[這是一個屈辱的回憶：兩千年的六月，我們系上兩個老師聯手把我趕出他們的研究，理由很可笑。他們還說我沒有能力唸博士，叫我拿個碩士就可以滾了。他們也蓄意封殺我，期末報告規定交七頁雙行格式，他們卻用四頁單行的篇幅（等於是八頁）來批評我的報告，批得一無是處，然後給了我這輩子最低分的成績：2.9分（滿分四分，2.9是研究生剛好及格的成績）。面對這樣的羞辱，我仔仔細細地一點一點反駁，以十頁的篇幅來反擊。不過最後我還是失敗了，他們不肯認錯改成績。我是有認真考慮過要去學校告她們，可是想到要繼續在系上生存下去，就放棄了這個念頭，要不然以後誰敢收我呀？！以下就是我反擊他們的內容，我孤獨沈痛的吶喊…]
Dear Susan:
First of all, thanks for your four-page single-spaced comments to my term paper for the course of Epi 533 (Pharmaco Epi). In response to your comments and clarifying some misunderstandings, I decided to write this e-mail to you.
The most concerned point in your comments of the report was about the citation of proper references. I made a mistake in citing UKPDS 38. What I meant was UKPDS 34 (Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes, UKPDS 34. UK Prospective Diabetes Study (UKPDS) Group. Lancet. 1998; 12; 352: 854-65.), the one Eric gave me and we discussed in the independent study. So, the text was right. I just chose a wrong one from Medline. In UKPDS 34, it was right that the randomized subjects were 753 (because the trial enrolled overweight DM patients only). The reason why I cited it was to explain why the risk and public health importance of hypoglycemia was rising (because the benefits of intensive therapy to DM patients were proved by UKPDS 34). Actually, on June 2nd, in the meeting with you, Emily and Eric, was the day I FIRST learned about UKPDS 39, the clinical trial about beta-blockers. So, it is NOT true that I cited UKPDS 39 but failed to explain its results properly. In the independent study with Eric, we studied and discussed 42 papers about DM, its treatment, hypoglycemia, and HCHP population, etc. I never knew that there was anything related to UKPDS and beta-blockers together. If you try to find anything about beta-blockers in Medline (http://
www.ncbi.nlm.nih.gov/entrez/query.fcgi), you will find 58,697 articles. If you were me, the one who wants to search articles about hypoglycemia, 16,402 single ones are right there waiting for you. If you type "UKPDS" as the keyword, you could get 75 papers. Or "hypoglycemia AND beta-blockers", you will get 418 papers. Even if you are lucky enough to know UKPDS 39 might be important, 6 papers could be found in Medline. I guess that's why I need a DIRECTOR in the readings of independent study. I am pretty sure that Eric didn't tell me that UKPDS 39 was about beta-blockers and hypoglycemia. On May 22nd (two days before my oral presentation), I showed the outlines of my presentation to Eric and asked him about how to pick up a proper example of clinical trials and I showed him the three randomized clinical trials we discussed (Kerr D et al. Br J Clin Pharmacol. 1990;29:685-93., Viberti GC et al. Metabolism. 1980;29(9):866-72., and Clausen-Sjobom N et al. Acta Med Scand. 1987;222:57-63.). He told me that the one with most subjects (Kerr D et al.'s) was proper. So, that's the whole picture how I didn't know anything about UKPDS 39. Maybe I should try to look at all the related articles in our discussion. But, there would be hundreds of them.
And, when we met each other on April 24th in F346-D to discuss my oral presentation, you told me that basic information about DM and beta-blockers could be decreased, but the part about public health importance could be increased in my presentation. You also told me that DCCT and UKPDS were good studies about the treatment of DM and its prognoses. I suspected that you didn't know anything about the clinical trial in UKPDS 39 then, either. If you did, why didn’t you tell me? I am your student!!
After the meeting on June 2nd, I was upset but still went to the library to study that paper (UKPDS 39). The main purpose of the trial was to determine the effect of tight control of blood pressure by a beta-blocker (atenolol) or an ACE inhibitor (captopril) in preventing the macrovascular and microvascular complications of type II diabetes. The side effect of hypoglycemia was not the major concern in the study. I found it was not a good clinical trial at all in addressing hypoglycemia as an outcome, although much more subjects were enrolled. The results about hypoglycemia contained possible biases. And, they didn't calculate relative risk of hypoglycemia, either, but concluded that beta-blockers and ACE inhibitors had similar effects in causing hypoglycemia (the same fault made by EUCLID study). My comments for UKPDS 39 are:
1. Randomization could really control ALL possible confounders? I don't think so. It could still happen by chance. In intention-to-treat comparisons (Table 1), 38% patients in captopril group were treated by sulphonylurea, but only 29% was treated by sulphonylurea in atenolol group. And, 20% were treated by insulin in captopril group, but 29% were treated by insulin in atenolol group. So the treatment of DM was related to the exposure by the process of randomization (by chance, in other words). And, it's surely related to the outcome, hypoglycemia, too. It could be a potential confounder. Unfortunately, they didn't try to control it in presenting the results of hypoglycemia.
2. The difference of compliance between captopril group and atenolol group was huge (P value less than 0.0001). 78% patients in captopril group took it till the end of trial. But, only 65% patients in atenolol group did, mostly because of intolerance by impaired peripheral circulation or bronchospasm. I believe it's non-selective (to the outcome of hypoglycemia). So, in the analysis of intention-to-treatment, this is a non-selective misclassification of exposure which results in relative risk shifting to the null. Since only 65% compliance was performed in atenolol group, the effect estimation of beta-blocker was very dubious.
3. I tried to calculate the relative risk of major episode for ACE inhibitor and beta-blockers relative to less tight control group. They were 1.49 (95% CI: 0.79, 2.93) and 1.29 (95% CI: 0.64, 2.61) separately. The estimations were suffered by non-selective misclassification of exposure (making the estimations shift to null, especially for the beta-blocker group). Although they were not significant, the differences were still unforgettable.
Then, I tried to calculate the detectable RR by PS program (Power and Samples Size Calculation, a free software made by William D. Dupont and W. Dale Plummer, Jr., Division of Biostatistics, Department of Preventive Medicine, Vanderbilt University School of Medicine, 1996). Under the condition of 80% power, at significance level of 0.05, sample size 400 in ACE inhibitor group, and 4.4% probability of events in controls (less tight control group), the minimum detectable RR was 2.21. In beta-blocker group (n = 358), it's 2.30. Insufficient sample size is still a crucial issue in this study.
So, you said that no one could perform a better study in addressing the effect of beta-blockers to hypoglycemia among DM patients might not be true. The study of UKPDS 39 was apparently not a sufficient one to reject the possibility of adverse effect from beta-blockers. But, it's true that, without plenty of time, money and subjects, it might not be feasible as a doctoral dissertation.
As to your comment 3), you said "In section F, you did not explain why you think the "important possible confounders" listed (race, antidiabetic drug therapy, etc) would confound the association between beta-blocker use and hypoglycemia. Would they be confounders or effect modifiers?" I think I had explained how race could be a confounder in such studies in the meeting on June 2nd. And, the study of UKPDS 39 could be a good example of how antidiabetic drug therapy could be a confounder (either by chance for clinical trials or confounding by indication in Epi studies). These possible confounders were suggested by the authors of related articles. They did not cite anything in the discussion to explain how they could be confounders, either. If you were the one of the reviewers for these articles, no one could be published then. Effect modification is another phenomenon which might be rare. Without further biological information/sense, it might be hard to decide if these factors could be possible effect modifiers.
About your comment 4), you said "In section G, item 2, you suggested alternate methods of defining exposure to beta-blockers. You did not justify why "30-day cumulation" or "90-day cumulation" would be appropriate or desirable ways to define beta-blocker exposure in a study of hypoglycemia. You were also not able to justify this proposal when I asked about it after your presentation." Actually, in my term paper, I wrote, "Differential exposure windows could help us to evaluate the effect of beta-blockers in a more standard way (FOR EXAMPLE, differentiating the exposure as "on the date of episode", "30-day cumulation", and "90-day cumulation"). And, I explained it in response to your question after my presentation. I said, "30-day is the general duration of a cycle of chronic prescription. So, 90-day cumulation could be the cumulative exposure of three cycles." In two case-control studies (reference 23. Thamer M, Ray NF, and Taylor T. Association between antihypertensive drug use and hypoglycemia: a case-control study of diabetic users of insulin or sulfonylureas. Clin Ther. 1999;21:1387-400. and reference 21. Morris AD, Boyle DI, McMahon AD, Pearce H, Evans JM, Newton RW, Jung RT, and MacDonald TM. ACE inhibitor use is associated with hospitalization for severe hypoglycemia in patients with diabetes. Diabetes Care. 1997;20:1363-67.), they defined the exposure by 90 days and 45-day exposure but did not justify why they did that at all. They defined it arbitrarily without any reason provided. If you were the reviewer of the papers, they could never be published, right? But these articles were published in the best journals!
About your comment 7), you said, "In section G, the suggestions for future studies were not specific. For example, in what setting could you study this question? How would you ascertain hypoglycemic episodes? Would the number of hypoglycemic events be adequate? What would be comparison exposure group?" However, in the section of handout for Term Papers and Presentations, Term Paper 4) was: “What are the new important lines of research?" and Grading 4) was: "Use of student's knowledge in pharmacoepi to interpret findings and suggest new lines of research/next steps." Thus, what you meant should be some suggestions to future studies. So I utilized the knowledge from your lecture about pharmaco exposure measurement and Eric's lecture about precision and validity and answered them as:
In measuring detailed exposure of beta-blockers, I suggested:
1. Sub-group analysis: in available studies, they merely try to discriminate the effect of selective/nonselective beta-blockers. Actually, in similar way, we could estimate the effects of long-term/short-term, lipid soluble/non-soluble beta-blockers.
2. Differential exposure windows: "on prescription" was generally the way they measured exposure. However, they did not define "on prescription" in the same criteria. Differential exposure windows could help us to evaluate the effect of beta-blockers in a more standard way (for example, differentiating the exposure as "on the date of episode", "30-day cumulation", and "90-day cumulation").
3. Effects of combined therapy: comparing monotherapy vs. combined regimens with diuretics, ACE inhibitors, calcium antagonists, or vasodilators.
In improving precision, we could:
1. Enroll more subjects in clinical trials,
2. Collect more cases in case-control studies, and
3. Generate more person-years with more events in cohort studies.
In improving validity, we should:
1. Control for possible confounders, including race, duration of diabetes, antidiabetic drug therapy, switching antidiabetic medication, multiple medication, experience of hypoglycemia, hepatic disease, renal disease, degree of glycemic control (regular blood glucose test and HbA1c), and
2. Utilize population-based study design.
So, I think I've already answered your question 4) in the handout properly. Thus, I suggest you to describe the questions more specific to avoid the same mistakes happening to students in the future. At last of your comments, you said, "The problems discussed above, related to omissions and misinterpretations of the literature in your topic, and the problem with understanding of key epidemiologic concepts, give me great concern about your ability to proceed with a doctoral project." I just started to try to learn something about Pharmacoepi for merely three months, and you've already decided that I am totally not eligible to be an Epidemiologist. I am a student who is still learning. How could you expect that I could do everything?
However, I really think you misunderstood my capability. I was one of the few students who could pass the National Studying Abroad Exam twice in the field of Epidemiology. However, at last, physicians always won the prize. (Most of Epidemiologists are also physicians in Taiwan... maybe 60-70%, I think.) Only three young Epi academics could pass the exam but one of them could gain the scholarship each year or more than a year. So, I didn't get the scholarship from my government before I came here. I also have two SCI published papers (as shown in my resume, one of them were published in the top one journal of Urology) and three on-going cooperative studies (almost done). I really don't think that I don't have the "ability to proceed with a doctoral project"!
Anyway, after all the misunderstandings, I still appreciate your and Eric's instruction. I learned a lot from you. Hope you don't hate me. I will keep on going with my dream to be an Epidemiologist.
Chin-Kuo
