwise she won’t arrive at her conclusion independent of the numbers. (iii) The $200 billion is expressive but doesn’t lead to the conclusion of privatization. (iv) The average household spendings are very unclear. It’s hard to compare since the other numbers are not relative to households. And, again, it doesn’t lead to the conclusion.

I don’t want to assess whether there was a massive privatization in the security sector. I’m just worried that the argument doesn’t help to solve this question. It’s sensation-seeking with big numbers. Intellectually totally unsatisfying and on the edge to be insincere. Although I’m aware that I just had a look at the very few pages, the book doesn’t look to be very reasonable – literally. It might be good for a emotional or pseudo-moral discussion about capitalism – but nothing more. A real final remark: Naomi Klein uses privatization of health care as one of these bad projects against public opinion… The current situation in the U.S. doesn’t look like this.