Unfortunately, I will have to postpone the start of my patient surveys. I will be out of the country for a week and a half starting tomorrow. My mentor explained that starting patient interviews today would result in potentially discreditable data. This is because starting surveys at this time followed by an eight-day interruption period (of no patient surveys) would result in data that would be considered compromised by the timing bias, a type of intervention (or exposure) bias.
In clinical research studies like mine, data should be collected consecutively as much as possible. Otherwise, the mere fact that several subjects who fit my research sample's inclusion criteria are not being surveyed,
AND fit no apparent exclusion criteria (e.g. not having diabetes), raises many questions as to whether the researcher is favoring the data collected before and after this supposedly arbitrary period of exclusion (i.e. whether the researcher has a timing bias against those perfectly eligible candidates who could have been surveyed).
Considering that I'm already heavily relying on having a large sample size to account for various uncontrolled variables, I think it is important that I limit whatever potential bias(es) that could compromise the credibility of my results. For example, I am measuring the patient's HbA1c (see Glossary) as an indication for how well the patient's diabetic treatment is going. Surely, the amount of time that the patient has been undergoing endocrinological treatment influences how well the patient's diabetic treatment is going; however, it is not feasible for me to take this, or other similar variables into account — if I limited my sample's criteria to 1.) diabetic patients, 2.) with no other health complications, 3.) who have been undergoing treatment for their diabetes for only A to B years, 4.) who have no family history of depression, 5.) who have not undergone treatment for depression, 6.) who are not on medications X, Y, Z, etc., 7.) who are between the ages of N and M, I would have a minuscule sample size from which I could make no fair generalization about the 25.8 million1 who comprise the U.S. diabetic population.
For this reason, I will be waiting until the week of March 12 to begin my patient interviews in order to ensure maximum credibility of my data — a minor setback, sure, but a worthwhile one nevertheless. Thank you for reading.
Lalani.
1American Diabetes Association. (2011). Diabetes Statistics. Retrieved from http://www.diabetes.org/diabetes-basics/diabetes-statistics/.
In clinical research studies like mine, data should be collected consecutively as much as possible. Otherwise, the mere fact that several subjects who fit my research sample's inclusion criteria are not being surveyed,
AND fit no apparent exclusion criteria (e.g. not having diabetes), raises many questions as to whether the researcher is favoring the data collected before and after this supposedly arbitrary period of exclusion (i.e. whether the researcher has a timing bias against those perfectly eligible candidates who could have been surveyed).
Considering that I'm already heavily relying on having a large sample size to account for various uncontrolled variables, I think it is important that I limit whatever potential bias(es) that could compromise the credibility of my results. For example, I am measuring the patient's HbA1c (see Glossary) as an indication for how well the patient's diabetic treatment is going. Surely, the amount of time that the patient has been undergoing endocrinological treatment influences how well the patient's diabetic treatment is going; however, it is not feasible for me to take this, or other similar variables into account — if I limited my sample's criteria to 1.) diabetic patients, 2.) with no other health complications, 3.) who have been undergoing treatment for their diabetes for only A to B years, 4.) who have no family history of depression, 5.) who have not undergone treatment for depression, 6.) who are not on medications X, Y, Z, etc., 7.) who are between the ages of N and M, I would have a minuscule sample size from which I could make no fair generalization about the 25.8 million1 who comprise the U.S. diabetic population. Lalani.
1American Diabetes Association. (2011). Diabetes Statistics. Retrieved from http://www.diabetes.org/diabetes-basics/diabetes-statistics/.
No comments:
Post a Comment