The Timing Bias

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 HbA_1c  (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/. 

PHQ-9 as a Diagnostic Tool

Seeing as I will start my patient interviews this week, it seems fitting that I should first tie up any loose ends relating to my survey. As I mentioned in my previous blog post, "Nine Questions," I will be using the Patient Health Questionnaire-9 (PHQ-9). Unlike HAM-D which is usually administered after a patient is diagnosed with depression, PHQ-9 is often used to diagnose depression.

Earlier, I said that I would only use my survey to rank the severity of depression symptoms. But after finding that PHQ-9 is well-validated as a diagnostic tool for depression (as opposed to merely a depression scale), I figure that I may as well take full advantage of the PHQ-9 scores. So what exactly does that entail?

Figure 4. Severity scoring of the Patient Health Questionnaire-9 (PHQ-9).

Figure 4 shows the corresponding provisional diagnosis and recommended treatment for a given PHQ-9 score. DSM-IV defines criteria for two depressive disorders: dysthymia and major depressive disorder. Though there are many other depressive disorders that have been defined within the scientific community with their own set of criteria, for the purposes of my research, dysthymia and major depression are the only ones I am concerned with. Scores equal to or above 10 are indicative of some depressive disorder, or, in the least, depressive symptoms that should not go ignored. 

So...
I can use this survey to answer the question, "Is there a higher incidence of depression (a PHQ-9 score greater than or equal to 10) in diabetics than in the general population on average?" So what is the incidence of depression in the general population on average? The National Institute of Mental Health puts this number at

6.7%* for Major Depressive Disorder

and

1.5%* for Dysthymic Disorder.

*of the U.S. adult population

You might ask why I am not going to retrieve my own version of these numbers by surveying the non-diabetics here as well. Since I am working at a specialist's office (an endocrinology practice), almost all patients that are seen here were referred by a primary care physician who felt it was necessary due to the patient's specific case. In other words, just because I can survey patients here who aren't diabetic, this doesn't mean that the non-diabetic patients here are very representative of the general non-diabetic population. This is because they're being seen for other endocrinology-related complaints (e.g. thyroid disorders) that could — for all I know — affect their PHQ-9 scores. It wouldn't be a good idea to trust their scores because there could be many extraneous variables at work other than my main independent variable, the presence of diabetes mellitus.

So keep these two numbers in mind! They will remain in the "largest"-sized font on my blog for the rest of my project. At the end of my last sixty days of high school, I will hopefully be able to compare my results against these numbers from the NIHM.

Lalani.

Which Comes First?

"I was in shock. Most days, I couldn't take my insulin shot. I would just sit for hours. I tried to live what I thought was a 'normal' life. I did not take care of myself, and I did not control my diabetes….I felt hopeless, overwhelmed. I didn't know what to do, where to begin, where to turn. And I cried a lot."

Judy Abendschein was diagnosed with Type 1 diabetes at the age of eighteen. In the years that followed, Abendschein became suicidal and was eventually hospitalized. It seems clear that regardless of which came first — her diabetes or depression — each debilitating condition perpetually worsened the treatment of the other to the point of hospitalization.

For this reason, there are many theories that depression and diabetes interact in a circular manner. Poor control over one's diabetes (e.g. poor glycemic control) can be a risk factor for depression; likewise, depression symptoms (e.g. poor dietary habits) can result in poor diabetic control (see Figure 3).¹

Figure 3.  Diabetes and depression are often thought to interact in a circular manner.²

After researching various symptoms of diabetes and depression, I created the diagram above to depict how these symptoms may interact in a circular, i.e. causal, manner. Hopefully, by the end of these last sixty days of high school, I can attest to this theory — that each of these chronic conditions often impairs the proper treatment of the other. If this is indeed the case, I suppose the question of which condition comes first is irrelevant. With the presence of either one, prevention and detection of the other should be a priority; otherwise, a situation like Judy Abendschein's may arise.

Bottom line: diabetes and depression combine to form a potentially fatal mix.

Lalani.

¹American Diabetes Association. (2011). Living with diabetes. Retrieved from http://www.diabetes.org/living-with-diabetes/complications/mental-health/depression.html. 
²Burton, M. (1997). Depression and diabetes. (cover story). Countdown, 18(3), 20.

Nine Questions

My survey has been completed and approved! You can view it here or click on the tab entitled "Patient Survey." This means I will likely start conducting patient interviews within the week. Originally, I was planning on using the Hamilton Rating Scale for Depression (HAM-D) to survey patients. However, HAM-D is an extensive twenty-one question multiple-choice questionnaire.

My mentor felt that the more brief, but well-validated, Patient Health Questionnaire-9 (PHQ-9) is more appropriate for the scope of my research. As I indicated in my proposal, the scores from the survey would be used to rank the severity of depression symptoms; not to declare patients who score above a certain point as depressed and those who score below that point as not depressed. PHQ-9 scores each of the nine DSM-IV criteria for depression symptoms from “0,” i.e. "not at all," to “3,” i.e. "nearly every day" (see Figure 1).

Figure 1. The Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV) contains criteria for two major depressive illnesses: major depressive disorder (MDD) and dysthymic disorder (less severe than MDD).¹ PHQ-9 contains nine questions that address each of these depressive symptoms.

Figure 2. Hemoglobin A1c is a component of hemoglobin 
to which glucose is bound.

Earlier this week, I posted about how I ultimately want to show the implications of a relationship between diabetes and depression. You will notice on my survey that I am recording the patient's HbA_1c level, or glycosylated hemoglobin (see Figure 2). This is the average amount of sugar in blood over three months, i.e. the level of glycemic control (blood glucose control) that the patient has maintained over the previous three months. Because glycemic control is the most important aspect of maintaining control over your diabetes, the HbA_1c level is a very good indicator of how well the patient's treatment is going. It is for this reason that I have chosen to include the patient's HbA_1c on my survey. The recorded 
HbA_1c values will hopefully provide for a comparative analysis of how well the treatment is for diabetics with depression and diabetics without depression.

As I continue to update you on my progress, check out the “Glossary” tab at the top of the page where I have compiled a list of various terms I use. And as always, thank you for your readership.

Lalani.

¹ Mondimore, F. M. (2007, September 15). Never-Ending Winter: Chronic Depression. Psychiatric Times. Retrieved
  from http://www.psychiatrictimes.com/display/article/10168/54361.

So What?

"They say, 'you'd be depressed too if you were this sick.' People just assume that depression is a kind of natural reaction, a part of the disease."

Today was my first day at the Scottsdale Healthcare Endocrinology Institute. As my proposal indicates, the majority of my time here will consist of one-on-one patient interviews. But before I begin, I need to finalize the survey that I will use with my on-site mentor so that it is appropriate and comprehensive for the scope of my research - all of the data that I want to use must solely be attained from this survey (to be in accordance with 45 CFR 164.502(d), and 164.514(a) — (c) of the Health Insurance Portability and Accountability Act (HIPAA) Privacy Rule). That is, everything from age to type of diabetes must be on the questionnaire. Once it is completed and approved by my mentor, I will be sure to post a link to it.

Before I go any further, I want to address perhaps the most relevant question to my research that I feel I have overlooked. Ultimately, the purpose of my research is to attempt to evaluate the implications of the relationship between diabetes and depression, not just the relationship itself. In other words, assuming I gain a better understanding of this relationship over the next sixty days, so what?

Diabetes demands a very active life: checking blood pressure and blood sugar level as needed, scheduling regular appointments, making dietary changes, and incorporating exercise into your life are all vital activities that are required for proper treatment after a diagnosis of diabetes. The outcome of treating an infection is merely getting rid of it. But diabetes is not an infection. It's a chronic condition. And as of right now, the outcome for treating it is control — control of all of its hindering effects. According to my mentor, an endocrinologist, unless you treat a depressed diabetic for their depression, you won't get a person who is ready to treat their diabetes.

This brings me to the above quotation from Patrick Lustman, Ph.D., associate professor of psychiatry at Washington University in St. Louis, who reported in the June issue of Psychosomatic Medicine. Unfortunately, as Dr. Lustman indicates, symptoms of depression are often downplayed in diabetic patients. So as far contributing to the scientific community, my long-term goal for this project is to not only attest to the relationship between diabetes and depression, but also convey the importance of getting proper treatment for depression to effectively treat diabetes when both conditions are present.

Lalani.