Research Projects for non-Scientists

There is a tremendous interest in Evidence-based Design these days. I have had several people suggest to me that I conduct some research studying the impact of my art on patients. If I wanted to do do research, what would be involved regarding time, people and money? When I say research here I mean real research; high-value research that will stand up over time. I would want it to be research worthy of being  published in a peer-reviewed scientific journal.

To help begin to explore what needs to be considered for a research project I asked someone who does this for a living, Dr. Debra Harris. Debra is an Evidence-based Researcher for RAD Consulting. The following is an interview conducted by email:

Let's say I want to study the impact of black-and-white vs. color nature photographs in ICU patients diagnosed with a heart attack. I want to see if it alters their heart rate, blood pressure, duration in the ICU and amount of pain medication requested. I was thinking there could be three groups. One group with no art. The second group with a 30 x 40-inch color landscape and the third group with a 30 x 40-inch black-and white landscape.

First, is that a reasonable way to propose a research topic? Or should it be refined?

Yes, this is an excellent proposed study. Your research question is… Does natural photographic artwork influence the physical responses of ICU patients who have suffered a heart attack?

The dependent variables are heart rate, blood pressure, amount of pain medication used (not requested) and average length of stay. These data points are easy to collect from the patient chart. Statistical analysis will show if art is having a significant impact on the patient’s condition based on your four data sets.

What would I want to budget for a project like this if I wanted to make sure it was done properly?

This is always a loaded question. If you are hiring a research firm to conduct the study, then the cost is the expense of the research team, time and materials (equipment, expenses). If using an academic research team, the costs include the time of the research team (lead investigator, graduate students), tuition, indirect cost to the university which is an additional 25% - 53%, equipment, travel and other expenses.

If I were to guess and make certain assumptions like you were providing the artwork ready to install at no cost to the study, data collection took maximum 4 months and that the hospital was local to the research team, minimizing travel expenses, I would guess that this study could be conducted for about $40,000.

Interested in finding funding? It sounds like an interesting study!

How long would a project like this typically take from inception to publication?

The time to complete a project like this from research design to final research report would take from 8 months to a year. This would include research design, receiving approval from the IRB, collecting data, analyzing data and assimilating it into a meaningful report. This does not include publication. That is a different animal. A study of this nature could be completed in one year at an academic institution, but may require less time if a professional research firm was conducting the study. 

In order to get good evidence, how many patients would need to be studied? Should it include men and women?

Gender is a variable that should be determined while defining the research design. There may be reasons for limiting the study to one gender or the other, but for this study, I would include both men and women that suffered a recent heart attack and is recovering in the ICU.

Determining sample size requires a statistical power analysis. One cannot guess at the sample size. In order to determine sample size, one must know how big the difference needs to be to be meaningful for each variable, the confidence level (.95 is typical for peer review), the variability, and the effect size. Once you have that information, you can run the equation to determine sample size; then you need to add a percentage to account for invalid responses.

Since I am not affiliated with a University and since I have no idea how to do research properly, who do I need to hire? Where do I find them?

You could contact a university that has researchers interested in your topic which may lead to an academic research study utilizing available resources. As mentioned before, it may add time to the study plan. Another choice is to hire a professional research consulting firm like RAD Consultants that can allocate time and resources.

Are there any common road blocks I should anticipate? Are hospital ethics committees likely to veto the project?

Ethics committees at hospitals are called the Institutional Review Board (IRB). The study will have to go through an approval process with the IRB in order to collect data on your patient population.

Most medical IRBs require that the principal investigator complete an ethics course and test for certification.

The IRB process can take as little as three weeks or take months, depending on the data to be collected and the board’s satisfaction that all measures are taken to protect the health and privacy of subjects.

I do not think that an IRB board would veto this project as long is the risk to patients is very low and the research design has enough power to test the hypothesis.

Once the research is done, how does one go about submitting it for publication?

Publication to a peer review journal takes a significant amount of time. First, you have to write the article based on the methodology and the findings in a way that meets the criteria of the journal. Then, the article is submitted and may be rejected, accepted with provisional revisions or accepted with minimal changes. Once accepted, the article will be placed at the publisher’s discretion in the cue for publication. The process may take several months to a year or longer.

There was a previous post about Debra's work at RAD consulting on this blog. To see the article click here. 

Debra Harris, Ph.D., AAHID

Comments

[Bernard Ewigman, MD MSPH]

One of the critical design elements in such a study would be the random assignment to the exposure groups. Random assignment is the key strategy used in randomized clinical trials and other experimental designs to insure that the groups being compared are similar in every way except the exposure (the art). Without random assignment you may end up with sicker (or healthier) patients in one group or another; smaller more crowded ICUS in one group; different nurses and doctors; different disease processes or injuries (think of neuro ICU vs cardiac ICU vs surgical ICU). Any of these or other variables may effect the outcomes of the study, making it impossible to attribute the outcomes to the intervention. I think the value of the study and the liklihood of publication in a peer review journal would depend heavily on how the problem of random allocation was solved. I also wonder if ICUs are the best setting to test this hypotheses. There are so many factors effecting the course of events for ICU patients that are likely more powerful than viewing art. In contrast, patients on a general hospital ward or in a mental health facility may be more susceptible to any influence that art may have on the course of their illness.

[Upali Nanda]

I am in full agreement with what is written there. About the projected cost- this is definitely the kind of budget you would draft if applying for a grant- and what a research project could cost. Unfortunately hard dollars for research in our field is hard to come by. If no grant is available some of those costs get taken on by the hospital, or the researching agency- if the project outcomes are of interest. My point is that often research can get done for cheaper, as long as people are willing to put up some time and resources.

Also, with a variable like colored and b/w photography in landscape images, I am not sure that you would be able to get much significant difference through routinely collected physiological measures in a patient room. I would suggest something more controlled, and more high stress- like a pre-procedure room where physiological measures are collected at carefully monitored intervals.

Finally, yes we will frequently do research projects for others. Like I said, to keep research clean we try to keep it separate from just an analysis of our own design projects. So pretty much all the research we have done has been for someone else.

I don't work as an independent consultant, but if a research project was in the works that was pertinent to our research interests (such as the study you propose), then I would be more than happy to work as a collaborator.

One of the challenges Dr.Harris didnt mention is finding a "site". In my experience, the research design, availability of researchers, and the time issue, is all secondary to finding a hospital willing to be the site for the study and incorporating it within their crazy schedules.

[Debra Harris]

All good comments. My initial response to Henry's question was sort of a big picture look for people interested in knowing a bit more about what has to be considered when developing a research idea. Getting down to specifics for a research design would require much more time and thought, not to mention finding a site (or two or three or...), identifying variables, randomizing participants, and determining data that is or could be collected from patient charts and the phsyical environment.

It is an interesting proposition...if we could identify funding sources, I would be happy to coordinate a team to do a study such as this. I think funding has to be pursued creatively. I also think that we could find hospital sites without too much bother...