Sorry

for my lack of updates on this recently guys. I promise to get right back onto it. The good thing is that I PASSED (whhooo!!!) my defense last week, so I can get back to updating this sight daily in the way that I want it.

You will start to see a little bit of a change with this blog. Not only will I post professional reviews, but also give Gastroenterologists the inside scoop on interesting things in the field.   Stay tuned, tomorrow will be lots of stuff. Can't wait to get back on this.   

Important to Notify Patients About Adverse Effects of Colonoscopy

A great study from the Netherlands detailing the negative effects of a Colonoscopy.

Summary: The main objective was to study was the adverse effects of a Colononscopy after the procedure. The authors found that about 1 percent of the accrued patients (n=1114) had experienced a major adverse event (hospitalization) and about 30 percent (!) of the patients had experience a minor adverse event (no hospitalization) 30 days after their colononscopy.  The study showed that those patients who experienced an adverse event were found to have a more negative view of Colononscopies and consequently less likely to return for another colonoscopy in the future.

Implications for Practice: The finding that 30 percent of patients were found to have a minor adverse effect from a Colonoscopy should raise an alarm for any Gastroenterologist.  This underscores the importance of notifying patients about the potential adverse effects of a Colonoscopy and following up with them after the procedure.  Still, Colononscopies are deemed rather safe and highly recommended.      

It is interesting that females and those under the age of 50 were shown to have significantly higher chance of experiencing an adverse event compared to males and those over the age of 50 respectfully.  It is difficult to elucidate the exact reason.  Another study did show that females were more likely to experience colonic perforation compared to men.  Follow up studies on this would be compelling.

Affordable Care Act is Finally "Safe"

Well, it's been 3 days since the decision, and it's honestly all I can think about. I can't tell you how many victory laps I have taken since the decision was announced. Like many of you, I was watching CNN and had a true "Dewey vs. Truman" moment.  CNN first in-correctly said that the law was turned down and then later correctly said that the law had been up-held.  Wow, what a swing of emotions in such a short period of time.  You are talking about the difference in countless lives and trillions of dollars between the two outcomes.  No, it's more than a "Dewey vs Truman" moment.  

But let's get to the law and it's implications - the important thing. There is simply no way to describe how beneficial this really is to the Gastroenterology community.  This article from USA Today sums up my feelings very well.  As I said in a recent post, there are many Gastroenterology patients who probably did not  receive the health care they needed due to a pre-existing condition, but for all children (and everyone else starting in 2014), they can no longer be discriminated against because of their pre-existing condition.  This is not only a win for patients, but also a win for Gastroenterologists everywhere.  Its a win-win across the board for everyone in the Gastroenterology community in more ways than just one.  Most importantly, private sector plans along with Medicare and Medicaid will now be required to cover colon cancer screening  tests with no cost to the patient. This is a great preventive step.  This is just one positive here.

However, there still is work to be done.  The American  Gastroenterology Society is busy detailing what they will be doing through their political action committee.  Most importantly, they will be trying to close a loop-hole which makes the patient pay out-of-pocket for the removal of polyps which are detected during colonoscopy.  A big thanks to the political action committee for all of the hard work they are doing over there.

But for now, let's just sit back and bathe in this win.  It's good to know that there is a small chance that this bill will get repealed.  We can talk about the ramifications later.  For now, let's just enjoy this moment.  Health Care for all at-last in the USA - what a win!

   

Great time to be a Gastroenterologist!

Well, as you all know, the Supreme Court came down very favorably yesterday for people with pre-existing conditions, the type of patients that Gastroenterologists typically care for.

Right now, I am just soaking this whole decision in right now, and I will have a very lengthy write-up later.  This has been a day-in-the-coming for many years now.  Just a great day for Gastroenterology. There should be a big write-up by tomorrow night.

In related, news, I look forward to updating this blog again more on a regular basis.  Thanks to all of you who read my blog, it's been fun!

Gastroenterologists need to be cautious of prescribing metoclopramide for non-diabetic patients

Reference: Parkman, H.P., et al., Clinical response and side effects of metoclopramide: associations with clinical, demographic, and pharmacogenetic parameters. J Clin Gastroenterol, 2012. 46(6): p. 494-503

A study out from Temple on the side effects of metoclopramide. 

Brief Summary: Gastroperesis is a disorder indicative of slow stomach emptying during digestion and is often associated with diabates.  A common drug prescribed for this condition is metoclopramide – an antiemetic drug effective against nausea and vommitting.  The main aim of this study was to determine the factors which are associated with response vs. no-response and side effects vs. no side effects with the use of metoclopramide.   100 patients from a hospital near Philadephia were accrued into the study, and clinical features were recorded including age, dosage, etiology (diabetes or no-diabetes) and genotype testing for a number of various genes though to implicate in the main find. 

Results: The patients who responsed were older and had a heavier body mass index.  Genetic polymorphisms in KCNH2 and DRA1D genes were associated with clinical response.   The patients who hads side effects to metoclopramide tended to be nondiabetic with normal gastric emptying.  Genetic polymorphsisms occurred in CYP2D6, KCNH2, and 5-HT4 receptor HTR4 genes.  Some of the side affects were very severe.      

Implications for Practice: Since some of the side effects were severe, gastroenterologists should be very wary of prescribing metoclopramide to nondiabetic patients.  For these non-diabetic patients, if genetic testing can be under-done, it may be very helpful in determing the effectiveness of metoclopramide.

Discussion: Really good paper here, and I had a really fun time reading it.  These types of studies are important in order to determine the effectiveness of various metabolic drugs.  I am glad to see that the authors included the genotype variables.   Furthermore, I really liked how the authors went in depth in the discussion section on the various genes which were under investigation and the potential biological mechanism at play.

Also, at the beginning of the discussion section, it was good to see that the authors included a brief section on the implications for practice in Gastroenterology. This is always important and something that many authors do not do.

Overall, just a very well investigated  and written paper with very interesting finds which every Gastroenterologist who prescribes metoclopramide should know really.

Commentary on Statistics and Study Design: My biggest suggestion has to with the absense of any multi-variate analysis, which should always be done in any type of risk factor study.  The authors only included a uni-variate (un-corrected) analysis, but it would also would have been helpful to conduct a multi-variate (corrected) analysis with logistic regression using either a forward or backward stepwise regression approach.  Doing this, the authors could have presented a set of clinical factors which were most associated with the outcome variable: having a positive response.  For instance, it may be that one of the clinical factors (body mass index) is not associated with the outcome variable (response vs. no-response) while controlling for one of the genes.  Due to the very large number of genes, it probably would not have been possible to include all of the genetic factors at once, but you could test just a sub-sample at a time.  You always want to conduct this type of multi-variate analysis, and report both the uni-variate and multi-variate results. 

Also, in the statistical analysis section, it seems as if three separate statistical techniques were used (Fisher’s Exact Test, ANOVA, and uni-varite logistic regression); however, it seems as if the entire paper could have been performed with just the logistic regression.  I don’t even see any of the results for the ANOVA analysis.  I’m not too sure here. It would have been helpful if the name of the statistical test used was under the tables.  This could be very helpful.   

Overall, a good investigation. Thanks to our buddies at Temple for doing this!        

Health Care Decision To Come Tommorrow

So, the big health care decision will be coming down tomorrow from the Supreme Court.   Time will be inching along until then. As expected, I will be following this decision very closely on this blog, and I look forward to the decision.  Expect a good and lengthy post on here about it some time tomorrow.  

Hopefully, the SCOTUS decides not to over-turn the part of the bill which outlaws insurance companies from denying people based on pre-existing conditions.  That is the important part which is most relevant to the Gastroenterology field.  Here is hoping!  

The Effectiveness of Fecal Immunochemical Tests May Decrease After Repeated Diagnostic Rounds For Colon Cancer

Another great experimental study out of the Netherlands on the effects of the Fecal Immunochemical Test (FIT) after repeated rounds of testing.

Summary: The main objective of the study was to determine the effects of repeated testing using the Fecal Immunochemical Tests (FIT).  Tests such as these can often have low evaluative test measures (high false negatives and positives) so it is important to do repeated testing - especially after an initial positive finding.  The study was conducted in two concurrent stages in order to determine the effectiveness .  Somewhat alarmingly, the Positive Predictive Value (PPV) was found to decrease from the first round to the second round of testing for the FIT test for those patients which had a negative result in the first round.  The PPV is an evaluative measure for a diagnostic test which divides the number of people who actually have the condition by all of those who tested positive for the condition: True Positives/(True Positives+False Positives)).  

Implications for Practice: The fact that the PPV decreased for the Fecal Immunochemical Test should come at somewhat of an alarm for Gastroenterologists. This result just helpsenforce the importance of getting a Colononscopy for patients if there is suspect of any bowel instabilities, as the the other diagnostic tests really can not measure up to a Colonoscopy.  Interestingly, the authors also found that there was no significant difference between a fecal immunochemical test (FIT) and a fecal occult blood test (FOBT: another diagnostic measure).  Thus, there is no reason to switch from a FIT based program to a FOBT based program.

Interestingly, the authors showed that there were several different sub-types of colon cancers which were diagnosed, and this suggests the potential that some sub-types may be more easily detected than others.  Another study did show that the sensitivity of the Fecal Immunochemical Test (FIT) is higher for cancers which are detected distal rather than proximal, and this may have something to do with it.  

Affordability Care Act Ruling Soon

Hey guys, sorry for not posting in some time. I have just been so busy here working on my dissertation and what-not.  But for sure, this blog is a high priority on my list, and I want to start updating it daily. I love this blog, and really hope to continue with it.

Right now, the focus is on the Patient and Affordable Care Act.  I wrote a post a few days ago on my personal opinions of the case and it's implications on Gastroenterology.  The USA Today recently had a really good article on the possible outcomes that can come of the Supreme Court's decision. Essentially, the outcome can range anywhere from the law being upheld to total defeat of the bill.  What may be likely is that the individual mandate that everyone have health insurance gets killed but the rest of the law remains including the part about illegal discrimination against those with pre-existing conditions.

I don't know what will happen, but one thing is for sure: for the good of Gastroenterology patients and doctors everywhere, it is important that the provision making it illegal for insurance companies to discriminate against pre-existing conditions to stay in place.  Here is thinking and hoping that it will stay in place.  As I said previously, it's very likely that your typical Gastroenterology patient would be prone to be discriminated against because they have a pre-existing condition. This is not good for the field of Gastroenterology, and this practice of discriminating against people because they have a pre-existing condition like Chron's Disease - through no fault of their own - needs to end.  Let's all pray that the results come out good on this.  A ruling should come out any second here.

The risk factors for an inadequate bowel preparation

Reference: Hassan, C., et al., A predictive model identifies patients most likely to have inadequate bowel preparation for colonoscopy. Clin Gastroenterol Hepatol, 2012.10(5): p. 501-6.

A study from Italy on the the factors which are most likely to lead to inadaquate bowel preperation for Colonoscopy. 

Brief Summary:   An inadaquate bowel preperation can severly affect the safety and efficacy of Colonoscopy – particularly for the detection of polyps and cancer lesions.  The main objective of this investigation was to identify the risk factors for an inadaquate level of bowel preperation.  Both patient (gender, age, etc.) and non-patient (type of regimine, time of preperation, etc) risk factors were considered.  A  total of 2811 patients were recruited for the study from a number of clinics throughout Italy.  The patients were accessed to have either an adaquate or inadaquate bowel preperation by an endoscopist at time of colonoscopy. 

Results: The patient factors deemed to be predictive of an inadaquate bowel preperation were being over weight, male sex, a high body mass index, older age, previous coloretal surgery, cirrhosis, having Parkison’s disease, having diabaetes and a negative result in a fecal occult test.   Among the non-patient risk factors, some of the factors which were deemed to be predictive were the type of bowel preperation regimine , the difference in time between the preperation and the actual colonoscopy (less time is better), and information on the preperation (oral+written info is better than just oral).   The authors found a moderate predictive accuracy on a validaiton test set (AUC=0.63).   The clinical detection rate of polyps and cancer lesions along with the cecal intubation rate was significantly higher in patients with an adaquate bowel prep vs. an inadaquate bowel prep.      

Implications for Practice: It is important that patients undergo an adaquate bowel preperation, and the probability of an adaquate bowel prep can be enhanced if the patient is given both a written and oral information for the bowel prep. 

Discussion: Really interesting finds here and very comprehensive to say the least.  It is alarming the difference in clinical detection rate for polpys and cancer lesions between adaquate vs. inadaquate bowel preperations.  For adaquate bowel preperations, the rate was 41% and for inadaquate bowel preps, the rate was 35% which is a rather significant difference. 

Some of the risk factors that came up positive were rather surprising.  It is interesting to see that males and those with Parkinson’s and diabates had a higher chance of an inadaquate test.  It would be neat to do a follow up explanation on many of the risk factors that came up.  It seems like the best bowel prep regimine was the Sennosides.  This seemed to be better than the use of any regimines which used the polyethylene glycol (PEG) solution or sodium phosphate. 

Commentary on Statistics and Study Design:  The study design and presentation was very good. This is an exploratory based perspective study.  In any type of exploratory based study, there are often many variables at play in the model building step, and the investigators handled them well here.  I particularly like how the authors included the results of both the uni-variate and multi-variate analysis for the logistic regression. It is always helpful to see the results of both, because it gives the reader a sense of what variables are independently and also dependently associated with the outcome variable.  The multi-variate results also tell you which variables (or risk factors) would be most likely to be associated with the outcome variable.  The only suggestion I have for the authors here would be to also include the odds ratios for the uni-variate analysis.  This could help the reader and statisticians like me.

            Going on, I also liked how the authors left out data for a validation sample.  In many exploratory based investigations, this is not done. Fortunately, the authors were able to collect a large among of data points from a quite a few clinics in order to do this, and this is why collecting a sufficient data sample size can always be a big plus.  In short, the investigator should always collect as much data points as possible for the investigation - regardless of what the power sample sizes might tell you.  More is always better if possible.      

            Overall, great layout and study design for this investigation.  I really don’t have too much to suggest.  I always try to assemble in my mind the best study designs and layouts of various papers in order to bring everything together, and this is definitely a paper I hope to reference in the future. 

Patients over age 60 are considerably more likely to have a 30-day mortality

Reference: Bae, S., et al., Incidence and 30-day mortality of peptic ulcer bleeding in Korea. Eur J Gastroenterol Hepatol, 2012. 24(6): p. 675-682.

A study out from Korea on the incidence and 30-day mortality rate of peptic ulcer bleeding in Korea.

Brief Summary:  A peptic ulcer is an ulcer which occurs somewhere in the gastrointestinal tract (large or small intestine) and can provide complications in the form of peptic ulcer bleeding (PUB).  The main objective of this study was to estimate the rate of peptic ulcer bleed bleeding and the risk factors for 30 day mortality from a PUB event.  An improved PUB diagnostic algorithm which relied on a prescription of histamine type-2 receptor antagonists (H2RA) or proton pump inhibitors (PPI) was used to diagnose the PUB patients.

Results:  On the basis of the aforementioned diagnostic algorithm, the incidence rate for PUB was 22.1 per 100,000 individuals.  The 30-day mortality rate for patients with presenting PUB over 80 was 7.65%, between the ages of 60 and 79, 2.87%, and for those less than 60, 0.83%. The overall 30 day mortality rate for all patients was 2.15 percent.  Overall, in a uni-variate analysis, the risk factors which predicted a greater likelihood of mortality for PUB were over age 60, female sex, ulcer-related drug use (aspirin, oral glucocorticoids, vitamin K antagonists,etc.), and antiulcer drug use (proton pump inhibitors and H2 receptor antagonists).   When adjusting for all the factors, only age was found to be a significant 30-day risk factor.  

Implications for Practice: Doctors need to be more wary of patients who are older than 60 and present with a peptic ulcer bleeding, because these patients are considerably more likely to have a 30-day mortality event. 

Discussion: Really interesting study here.   The incidence of gastrointestinal conditions can really vary from one geographical region to another, because the environmental and dietary variables can really vary from one region to another, and thus studies like this is important.  Whenever this geographical dependency occurs, it is important to limit the analysis to a certain sub-population in the world as this investigation did.   

It was surprising to see that the PPV jumped so significantly with the use of a more restrictive diagnostic algorithm (prescription of the PPI or H2RA).  The PPV went from 53 to 88 percent – astounding increase! I am interested to see whether other epidemiological based studies could take advantage of this prescription fix. 

Also, I liked the discussion section that the author’s wrote-up.  In short, they gave a detailed but succinct description of the various risk factors, and their influence on the eventual disease.   The authors noted that patients of age greater than 75 are more likely to have a peptic ulcer than those between the ages of 25 and 44 due to consumption of NSAID/aspirin or the high prevalence of H. pylori among the elderly.  This would seem to be the main causative factor here.      

Also, the author’s noted that the result for gender was different for this study compared to past studies.  Past studies showed that males have a higher incidence rate due to higher rates of alcohol consumption and smoking among men, which are known to be risk factors of PUB. It seems as if a significantly more work needs to be done to elucidate the exact effect of gender here – especially given the confliction results from this investigation to past.  It no doubt has something to do with the inherent differences in the sub-populations under study.    

Commentary on Statistics and Study Design: Overall, the investigators did a really good job with the statistical analysis.  I like how the author’s reported both the un-adjusted (uni-variate) and un-adjusted (multi-variate) results.  This is really important in a study like this, because it’s important to give the reader a final set of risk factors (variables) which are most associated with the response (morbidity).   For instance, it’s highly suspicious that the use of an ulcer-related drug – which was significant in the uni-variate analysis – would also be significant in the multi-variate analysis when also adjusting for age. 

Also, I really like how the authors used the Charlson co-morbidity index and used this as a controlling factor multi-variate analysis. This is obviously crucial to do, since the presence of another disease could seriously confound the experiment.  Typically, many investigators would just totally exclude any patients from the study all-together with the presence of another disease, but in doing this, the analyst is essentially losing data samples. It is better just to control for the confounding factor if possible. 

I guess my only suggestion for the authors would be to include a more explicit description of the odds ratios which were presented in Table 3 which detailed the individual risk factors for the un-adjusted (uni-variate) model.  For instance, say something like “the odds of a 30 day mortality increase by 350% for patients who are between the ages of 60 to 79 compared to those patients who are less than 60.”  This type of succinct yet explicit detail can be really helpful.  Other than that, everything looks good!  I really liked the tables and figures – not too much to say there.

As the authors alluded to in the discussion section, there are several lifestyle variables (smoking, alcohol consumption, etc.) which could have been included in the study.  In fact, the introduction of these lifestyle variables would probably change the relationship with 30-day co-morbidity significantly.  For instance, I bet you that sex will not even be significant if you introduce these lifestyle variables, but this is another good study for the future.

A big thanks to our Korea pals for doing this!        

    

  

How will the new Patient Protection and Affordable Act Affect Gastroenterologists?

If you have been following this blog, you know that this blog tends to be more of a professional related blog.  Most of the posts are summary reviews of research articles, and this is how this blog will be continued into the future.  However, I do like to mix it up a little bit, and this post will serve as a good example.

Most notably, the entire country right now - especially the entire health care community - is on edge about the Supreme Court's ruling on the Patient Protection and Affordable Care Act (PPACA) which should be released some time later this month most likely.  It is no exaggeration to say that this decision is the most influential one to come down from the Supreme Court since perhaps Bush. vs. Gore back in 2000. And before that decision, you have to go way back to 1972 for the Roe vs. Wae decision to find any decision that has real influence on your typical American. In short, there are few rulings from the Supreme Court of the United States (SCOTUS) that have any impact on your average American - once every 20 years maybe - but this one definitely does have a major impact on everybody.

I have been following this case extremely closely, and I have absolutely no idea how the court will eventually rule.  The general consensus is that the Obama administration lawyers did not rack it up well in the 3 days of arguments at the end of March against the opposing lawyers.  Particularly, the questions from the 5 conservative judges were rather harsh.  However, given the nature of the case, one would expect tough questions regardless for the PPACA, so you can't put a whole lot of weight on the questions necessarily. The way it breaks down is this: you have 4 justice which will definitely vote to uphold the law, 1 justice who will definitely vote against it, and 4 other conservative leaning justices who could go either way.  The key here is that the Obama administration only needs 1 of those 4 justices to vote for it,  and I have a feeling this will occur - although I am not as confident as I was two months ago before the hearings began.  If I had to guess, I would say there is a 75% chance that the SCOTUS will up-hold the PPACA when the ruling comes out later this month.

But anyway, enough legal rambling, this post will assume that the PPACA will in fact be supported by the SCOTUS.  So, a good question is this: how will this ruling affect your average Joe Gastroenterologist?  A-lot has been written about how it will affect the patient, and this is a good article from the CCFA on the ramifications from a patient's perspective.  

My answer to this question is this: anything that is good for the Gastroenterologist patient is also good for the Gastroenterologist.  Simple and straight forward as this really.  It is no secret that your typical Gastroenterologist patient would probably be prone to be lumped in that group of people who have a "pre-existing condition" and get discriminated against by insurance companies and subsequently be denied coverage.   I have no statistics to back this up, but I would guess there is a rather large pool of patients out there with Gastroenterology problems who limit (if not completely) their access to a Gastroenterologist in order to reduce their own personal cost. This is bad for the patient, bad for the Gastroenterology community, and also bad for Gastroenterologists in the end.  These are patients which could be brought into greater Gastroenterology care and help support your typical Gastroenterology practice.

So, in short, all Gastroenterology related people should be hoping that the PPACA gets passed. It would be good for the patients and also good for your typical Gastroenterologist.  Feel free to comment on this post with any thoughts or opinions. Here is hoping 9 justices don't think otherwise.

The Clinical Factors associated with upper gastrointestinal hemorrhage

Reference: Freedman, S.B., et al., Predictors of clinically significant upper gastrointestinal hemorrhage among children with hematemesis. J Pediatr Gastroenterol Nutr, 2012. 54(6): p. 737-43.

A study out from Toronto on the clinical risk factors which predict a significant upper gastrointestinal hemorrhage. 

Brief Summary: Hematemesis is the vomiting of blood and there is currently no study which details the clinical findings which predict upper gastrointestinal hemorrhage (UGIH) – defined by bleeding in the upper gastrointestinal tract and which often requires surgery.  The main objective of this retrospective study was to determine the percentage of children with hematemesis who have UGIH and to identify the clinical features which predict UGIH.  A total of 613 eligible children with determined hematemesis from a tertiary care center in Toronto were accrued into the investigation. 

Results:  A total of 27 of the 613 hematemesis patients (4%) had upper gastrointestinal hemorrhage (UGIH).  The clinical features which were deemed to be predictive were: older age (9.7 vs. 2.9 years), vomiting moderate to large amounts of fresh blood, melena, significant medical history, unwell appearance and tachycardia.  Furthermore, children with a medical history of esophageal/gastric varices and a low hemoglobin and platelet count in the blood are also highly indicative of UGIH. 

Implications for Practice:  A patient with hemorrhage and the following clinical findings (older age (9.7 vs. 2.9 years), vomiting moderate to large amounts of fresh blood, melena, significant medical history, unwell appearance and tachycardia) should be wary of UGIH and perform the appropriate diagnostic tests.  If a child presents with high risk features and there is suspicion of UGIH, then a complete blood count is the only test likely to yield clinically helpful information.

Discussion: Good study here. It is very important that Gastroenterologists know these types of clinical risk factor statistics like a baseball manager knows his own players. 

It was good to see that the author’s included the laboratory investigative tests such as hemoglobin and blood count.   You don’t see this in many clinical risk factor studies. 

Commentary on Statistics and Study Design:  The author’s did several things in this study that I really like. I really liked the sample size calculations – good to see that. Also, the analyst checked for normality and used a different test based on whether the data was normal or not.  Also, it was good to see a correction for multiple comparison testing – you rarely (if ever) see this done, even though it should be in most risk factor studies.

                I guess my first major suggestion comes in the form of a presentation standpoint.  The authors essentially broke up the clinical factor results into historical and clinical features (Table 2), medical history (Table 4), and laboratory investigations (Table 5). Personally, I would have combined all these results into one table.  In the table, one could then just have demarcated the different types of clinical features.  It is just a personal preference of mine – it’s fine either way really.

                My other suggestions have more to do with the actual statistical analysis.  The authors did not perform a logistic regression due to the perceived class imbalance of 4% (positive for UGIH) vs. 96% (negative for UGIH) for the response variable. However, I don’t think this should have prevented the use of the logistic regression model. Generally, as long as the class distribution is the same as the total population (which it’s assumed that it would be?), then only having 4% as the positive class would be OK.  You want to read the paper titled “Sampling Bias and Class Imbalance in Maximum-likelihood Logistic Regression” by Freedman et. al. which explains all this well.  Conversely, what the author’s could done is compared the results from the logistic regression to the 2 sample statistical methods that were used (Mann Whitney and 2-sample t-test).  The results should have been roughly the same.  If they were, then it would be OK to just go ahead and use the logistic regression.  In fact, the 2 sample statistical methods would be just as susceptible to a class im-balance as the logistic regression technique if I’m correct – there is really no way around this.

                Also, I see that the author’s excluded blood pressure and oxygen saturation due to missing data with these variables.  This was probably the best approach here, assuming a large percentage of the data was indeed missing. Generally speaking, if you have missing data for greater than 5% of the data for a given variable, then it is best to just remove the variable if possible.

                As the author’s noted, there is a-lot of potential selection and measurement based limitations with this investigation, and I won’t go into all of them – the author’s would certainly be much better at identifying the appropriate variables than I would.  It’s always important that you be as précises as possible with the measurements on both the predictor and response variable.

                A big thanks to the guys up in Toronto for doing this!        

Quality of life factors are reduced in patients with Hepatitis B

Reference: Wang, L., et al., Quality of life and the relevant factors in patients with chronic hepatitis B. Hepatogastroenterology, 2012. 59(116): p. 1036-42.

A little bit different of an article that I am used to publishing on this blog, but I like difference.  Here is a study quantifying the quality of life factors for patients with Hepatitis B. I can’t tell you how much fun I had reading this article!    

Brief Summary: Hepatitis B is an infectious inflammatory disease of the liver caused by the Hepatitis B virus.  Currently, little is known regarding the health-related quality of life (HRQL) factors which are affected by Hepatitis B.  The main objective of this study was to access the HRQL factors for patients with Hepatitis B.  The HRQL factors were physical functioning (PF), role physical (RP), bodily pain (BP), general health (GH), vitality (VT), social functioning (SF), role emotional (RE) and mental health (MH) from the Chinese version of the medical outcomes study 36-Item Short-Form health survey.  A total of 407 patients were accrued into the study, and they were compared to a literature based control group.  Secondary objectives were to ascertain the effect of genotype differences in the ACE and DRD4 genes on the HRQL factors, and also to study the effect of anti-viral therapies on some of the HRQL factors.  These factors included physiology function (PHD) and psychology function (PSD)

Results: The HRQL factors were significantly lower for the patients with Hepatitis B compared to the literature-based control group, and they were lower for all 8 previously mentioned factors.  Furthermore, the genotype of the ACE and DRD4 genes were found to be associated with HRQL, and the anti-viral therapies were found to improve several HRQL factors which included physiology function and society function.   Other therapies (hepatoprotective, jaundice eliminating and supportive treatment ) were not shown to add an improvement.   

Implications for Practice: Patients with Hepatitis B have significantly lower quality of life factor scores.   If the Hepatitis B patient has a highly negative physiology function psychology function, then anti-viral therapies can help improve these quality of life factors.     

   

Discussion: I found it really interesting how genotype can have such a drastic effect on the quality of life factors.  I imagine this relationship would hold for other types of Gastroenterology based diseases as well. Just very fascinating.   

In the discussion, the authors stated that clinical therapy has a minimal effect on the emotional state of a patient Hepatitis B.  Rather, family and community care are more effective than any therapy at increasing the quality of life. All of this is probably true, and I’ve seen some past quality of life studies say the same thing.

 

Commentary on Statistics  and Study Design:  Overall, the statistical analysis was very good. I like how the author’s performed both a uni-variate and multi-variate analysis between the factors and the main response variable which was the HRQL scores.  Since the version I read was not in current print, I could not see the figures, but I assume that both the uni- and multi- variate analysis was shown.  Also, it was good that the author’s pointed out the limitation with using a literature-based control group.  This is a definite limitation. I didn’t see a sample size for the control group, but I’m assuming it was rather large, so this technique should have been fine. 

My one major suggestion would be in the interpretation of the coefficients. I am assuming that the multi-variate table (which again, I couldn’t see) included the coefficients from the multi-variate analysis. If it did not, then it needs to be there, and there should be an explicit and detailed explanation of the coefficients. For instance, if the coefficient is -20.0 for one of the factors (let’s say physical functioning), then an explicit interpretive statement needs to be stated such as “this coefficient means that the patients in the Hepatitis B group had a physical functioning quality of life score that was 20 less those patients in the control group, and this coefficient is significantly different than 0.0.”  This type of explicit and descriptive statement can really help out your non-statistician tremendously, and remember, your average Joe doctor is not a statistician.

Other than that, everything looks good. I assume the sample size was good enough. 

It was good to read a paper way over in China. Thanks!             

Concomitant Gastroparesis has a negative affect on children with Biliary Dyskinesia

Reference: Chumpitazi, B.P., et al., Concomitant gastroparesis negatively affects children with functional gallbladder disease. J Pediatr Gastroenterol Nutr, 2012. 54(6): p. 776-9.

A study out from Baylor on how concomitant gastroparesis (CG) can negatively affect biliary dyskinesia (BD). 

Brief Summary: Gastroparesis is a medical condition of the stomach where the food remains in the stomach for too long. It is thought that this condition can negatively effect patients with biliary dyskinesia (a disturbance of the biliary ducts).  The main objective of this retrospective study was to determine the prognostic effect that concomitant gastroparesis (CG) has on biliary dyskinesia (BD).  A total of 35 children were accrued into the investigation who had biliary dyskinesia (BD).  Of these 35 patients, 20 had CG, and the other 15 did not have CG.  The time of the last follow-up was then checked, and assigned to 4 clinical outcome groups based on the symptoms: poor, fair, good, excellent.  The children with CG were then compared to the children without CG. 

Results: The children who had CG were more likely to have un-favorable clinical outcome (defined as poor or fair) than children who did not have CG.   Furthermore, in children undergoing a cholecystectomy, the clinical outcomes were worse for children with CG.   

Implications for Practice:  Children who have both CG and BD need to be monitored closely as their clinical outcome may be worse than children who only have BD.  Particularly, the Gastroenterologist needs to be wary of mis-diagnosing CG patients as BD.   

Discussion:  Interesting study here on the interaction between CG and BD on clinical outcome.  I liked how the authors gave an explanation for the reason why children with both CG and BD have a less favorable outcome.  The authors noted that this may be due to miss-directed therapies.  Therapies that are directed solely against BD may not address the symptoms against CG.    

It is interesting to see that the author’s used schools absences as an indicator for limitations in activity.   I have never seen this variable used before to determine the level of severity, but I suppose that it would be a good indicator of severity.  As good as anything else I suppose for a more objective indicator.  

Commentary on Statistics and Study Design:  I guess I have two major suggestions for the authors: one related to the statistical technique used and one with the presentation of the results. 

First, let’s start with the statistical technique. The author’s took what was originally a 4 category response variable for the clinical outcome (poor, fair, good, excellent) and converted it into a 2 category (binary) response (favorable and un-favorable) and then performed your typical 2 category logistic regression analysis.  However, the author’s just have used a multi-variate outcome logistic regression technique for an ordinal response.   This is called polytomous logistic regression.  Polytomous logistic regression is incredibly easy to implement in SPSS.   In fact, the interpretation is just about exactly the same as your standard binary response logistic regression.    But back to the point: running this polytomous logistic regression would have prevented the loss of information from going down to 2 categories from 4 with the response variable.   The authors then could have reported the individual p-values for each level of the response variable.  Simple and easy to do really. 

As for the presentation, the author’s performed a multi-variate regression technique, but it would have been helpful to see the results in more of a table-like format.  I wish I would have my primer up already on how to visually report statistical results up correctly in a publication, but alas, in my busyness I don’t. The typical procedure here is to report both the uni-variate and final multi-variate (adjusted) results.   The reason you want to also report the multi-variate results is that it gives the reader a good idea of what final predictors are most indicative of outcome.   Come back when I have this final primer up – it could have helped this paper tremendously. 

Also, it may have been helpful to obtain more samples.  Given the small sample size - 35 children - it may have been difficult to make confident conclusions. This is one limitation that the author's should have probably noted at the end.  

A big  thanks to our Baylor pals for  doing this. Love RG3 – go Skins!            

Probe-Based confocal laser endomicroscopy shown not to be as effective as other advanced imaging techniques for the differentiation of colorectal lesions

Reference: Kuiper, T., et al., Feasibility and Accuracy of Confocal Endomicroscopy in Comparison With Narrow-Band Imaging and Chromoendoscopy for the Differentiation of Colorectal Lesions. Am J Gastroenterol, 2012. 


A study out of the Netherlands on the effectability of various advanced imaging techniques for the differentiation of colo-rectal lesions.

Brief Summary: The main objective of this study was to compare the accuracy of several advanced imaging techniques for the differentiation of colon cancer. Particularly, the advanced imagine techniques compared were Chromoendoscopy (CE), Narrow Band Imaging (NBI), and probe-based confoscal laser endomicscropy (pCLE).  Specifically, two different variations of the Narrow Band Imaging were analyzed - one that uses the Kudo pit pattern (NBI-Kudo) and another that used the vasculor pattern intensity for differential diagnosis (NBI-VPI).  

Results: The authors found that the pCLE imaging technique was significantly less effective than any of the other advanced imaging techniques (CE, NBI-Kudo, NBI-VPI). Specifically, the authors later showed that the CE and NBI-Kudo methods were the two best advanced imagine techniques.  

Implications for Practice: For the purpose of advanced imaging, Gastroenterologists can not yet place a high degree of confidence on the probe-based confoscal laser endomicscropy technique  for the differential diagnosis of colon lesions. Rather, they should place a greater priority on the Chromoendoscopy and the Narrow Band Imaging with the Kudo accessment advanced imaging techniques for the purpose of differentiating colon lesions.  

Brief Discussion: One of the advantages of the pCLE method is that it can provide a direct diagnosis of individual cells without the need for any biopsy - thus reducing the total time and cost of a colonoscopy considerably. However, in order to use the pCLE method, it needs to show a greater degree of effectiveness than other advanced imagine techniques, and this study shows that it does not.  

It was interesting to note that the authors found that the accuracy of the pCLE method increased for the high quality videos vs. the lower quality videos.  If more attention could possibly be put towards getting higher quality videos, then this could significantly enhance differential diagnostic accuracy.       

Brief Commentary on Statistics and Study Design: Overall, the statistical design of the experiment looked good.  The statistician decided to use the McNamara's test in order to compare the contingency tables derived from the various imagine techniques, which is the correct test to use for this.  One can obviously not assume that the evaluative results from the various imaging techniques are independent, so the McNamara's test is the correct one to use.   

It would have been interesting to see if the ratings from the different imaging techniques were different depending on various patient characteristics. For instance, would a certain test be more effective than another test based on the age of the patient? Common sense would probably say no, but it would still be interesting to see.

Twice-Daily Dosage of Rabeprazole (RPZ) found to be more effective than a Single-Daily Dosage for Severe Reflux-Esophagus Disease

Reference: Kinoshita, Y. and M. Hongo, Efficacy of Twice-Daily Rabeprazole for Reflux Esophagitis Patients Refractory to Standard Once-Daily Administration of PPI: The Japan-Based TWICE Study. Am J Gastroenterol, 2012.

A study out of Japan on the effect of a twice-daily dosage vs. a once-daily dosage for patients with severe Reflux Esophagitis disease who do not initially respond to a proton pump.  

Brief Summary:  About 10 percent of Reflux Esphagitis (RE) patients do not initially respond to proton pump inhibitors (PPIs) - especially common in those with severe RE.  The main objective of this study was to determine whether increasing the effective dosage of a type of PPI called rabeprazole (RPZ) would lead to significant healing.  Specifically, it was desired to see if a 10 or 20 m.g. b.i.d. (twice-daily) regime dosage is more effective than a 20 m.g. q.d. (once-daily) regime dosage of RZ.  About 330 patients were randomly assigned to three different groups consisting of a 10 m.g. b.i.d., 20 m.g. b.i.d., or 20 m.g. q.d. regime dosage.   The primary comparison end-point between the three groups was the rate of endoscopically confirmed healing after 8 weeks, and a second end-point comparison was the weekly heartburn resolution after 4 weeks.    

Results: In general, a regime of 10 or 20 m.g. twice-daily dosage of RPZ was found to be more effective than a 20 m.g. single-daily dosage among patients who did not respond to initial treatments of PPI. Specifically, patients with Los Angeles grade A or B RE were found to respond optimally to a regime dosage of 10 m.g. twice-daily RPZ whereas patients with Los Angeles grade C or D RE were found to respond optimally to a regimen dosage of 20 m.g. twice-daily RPZ.  The heartburn resolution also decreased for the two twice-daily regimens.  The Los Angeles classification system is typically used to classify Esophagitis patients into four different groups based on severity.  Group A is the least severe and Group D is the most severe.   

Implications for Practice: If a patient does not respond initially to the standard proton pump inhibitor (PPI) regimen, then the patient should be prescribed either a 10 or 20 m.g. b.i.d. regimen of Rabeprazole (RPZ). If it is known that the patient has type A or B, prescribe 10 m.g. twice-daily of RPZ; if it’s known that the patient has type C or D, prescribe 20 m.g. twice-daily of RPZ.

Discussion: More studies like this need to be done in order to elucidate the optimal dosage for a given condition.  It was interesting that Los Angeles group A and B patients were found to have a different optimal dosage than group C or D patients.

Brief Commentary on Statistics and Study Design: The authors used the Cochran-Armitrage test which can be used for this purpose. However, since the outcome variable was continuous, it would have been more simple and easy just to use simple general linear regression (GLR) or an ANOVA model, and this is what really should have been used. The advantage of using a GLR is that it is fairly simple to control for any other potential confounding variables.  Indeed, there are several other demographic variables that could have been included.  The ANOVA model could have then been used to compare the means from the three different groups and are setup specifically for experiments like this.  In statistics (as in most anything in life), the simplest solution is usually the best.  

Pancreatic stents (PS) should not be left in situ for more than 7 to 10 days during endoscopic retrograde cholangiopancreatography

Reference: Hill, S.K., C. Bhalla, and A. Thomson, Risk of bacterial colonization of pancreatic stents used in endoscopic retrograde cholangiopancreatography. J Clin Gastroenterol, 2012. 46(4): p. 324-7.

Brief Summary: Pancreatic stents (PS) are often used during endoscopic retrograde choloangiopancreatography to reduce the risk of pancreatitis.  However, if left in situ for more than 2 weeks, there is evidence that they may lead to pancreatic duct injury via micro-organism colonization.  The main objective of this study was to investigate the incidence of pancreatic stent colonization and study the relationship between the pancreatic stent colonization with type, length, duration and time the PS remained situ.   The authors were able to obtain 47 pancreatic stents in situ, and these stents were then sent away for microbial culture where the pancreatic stents were then classified as being either significant or insignificant based on their microscopic colonization.  The cultures were classified as being insignificant if there was no growth of pathological organisms with clinical significance. 

Results: Pancreatic stents left in situ for more than 10 days were found to have a 3.3 times greater risk of harmful microscopic colonization compared to just 2 days.  There was no significant relationship found based on type or length of pancreatic stent. 

Implications for Practice: Current recommendations suggest that pancreatic stents remain in place for no more than 14 days. However, this study shows that the Pancreatic Stents should not remain in place for more than 10 days.  More generally, the Pancreatic Stents should be removed as soon as their preventive value for post-ERCP pancreatitis is exhausted. 

Discussion: Studies like this which detail and optimize best practices are important and more should be done.  It would be interesting to find out why 10 days seemed to be the cutoff point.  This probably has to do with the rate of colonization of the bacteria, and there are probably some interesting follow up studies to elucidate this.

Brief Commentary on Statistics and Study Design:  The overall study design and statistical approach seemed to be good.  A multi-variate logistic regression model was fitted which is the correct statistical model to use for this.  It would appear as if the predictor variables in the model were duration and length of the Pancreatic Stent, and all other variables were thrown out of the logistic regression model due to their non-significance with the outcome variable.  The authors did say that the other variables (age, gender, indication, or diagnosis) had no significant correlation with bacterial colonization, and this was harmful.

I am curious as to why there were no power calculations done.  I assume that 47 samples would be ample though. 

A report on the incidence and mortality of liver cirrhosis - underscores why early diagnosis is important -

Reference: Fialla, A.D., O.B. de Muckadell, and A. Touborg Lassen, Incidence, etiology and mortality of cirrhosis: a population-based cohort study. Scand J Gastroenterol, 2012. 47(6): p. 702-9.

It’s not often I get to comment on an epidemiological based study, so this is exciting!  A study from Finland on the case incidence of cirrhosis. 

Brief Summary:  The main objective of this study was to detail the incidence and mortality of cirrhosis patients.   A total of 1369 patients were identified as having cirrhosis from the discharge diagnoses from the Funen Patient Administrative System (FPAS) from Funen, Denmark.  The patients were identified and differentiated based on their age, etiology (autoimmune hepatitis, alcohol cirrhosis, viral hepatitis, cryptogenic hepatitis, etc.), gender, and complications at diagnosis (present, absent).  The incidence and mortality rates were then calculated for each of the previously mentioned clinical variables. 

Results: The incidence was twice as high for women as it was for men with all age groups.  As for mortality, the mortality rate was significantly higher for men, age above 70 years compared with other age groups, complications present at diagnosis, and etiology due to alcoholic disease or cryptogenic cirrhosis. 

Implications for Practice:  If a patient is diagnosed with symptoms and is male, age > 70, has complications and an etiology due to alcoholic or cryptogenic cirrhosis, then there is a higher mortality rate.  Most importantly, the authors noted that it is important to diagnosis patients as early as possible for cirrhosis, because the likelihood of diagnosis will decrease significantly.  This was notable since patients which had complications at time of diagnosis (and thus a more further-along disease progression) had a 28% higher odds of mortality than those who did not initially present with complications.     

Discussion: Good study here. Like I said before, I don’t get to review epidemiology based papers too much, so it was good to do one finally.  I can’t tell you how fun it was to read this paper!  It was interesting to find that males have a higher rate of mortality than females.  I am curious as to why this is. 

                As the authors mentioned in the discussion section, the time of diagnosis is obviously critical here; patients with alcoholic and cryptogenic cirrhosis generally have a worse mortality rate than those with other etiological based causes, because their disease-progression is generally worse at initial diagnosis.  It was stunning to see that of the 1369 cirrhosis patients admitted into the sample cohort, 1076 (~79%) were alcohol induced! 

Commentary on Statistics and Study Design:  Overall, good statistics and study design here. I would have analyzed the paper in a very similar fashion to how the authors actually did it.  I especially like how the authors included both a uni- and multi-variate analysis, and reported the final results (Table 4).  I would have removed any variables which were significant in the multi-variate model to produce more precise odds ratios for the most significant variables, but since there were so few variables in the analysis here, the approach the authors took is OK.  It would not have resulted in a big difference.

                In the methods section, the authors said that they tested for interactions, but they did not report on them in the rest of the paper. I assume this means those interactions were not significant?  This would be useful information to know.  Other than that, great paper overall.

                A big thanks to our buddies in Denmark for doing this!  Hope to get there one day.       

The distribution of fat in the visceral and gastroesophageal junction is associated with BE

Reference: Nelsen, E.M., et al., Distribution of Body Fat and Its Influence on Esophageal Inflammation and Dysplasia in Patients With Barrett's Esophagus. Clin Gastroenterol Hepatol, 2012.

A study out of the Mayo Clinic in Minnesota on the distribution of body fat and its influence on  Barret’s Esophagus. 

Brief Summary: It is known that body fat can increase the risk of Barret’s Esophagus (BE). However, it is not known how the distribution of the body fat (visceral, abdominal, etc.) in particular influences Barret’s Esophagus.   The main objective of this retrospective study was to study the distriubution of the fat in patients with Barret’s Esophagus – particuarly with respect to the distriubtion of the fat.  A total of 100 pateints were sampled for this study from the Mayo clinic: 50 of those with BE (case) and 50 without BE (controls).  Fat measurements were then taken from all of the patients according to the following speicific fat regions: subcutaneous fat (SF), visceral fat (VF), and gastroesophageal junction (GEJ) fat.  

Results: The distribution of fat in the visceral and gastroesophageal junction region is signficantly associated with Barret’s Esophagus (BE).     

Implications  for Practice: Patients with a high fat content in the gastroesophageal and visceral regions are at a higher liklihood of acquiring Barret’s Esophagus (BE) if shown in unison with other symptoms. 

Discussion:  A good observational study which shows how the distribution of fat influences BE.   I liked how the authors talked about the potential mechanisms between the body fat distributions and their influece on BE – many authors do not do this.   

In order to further elucidate the relationship between fat and BE, it would be nice to see a comprehensive multi-variate study which includes the full array of interaction variables.  For instance, the authors mentioned that BMI (body-mass index) alone does not explain the male and Caucasian pre-dominance of BE in males, and the authors further go on to explain that this may be due to males and Caucasians having a greater proportion of fat in the abdominal area (visceral and/or subcutaneous fact).  All this being well known, it would be nice to see if there was an interaction between gender and nationality in conjunction with fat distibution area. For instance, does the effect on abdominal fat tendencies with Barreth's E ?  These types of questions can be answered by throwing interaction terms in the model  and by utilizing a multi-variate regresssion technique which I will get into with the next section.

Commentary on Statistics and Study Design:  I really liked some of the statistical consideratiosn in this papers.  Most notably, it was good to see that the authors used the concept of matching.  Matching is done to balance out the effects of any confouding variable in an observational study and should always be done if the investigator is highly suspicious of any confounding variable in the study.   This is especially pertinent to this study since it’s thought that gender may be highly associated with BE.  You don’t see many investigators do this even though it should be done for many investigations.   Also, I liked how the author’s gave explicit detail on the fat measurements in the Methods section.     

                However, I would have taken a different statistical approach to analyzing the relationship between the fat distribution variables and the BE response variable.  The authors investigated the uni-variate and bi-variate relationship (while controlling for BMI) between the fat distribution variables and the BE response variable.  In doing this, the author’s have obviously flagged BMI as an essential variable due to past literature knowledge for BE, and this would  seem to be correct.  However, for the fat distribution variables, I would have investigated the nature of the relationship between the fat distribution variables and the BE response while controlling for all variables – not just BMI.  This can be done with a typical multi-variate regression technique which includes all variables, and then the investigators could have simpy eliminated the un-necessary variables from the model in a backwards-stepwise approach.  At the end, one would then report both the uni- and multi- variate results.  Taking this approach would allow the investigator to come to a small set of variables which are most associated with BE. For instance, if may be that GEJ fat is not associated with BE while also controlling for Visceral fat, and GEJ fat can thus be eliminated as a significant predictor variable.  Furthermore, using this approach, one could also look at the presence of different interaction variables - which as I alluded to in the discussion section - may be present here.  For whatever reason, many investigators do not include the presence of interaction variables even though they should.  

                From a presentation standpoint, for the second aim in Figure 3, the author’s decided to present a box and whisker plot instead of a table.  Since one is only looking an uni-variate associates here, this seems to be OK.  Personally, for uni-variate assocations, I like the use of a box-and whisker plot more than a table since the actual numbers are visually more explicit, but either one works well.  As long as the knowledge comes across, it’s all good.

                A big thanks to our Mayo Clinic buddies for doing this!

      

The Probiotic Saccharomyces boulardii does not provide any significant benefit in the alleviation of Antiobotic-associated diarrhea (AAD)

Reference: Pozzoni, P., et al., Saccharomyces boulardii for the Prevention of Antibiotic-Associated Diarrhea in Adult Hospitalized Patients: A Single-Center, Randomized, Double-Blind, Placebo-Controlled Trial. Am J Gastroenterol, 2012.

A study out of Italy describing the effects of the probiotic Saccharomyces boulardii in the prevention of Antiobotic-associated diarrhea (AAD) and Clostridium difficuile-associated diarrhea (CDAD). 

Brief Summary: Antibiotics can effectively be used to mitigate several Gastroentestinal diseases, but one of their negative side effects is Antibiotic-associated diarrhea (AAD) and Clostridium difficile-associated diarrhea (CDAD).  In the past, probiotics have been effective used to prevent AAD and CCD. The main objective of this study was to test the effectiveness of the probiotic Saccharomyces boulardii on the occurrence of AAD and CDAD.  A total of 562 patients were accrued in a double-blind, placebo controlled study. The accrued patients were assigned either the S. boulardii capsule or the placebo.

Results: The patients who received the S. boulardii capsule (the probiotic) did not incur a significant reduction in the incidence of AAD or CDAD compared to the plecebo.

Implications for Practice: Gastroenterologists should not prescribe priobiotics for the purpose of alleviating AAD or CDAC symptoms.

Discussion: Previous studies have shown that probiotics can be used to decrease the incidence of AAD and CDAC, and it is interesting that this study came out with a negative result.  It is important that all results - both positive and negative – such as this one be published to increase the overall understanding. 

It is interesting that the only patient characteristic found to be associated with AAD was a low level of serum creatinine.  Another retrospective analysis did find an increased level of serum creatinine for patients with extreme AAD.  More studies will need to be done to elucidate the cause and effect relationship between AAD and serum creatinine.  It would be interesting to look at the cause and effect relationship at play here, especially on a more of a molecular biology level.

Commentary on Statistics and Study Design: The statistics and study design from this paper was well planned out. The author’s used a multi-variate logistic regression model to study the main effect between the treatment and the outcome, which is the correct model to use.  The advantage of using this model – over the typical univariate statistics – is that other variables can more easily be controlled for.  Also, I liked how the authors included clear and precise power calculations.  And of course, the double-blind nature of the study alleviates any measurement bias concerns – always the gold standard in medicine.

As for the previous studies detailing the relationship between AAD and probiotics, the authors said that these studies have had several methodological drawbacks including “non-calculated sample size, inadequate study power, lack of generalizibility of the results due to manner of the selection of the enrolled patients, and unspecified information about the application of non-pharmacological infection control measures”.  I have not had the chance to go back and read these papers in depth to verify all of this (I will in the future soon), but the author’s points are well noted – especially the concern over the manner in which the patients were selected. One of the biggest study design errors I see in medical studies such as this is a mistake in the population which is samples. In any sort-of statistical inference study, you want the population you draw from to be as similar to the general population as possible. The reason being is that the nature of the statistical relationship can differ from the general population based on the sub-set population that you draw from.  This is known as a “selection bias.”  It is a very difficult bias to overcome, and statisticians can only alleviate the issue as best as possible.