In today’s Cox’s Conference, Dr. Karl Gordon Patti presented a case of chronic effusive pericarditis to expert discussant, Dr. Justin Grodin. Continue reading Cox’s Conference: Chronic Pericarditis
Today, Dr. Hunter Stone presented a case of atypical migraine presenting as left sided hemiparaesthesias, left-sided weakness and blurred vision of the left eye. Continue reading Parkland Morning Report: Hemiplegic Migraine
A patient with AIDS presents with acute dyspnea. An ultrasound probe is placed on his lateral chest wall producing the following image:
What is the name of the above finding and the associated diagnosis? What is likely the underlying cause of the above diagnosis?
Scroll down for the answer.
Frances Walocko, PGY1, presented her work at the Derm- Rheum Grand Rounds last week.
Here is Frances summarizing her work:
”The overall goal of this study is to identify patient risk factors and biologic markers that predispose patients with cutaneous lupus to develop systemic lupus erythematosus (SLE) through a longitudinal study of patients who start off with only cutaneous lupus. Our preliminary data have illustrated that patients initially presenting with generalized discoid lupus, a higher number of ACR criteria, and use of prednisone are at higher risk of progressing to SLE. The next step for this project is to further evaluate patient data for the follow up visits.”
Great job : )
Jeff Chidester, PGY1 and Cardiology Fellows (Bryan Lawlor, PGY-4, and Michael Loguidice, PGY-5) published an article last month in Circulation: Cardiovascular Quality and Outcomes titled, ”Multidisciplinary Clinic to Identify Near Misses and Decrease Readmission Rates After Hospitalization for Myocardial Infarction.”
For one week in October — Right Care Action Week — people around the country join together and take action to show what our health care system should look like.
The UTSWIM program is participating in Right Care Action Week, and will have ‘What Worries You?’ cards at each morning report site for you to fill out, on October 16th, Monday.
This is to initiate a conversation (with yourself, and your co-residents) about what cost issues face us today.
Josephine Harrington published an article in Circulation this week. She worked with primary mentor Dr. James de Lemos, to analyze the relationship of high-sensitivity troponin (hsTrp) with levels of sedentary activity.
Our last Journal Club on 10.2.17 was cardiology themed. Dr. Lanna Little presented the landmark MADIT-CRT trial, which showed that ICD+CRT significant reduction in HF events in patients with NYHA I/II heart failure with LVEF ≤30% and QRS duration ≥130 msec.
Thank you to our faculty discussants Dr. Khera at CUH, Dr. Das at Parkland, and Dr. Abdullah at the VA!
See below for discussions that occurred at all 3 sites:
Q: What are the important questions to ask ourselves when looking at a study?
- When you are evaluating a study like this, you should ask yourselves the following: 1) is the study valid? 2) Should I do it for my patient?
Q: What are we talking about when we evaluate the “validity” of a study?
- Internal validity (confounding, chance, bias)
- External validity
Q: Internal validity – What is “confounding?”
- Randomized controlled trials take care of the confounding variables. In an RCT, the only thing different between the 2 groups is the exposure, which you are assigning randomly. You are basically flipping a coin and that decides who gets what. So, confounding should not be an issue unless the groups are not balanced.
- Table 1 allows you to answer the question: “Do the groups look the same?”
Q: Internal validity – Chance: “Could they have found the same answer with chance alone?
- When evaluating this question, it is important to read the power statement in the paper.
- This study reports a power of “95% to detect a hazard ratio of 0.75 at a 2-sided significance level of 0.05.” – This means the study was powered to detect a 25% risk reduction and that you would find this by chance alone 5% of the time.
Q: Internal validity – Bias: What are the sources of bias?
- It is important to always read who paid for the study.
- Boston Scientific paid for this study.
- Open label design
Q: If industry sponsorship introduces bias, why have it?
- This and many trials are very expensive, so company sponsorship is a necessary evil
- Many NIH-sponsored trials are underpowered and can’t interpret because they didn’t have enough money, so industry sponsorship is incredibly common in cardiology.
Q: How do you keep industry’s bias from influencing your study?
- Create an explicit protocol on clinicaltrials.gov
- Once they pay, you keep them out of it, firewall the industry out of the study
Q: What is the significance of the difference in primary endpoint being driven primarily by HF events rather than mortality?
- You can’t cheat death as an endpoint (you’re either dead or not); however, hospitalizations for heart failure are less objective. Investigators and patients can both be biased. For example, if you know the patient is randomized to ICD without CRT, you might be more inclined to think their shortness of breath is due to heart failure.
- Mortality numbers in the study were expectantly low because of class I/II disease and every-changing medical management.
- Another thing to consider: The study investigators may not have been looking for improvement in mortality. Remember, that placement of ICD-CRT may have been associated with increased mortality. So, they could have been looking to see if the experimental group had increased mortality compared with the control.
Q: Why the open-label design?
- Had to be from a practical standpoint.
- This study would be a hard one to blind because you can see the wires, you can see the changes on EKG. Some studies do sham procedures (not done here).
Q: How did this study work to combat the bias introduced by the open-label design?
- They used an independent committee (3rd party), unrelated to the study to adjudicate the primary endpoints. This makes it better, but does not completely fix the issue. It is important to note that they only adjudicated those who showed up to the hospital.
Q: What is “intention to treat” and why is it important?
- Once randomized, always analyze as if there were in that original group no matter what happens to you.
- If you don’t use intention to treat analysis, then confounding creeps in and your balance between the groups is messed up.
Q: What is the problem with intention to treat analysis?
- The problem with intention to treat is cross over. When you cross over, you bias towards the null (decrease the effect size).
- Big studies should always account for cross over.
Q: Internal validity – What is generalizability?
- Generalizability asks: “who does this apply to?” You can think of generalizability as “Who was on table 1”
- The biggest downfall in looking at the generalizability of this study was that the study population was >90% white (so less generalizable to other races). There were more males in the study, but that’s the nature of the disease.
- Another important thing to look at in table 1 is the medications. You need to ask: “Were the study participants adequately treated?” In this study, the participants were pretty well treated with guideline-directed medical therapy.
Q: Should I apply the results to my patients?
- Important things to consider:
- Competing risks of the patient in front of you (i.e., Are they going to die of something else first such as malignancy, etc.?)
- Often when we talk about “Is it worth it” we are talking about number needed to treat. In this study, the NNT is ~12.5.
Q: What is the significance of the improvement in ECHO volumes measured?
- There are a couple sources of bias in the ECHO section of the endpoints. For example, what happened if people didn’t get echos that if people died? What if only sick people got echos?
- Also they did not tell us if the echo readers knew which patient got what.
- The way you’re supposed to do these is that you should dump them all together, shuffle them, and read them blind. If you put the echos side by side, then the echo reader may say “oh this could look a little better”
Q: Why is class I not in the guidelines?
- The NNT is likely way too high with minimal benefit given minimal hospitalization/symptoms already. You can always add it later if they progress to more severe symptoms
Q: Are the procedural risks changing?
- The procedural risks are getting better. We are now more conscious of infection risks and have more trained individuals because of more devices being placed
Dr. Katherine Gavinski, PGY2, has been selected to receive the L. Randol Barker Award for her abstract entitled “Best Evidence for Teaching Illness Scripts to Preclinical Students.”
She will be presenting her work at the 12th Annual Hopkins GIM Housestaff Research Awards National Competition in December this year.
She worked with Dr. Yvonne Covin (former GIM fellow at UTSW) and Dr. Palma Longo (faculty in the Health Care Sciences department).
A 22-year old man presented with progressive headache, painless unilateral vision loss, fever, and axillary lymphadenopathy. CT head and lumbar puncture were negative. RPR and Lyme titers were negative. Fundoscopic examination revealed the following:
What is this finding and the likely etiology in the context of the patient’s symptoms?
Scroll down for the answer. Continue reading Image Challenge of the Week!