Research forum - Pediatric Oncall

Posted By :

Posted On :

30 Sep 2010

Blogs:- A reader has asked whether we have to know the prevalence of a disease before we calculate sample size. Here is what I think the reader is asking. What do we do if we do not have adequate number of patients to study_? That is a question which will automatically come once you calculate the sample size. If your clinic does not follow adequate number of patients with that particular problem, you have to recruit patients from other centers. This is why research on less common diseases require multi-center studies.
The situation is slightly different in the hypothetical study we discussed in the earlier essay. The investigator is trying to find out whether the incidence of paradoxical reaction {to treatment for TB} is higher in the group with TB and HIV as opposed to a group with TB only. We assume that he, she has adequate number of patients with both TB and HIV. If she has a large group of patients it is not practical or necessary to study the entire group. The question is to decide...

Total Replies:-

Section on Research articles

Under Category :

Posted By :

Posted On :

18 Aug 2010

Blogs:- Greetings to the readers. Please note the presence of a new section on Research Articles in this Research Forum. You will see my response to a specific question on calculating sample size and also an article on the history of medical research in this new section.

Looking forward to your responses.

Thank you Balu Athreya

Total Replies:-

Sample Size

Under Category :

Posted By :

Ira Shah

Posted On :

08 Jul 2010

Blogs:- Thank you Dr Athreya for your comments. Yes, we are looking to see incidence and type of paradoxical reactions in children who are being treated for TB. This includes both HIV and non HIV infected children. Paradoxical reaction means worsening clinical or radiological symptoms inspite of response to anti TB therapy. This is similar to a condition called as IRIS in HIV infected individuals. This is known as immune reconstitution. Paradoxical reactions are seen in form of increase in size of lymphnodes or appearance of tuberculomas on treatment inspite of patient having response to therapy.

So now that we have defined the study purpose, how should be calculate sample size.

Total Replies:-

Calculating Sample Size

Under Category :

Posted By :

Posted On :

28 Jun 2010

Blogs:- Thank you Dr.Mehta for your response and starting the discussion. Yes, we need to know more about the study details.
Here is my thinking. The way the question is framed, the author wants to study paradoxical reaction {PR} in TB in children. The first step is to define paradoxical reaction. I think the author means the incidence of worsening of TB after starting appropriate treatment. My first response is that this data is already available in the literature. One number I have seen is 15 percent.
If the author does not want published data, but wants to establish the incidence of PR in her population of patients, the study is a simple one. She has to establish the criteria for the diagnosis of TB, define what subtypes of TB she wants to get data on {eg: primary TB, disseminated, pulmonary, any kind of TB}, define paradoxical reaction {including time interval from the onset of treatment, drug-resistance etc} and exclusion criteria. Then she has to prospectively follow the n...

Total Replies:-

Calculating Sample Size

Under Category :

Posted By :

Posted On :

14 Jun 2010

Blogs:- Dear Reader, I was waiting for someone to start a discussion on the problem posed by Dr.Shah. No one has responded. Therefore, let me start the discussion.

Dr.Shah wishes to study paradoxical reaction in Tb in children. She wants to know how to calculate sample size. What do you think the correct answer is_? Alternately, how can this question be refined, if the real question is about sample size calculation_?

Looking forward to a discussion group.

Total Replies:-