I felt that my experiences with testicular cancer involved many of the decision biases discussed in class and made an excellent topic for the paper. While the frequent discussion of decision biases makes the story a little dry, I think it might help people to understand and compensate for these biases.
In August of 1992 I noticed something strange about my left testicle. It felt as if there was a ball bearing inside the testicle. I could not remember if it had always been that way or if this was something new. It did not hurt so I decided to wait and see if it would go away on its own. After a few weeks I looked up testicular cancer in an old medical guidebook we had in the house. It did not mention any symptoms, but it did mention that testicular cancer was very rare. I interpreted this to mean that it couldn't happen to me. My wife, who often thinks I'm a hypochondriac, agreed.
Gilovich writes that, "Truly ambiguous information is often simply perceived in a way that fits our preconceptions."[1] The fact that testicular cancer is `very rare' is ambiguous and not altogether correct. "Although testicular cancer is a relatively rare tumor, it is the most common cancer among men aged 15 to 35 years. In the United States in 1989 approximately 5,500 men will be diagnosed with testicular cancer."[2] I had a preconceived notion that I could not possibly have cancer. Consequently, I equated `rare' with zero probability. I did not go to see a doctor.
After learning of my mistake, I addressed the issue by removing the source of ambiguous and outdated information from the house. I got rid of the old guidebook and replaced it with something newer and more comprehensive, which is not to say that the problem of ambiguous information will ever go away. The new book has a correct and informative section on testicular cancer, however, I know that it may turn out to be just as outdated and incomplete as the old book the next time we need it. Since we will all have to deal with ambiguous information sooner or later, I hope not to let my preconceived notions unduly bias that information.
During this period my wife and I were trying, without any luck, to have a baby. For this reason my wife wanted me to see a urologist. I agreed, however, I took my time. For a variety of reasons I did not see the doctor for almost two months. While I was still a little concerned about my condition, I wasn't in any pain so I didn't think that it was anything serious. I planned on asking the doctor about the testicle, but in reality I was more interested in getting a sperm count.
I finally saw the doctor on Tuesday, October 20th. After examining the testicle for 10 seconds he told me that there was a 95% chance that I had cancer. Two days after that an ultrasound confirmed the diagnosis. A sperm count also revealed that I was essentially sterile. Five days after that I had surgery to remove the cancerous testicle. I also had a variety of bloods tests, x-rays, and CT scans to determine the extent of the disease. Two days after the surgery the pathology report came back. It turns out that I had what is referred to as Clinical Stage I Nonseminomatous Germ Cell Cancer of the Testis. The cancer was clinically Stage I because they didn't think it had spread outside the testicle, but they couldn't be certain without opening me up and looking. The urologist advised me that because of the type of cancer cells found in the tumor (embryonal), he recommended that I undergo surgery to remove some of the lymph nodes in my abdomen (in an operation called a Retroperitoneal Lymph Node Dissection or RPLND) and undergo two years of chemotherapy. It would be an understatement to say that I was upset.
Up to this point I had not done any research of my own nor had I sought a second opinion. I was in an HMO so I didn't think I had much choice about who did the surgery. Besides, removing the testicle was a simple operation, and I did not want to put it off. However, the urologist's recommendation shocked me, and I quickly sought some other opinions. I began to do some research at the medical school library, and I talked to a number of oncologists and urologists from around the country.
I learned that my fate was not nearly so bleak. However, I did need to choose between two distinctly different methods of treatment entailing different levels and types of risk. Not risk of death, but risk of misery, sterility, uncertainty, and chemotherapy. It was a complex decision that kept me busy for almost two months. Here is how I described my situation on November 28, 1992 in a post to the Internet:
I'm kind of in an awkward position, and I'm interested in how other people may have dealt with this kind of problem. Basically, I am very lucky. Testicular cancer is one of the one's they have managed to "cure." However, since the cure rates are so good for people at my stage (Stage I, an encapsulated tumor with no apparent vascular invasion) I have to choose my own treatment. Given my pathology, I probably have a 15-25% chance of relapse. I can either enter a surveillance program involving monthly chest X-rays and bloods tests combined with quarterly CT scans for a year and a reduced schedule thereafter for at least 5-10 years, OR I could have a retroperitoneal lymph node dissection that would lower the risk of relapse to below 5%. I would still need to follow up with some surveillance, but I wouldn't need CT scans. In either case, if I relapse, I would be treated with 3 or 4 cycles of chemotherapy which would cause a complete remission greater than 90% of the time.
To add further fat to the fire, I am currently with an HMO. If I had the surgery tomorrow, they would pay for everything but I would have to use a local doctor that has never done this operation. If he does it wrong, I end up sterile. If I can wait until January, then I could have the operation anywhere I want because I switched insurance plans.
There's a lot more, but I think you have the gist of it. I've got an 80% chance that I am already cured. Should I go the surveillance route because I should avoid unnecessary surgery and chemo will cure me, or should I risk the surgery because chemo is to be avoided at all costs?
For the sake of clarity, my immediate options were surveillance or surgery. The surveillance approach entailed frequent checkups, blood tests, and CT scans. If the cancer relapsed, I would probably need three months of chemotherapy, and I might also require RPLND surgery. The other approach entailed RPLND surgery followed by surveillance to check for relapse. The chance of relapse was reduced, though, because the lymph nodes have already been removed. If the cancer relapsed, I would still require three months of chemotherapy. The risk of death is roughly the same, though it may be slightly higher for the surveillance approach. In either case, the 5 year cancer-free survival rate for both treatment option is over 95%.
The surgery involved is a major operation that requires a week in the hospital and six weeks of bedrest. If the surgery is not done carefully, it can cause sterility. The chemotherapy is conducted on an inpatient basis and is particularly brutal. Its major side effects can include lung problems, kidney failure, sterility, neuropathy, and hearing loss. However, it is extremely effective for this type of cancer.
In my search for a decision, I found myself dealing with many biases. Some of the biases were mine, some were in the literature, and some were in the doctors and other patients. I started out my search for a decision with a clear preference for surveillance. It avoided any unnecessary treatment, and it maximized the probability that I would regain my fertility.
Since this goal completely disagreed with my urologist's recommendation, I felt that I needed to justify my preference. In doing so, I succumbed to the tendency to seek confirmatory information. Gilovich writes that, "When trying to assess whether a belief is valid, people tend to seek information that would potentially confirm the belief, over information that might disconfirm it."[3] In my initial research, I was particularly interested in finding articles that would support my decision. I imagine that I unconsciously selected the research supporting my beliefs out of the abstracts. I know that I read the articles supporting my beliefs long before I read the articles disputing my beliefs.
I recognized this bias quickly. I realized that I could be making a serious mistake by discounting people and articles that did not support my beliefs. Consequently, I made a concerted effort to read and understand the contradictory information. Gilovich predicted I would do this: "Rather than simply ignoring contradictory information, we often examine it particularly closely. The end product of this intense scrutiny is that the contradictory information is either considered too flawed to be relevant, or is redefined into a less damaging category."[4] However, Gilovich did not predict my reaction. I found myself easily swayed by whatever I was reading at the time. I suppose this means that I had not really come to a decision. Alternatively, it could be evidence of the `ease of recall bias'[5]. There was so much information swimming through my brain that I was overly influenced by whatever I had read last. Once I recognized this tendency, I rearranged my reading so that the last article I read would support my initial belief. Whether this was the right thing to do, I don't know. It did, however, put me in a better mood. In any event, I feel that this bias still affects me to a great extent. I am much more aware of the bias now, but it requires a concerted effort to overcome it.
I eventually accumulated over 80 articles, 6 second opinions on the treatment, and 1 second opinion on the pathology report. I gathered this much information to guard against the tendency to believe what we want to believe and to guard against the confirmation trap. All 6 second opinions agreed that I wasn't taking on unnecessary risk by go through with surveillance. Furthermore, the second pathology report did not show any embryonal cancer and therefore supported my decision. As for the articles, it's not hard to be thorough when your life depends on it.
My decision may also have been biased by the certainty effect. Bazerman writes that, "Under the certainty effect, we are more apt to be interested in reducing the likelihood of certain events than uncertain events."[6] In my case, there were so many branches in the probability tree that it was hard to determine the amount of risk I was taking. The only thing that was certain was that if I chose to have surgery, then I would have a scar from my pubic bone to my sternum for the rest of my life, and I would be out of commission for almost two months. Accentuating this problem was the ease of recall bias: I had just had minor surgery to remove the testicle. I was only in the hospital one night, and yet everything hurt much more than I expected, and that weighed heavily on my mind. I tried to offset the certainty effect by talking to a couple of people who had undergone chemotherapy, but I failed since the probability of needing chemotherapy was hard to compare with the certainty of having major surgery. This effect is definitely not universal. I exchanged email with another guy who had the surgery. He felt that the certainty of surgery was far superior to the increased risk of chemo that goes with the surveillance approach.
I also ran into evidence of the false consensus effect. As Gilovich writes, "It is a fact of social life that we are selectively exposed to information that tends to support our beliefs. Conservatives read conservative periodicals and thus receive support for a conservative political agenda.... Furthermore, in addition to a biased set of arguments relevant to a particular belief, we are also exposed to a biased sample of people and their opinions."[7] As I have mentioned, in my search for a decision I heard a strong opinion favoring surgery from my urologist. I later figured out that urologists are surgeons and, since they know more about surgery and put more trust into surgery, they are far more likely to recommend surgery than anything else. On the same note, medical oncologists use drugs to cure cancer and are therefore more likely to have confidence that the chemotherapy will work. I don't think these doctors would recommend a specific treatment because it would be to their financial advantage. I think that they all have a natural confidence and trust in their own specialty. This kind of bias is probably pervasive in all fields, and it lends credence to the concept of the second opinion even outside the medical profession.
I also noticed that there was a geographic bias to treatment options. Despite a common research language and a sharing of information, the treatment for this type of cancer has typically been very different in Europe than in America. A patient is far more likely to have an RPLND in America than in Europe. I was told that this is because there aren't very many good surgeons in Europe. Whatever the reason, it is an important factor to note.
Finally, I am proud to say that I suffer from a case of the hindsight bias. Nine weeks after the initial surgery, Julia was conceived. Every time I see her I say to myself, "I knew that was the right decision." Every time I go back to see the results of my latest CT scan or blood tests I can say the same thing.
Epilogue: It has been more than twenty five years. I now have two beautiful daughters and a very cute little boy who isn't so little anymore. I am statistically out of the woods. I still go back to get a chest xray and a blood test every so often. Just to be sure. I've talked to too many people in the unlucky 1% to assume that I am completely cured.
[1] Gilovich, T. (1991) How We Know What Isn't So: The Fallibility of Human Reason in Everyday Life, New York: The Free Press., p.54.
[2] Ozuls, R. F., and Williams, S. D. (1989) Testicular Cancer, Current Problems in Cancer, 13 (5), p.289.
[3] Gilovich, T. op. cit., p.33.
[4] ibid., p.55.
[5] Bazerman, M. (1994) Judgment in Managerial Decision Making, New York: John Wiley & Sons., p.17-19.
[6] ibid., p.64.
[7] Gilovich, T. op. cit., p.115.
