Category Archives: Research design

Evidence based nursing, RCT, reading research, Research design, Uncategorized

Telling the Future: The Research Hypothesis

Leave a comment

What is a research hypothesis?   A research hypothesis is a predicted answer; an educated guess.  It is a statement of the outcome that a researcher expects to find in an experimental study.Hypothesis

Why care?  Because it tells you precisely the problem that the research study is about!  Either the researcher’s prediction turns out to be true (supported by data) or not!
A hypothesis includes 3 key elements: 1) the population of interest, 2) the experimental treatment, & 3) the outcome expected.  It is a statement of cause and effect. The experimental treatment that the researcher manipulates is called the independent or cause variable.  The result of the study is an outcome that is called the dependent variable because it depends on the independent/cause variable.

For example, let’s take the hypothesis “Heart failure patients who receive exmeds2perimental drug X will have better cardiac function than will heart failure patients who receive standard drug Y.”  You can see that the researcher is manipulating the drug (independent variable) that patients will receive.  And patient cardiac outcomes are expected to vary—in fact cardiac function is expected to be better—for patients who receive the experimental drug X.

Ideally that researcher will randomly assign subjects to an experimental group that receives drug X and a control group that receives standard therapy drug Y.   Outcome cardiac function data will be collected and analyzed to see if the researcher’s predicted answer (AKA hypothesis) is true.

In a research article, the hypothesis is usually stated right at the end of the introduction or background section.

If you see a hypothesis, how can you tell what is the independent/cause variable and the dependent/effect/outcome variable?question   1st – Identify the population in the hypothesis—the population does not vary (& so, it is not a variable).   2nd – Identify the independent variable–This will be the one that is the cause & it will vary.  3rd – Identify the dependent variable–This will be the one that is the outcome & its variation depends on changes/variation in the independent variable.

PRACTICE:  What are the population, independent variable(s) & dependent variable(s) in these actual research study titles that reflect the research hypotheses:

FOR MORE INFORMATION:  See SlideShare by Domocmat (n.d.) Formulating hypothesis at http://www.slideshare.net/kharr/formulating-hypothesis-cld-handout

 

Evidence based nursing, reading research, Research design

“What was the question, again?” Hypothesis vs. Research question

1 Comment

imagesCAGYW6WBWhat is the difference between a hypothesis and a research question?  I suppose some will ask: “Why should I care?”  

The answer is that knowing the difference is a clue to how strong the research findings are.  Experimental studies that use hypothesis create stronger evidence for practice, than do non-experimental studies that merely answer research questions.

  • A hypothesis is a predicted answer and focuses on cause-and-effect , such as “Those who take 80mg ASA every day will have lower incidence of MI’s than those who do not.”  When we know a lot about a topic already, then we can hypothesize (in other words make an educated guess about how the experiment will turn out).  A researcher may word the hypothesis in interrogative form with a question mark, but you can still recognize it as a hypothesis if it uses any terms like cause, effect, impact, increase, improve, or their synonyms and has at least 2 variables.
  • Researchers ask only a research question when they don’t know enough to guess about the cause and effect, & so they either 1) want merely to describe something or 2) to figure out whether 2 things are related to each other, but aren’t ready to identify one as causing the other. For example no one knows which came 1st, the chicken or the egg?—you can see that chickens and eggs are related 100% of the time, but you may win a Nobel prize if you can figure out which one originally caused the other.

Cause and effect (experimental, hypothesis) studies create stronger evidence for practice than do descriptive or correlational (non-experimental, question) studies.

Critical thinking question: One of the following is a hypothesis & one is a research question.  Which is wquestionhich and why?

  1.  The purpose of this study was to describe the expectations for pain relief of patients with abdominal pain and how their communication with providers relates to their overall pain relief. (Yee et al 2006)
  2.  We investigated whether a brief pain communication/education strategy would improve patient pain communication skills. (Smith et al, 2010)
Evidence based nursing, Literature review, Research design, Sampling

“That is so random!” But is it Representative?

Leave a comment

What makes a good sample in research?  One thing.  And it isn’t random selection.  (Surprised?)

Portrait of a diversity Mixed Age and Multi-generation Family embracing and standing together. Isolated on white background. [url=http://www.istockphoto.com/search/lightbox/9786738][img]http://dl.dropbox.com/u/40117171/group.jpg[/img][/url]It is representativeness.  No matter how the sample was picked, it must be representative of all those in the larger population, if the researcher wants to say anything about anyone who wasn’t in the study.  Now, of course, it is true that random selection is more likely to give you a representative sample, but it is no guarantee.  Only likely.

What is random sampling?  It is when every member of the larger population has an equal chance of being selected for the study sample.  Example? Drawing names out of a hat.  It is well-accepted practice to generalize research results from a random sample to others like those being studied (assuming that all other aspects of the study are strong).

In contrast a convenience (or nonprobability) sample is when some people are more likely to be chosen to be in the study than others.  You shouldn’t generalize the results of these studies because the samples may Not represent others.

Example of when random sampling doesn’t work: Let’s say you have a mixture of red, green, & yellow apples, and you select a sample that has only yellow apples.  (The red & green ones are going to be offended!–They’re left out.)  You now have a sample that is biased in favor of yellow apples!   Your sample does Not represent the larger population of apples…even if you used random methods to get it.  If you want to apply the study to red & green & yellow apples…well….you must get some of them in your sample, too. The yellow apples might not be at all like the other types and studying just yellow might mislead you into thinking something about the red & green ones that isn’t true!   Of course you could study all the millions of apples in the world and exclude none, but that would be pretty cumbersome and expensive.   So, it’s better to go for a representative sample!

When else doesn’t random sampling create a representative sample?   If I am doing historical research, say on the Nursing Department at California State University/Northridge, then I want to hand pick the specific RNs by name who were in charge of the Department from the beginning.  Randomly selecting nurses from those who worked at the University won’t represent those leaders.

QUESTIONCritical Thinking:  Take a quick look at the linked abstracts. How were the samples selected?  How representative are the samples of a larger population of interest?  Could you generalize the results to other people, and if so to whom?

Want more information on sampline? Check this out.  It takes < 5 minutes:

https://www.youtube.com/watch?feature=endscreen&v=be9e-Q-jC-0&NR=1

Evidence based nursing, reading research, Research design

“What it is.” – a primer on descriptive studies

2 Comments

What is a “single descriptive or qualitative study”?

A single descriptive or qualitative study is a study in which the researcher watches and listens, then describes what s/he sees and hears. In these studies the researcher does NOT try out a new treatment and measure the results. The descriptive researcher only describes.

The description may be reported in:

  • Numbers & Statistics (called a quantitative study) OR
  • Words & Themes (called a qualitative study).

If the researcher reports BOTH numbers/statistics AND word/themes, it is called a mixed methods study.

Descriptive studies are listed as pretty weak evidence for changing practice, but remember that it is weak only in terms of Not being able to show that one event is causing another event.   They are still excellent in terms of describing what is.    (For more on strength of evidence refer back to: “I like my coffee—and my evidence—strong!)

CRITICAL THINKING: (sing with me) “One of theseQUESTION things is not like the other. One of these things just doesn’t belong. Can you guess which one is not like the other?” Two are descriptive studies. One is not.

  1. Thomas, D., et al., (2015). Pediatric Pain Management in the Emergency Department: The Triage Nurses’ Perspective. The aims of the study were to describe the triage pain treatment protocols used, knowledge of pain management modalities, and barriers and attitudes towards implementation of pain treatment protocols.
  2. Ucuzal & Dogan. (2015). Emergency nurses’ knowledge, attitude and clinical decision making skills about pain. The aim of this study was to examine emergency nurses’ knowledge, attitude and clinical decision-making skills about pain.
  3. Harrison et al., (2015). Sweet tasting solutions for reduction of needle-related procedural pain in children aged one to 16 years.  The aim of the study was to determine the efficacy of sweet tasting solutions or substances for reducing needle-related procedural pain in children beyond one year of age.
Evidence based nursing, Literature review, reading research, Research design

Cohort & Case-controlled studies: Going forward & backward

2 Comments

Got a clinical problem?  You probably want to solve it with evidence—STRONG evidence.   Click on this link to see one well-accepted hierarchy from strongest #1 to weakest #7 (Melnyk & Fineout-Overholt, 2005).   Today let’s look at the 4th strongest level of evidence = Case controlled or cohort studies

First a quick review

Click here for a quick review of the strongest 2 levels of evidence (#1 Systematic reviews, Meta-analyses, or Evidence-based clinical practice guidelines based on systematic review of RCTs. #2 Randomized controlled trials)

Click here for a review of the 3rd strongest type of evidence (#3Controlled trials without randomization)

Now on to the new “stuff”  strong

All 3 of the top, strongest levels of evidence are experimental studies (or include available experimental studies). That means the researcher actually does something or gives a treatment to some of the subjects and then records the outcomes. 

The weaker 4 levels of evidence are non-experimental designs. This means that the researcher merely observes & does Not do anything to subjects. So how does that work?!

First, a cohort study (non-experimental). A cohort study starts with a group of people who have something in common and then the researcher observes only & keeps collecting data from them over a long time into the future. Data collection into the future is called a prospective study.  “An example is the Nurses’ Health Study, in which over 20,000 nurses were identified and followed-up annually with tests and surveys for over 25 years (this study is still ongoing). These studies provide very valuable information, but are obviously very expensive and time-consuming.”(OMERAD EBM course, 2008)

Now a case-controlled study (non-experimental).  In a case controlled study the researcher observes only & collects data over time into the past (not the future). Data collection into the past is called a imagesCAH6C8NTretrospective study. Again, from the OMERAD EBM (2008) site this example: “Patients with a disease are identified who have suffered a bad outcome such as death or recurrence, and compared with patients who have the disease but haven’t suffered the bad outcome. For example, a researcher might  identify a group of breast cancer patients who have died…, and compare them with a similar group of patients with breast cancer who are still living.”

Critical thinking: Which of these would be better for casQUESTIONe-controlled study and which for cohort study.

  1. You are a runner in the Los Angeles marathon and you are interested in how that race can improve cardiovascular health among those who finish. Question: Cohort or Case controlled?
  2. Some finishers of the LA marathon die of heart attacks 20 years later; many survive another 40 years.   Question: Cohort or Case controlled?

For more info see:

 

Evidence based nursing, Literature review, nonrandom trials, Research design

Of Mice and Cheese: Research with Non-equivalent Groups

3 Comments

Last week’s blog focused on the strongest types of evidence that you might find when trying to solve a clinical problem. These are: #1 Systematic reviews, Meta-analyses, or Evidence-based clinical practice guidelines based on systematic review of RCTs; & #2 Randomized controlled trials. (For levels of evidence from strongest to weakest, see blog “I like my coffee (and my evidence) strong!”)

So after the two strongest levels of evidence what is the next strongest? #3 level is controlled trials without randomization. (Sometimes called quasi-experimental studies.)

Here’s an example of a controlled trial without randomization: I take two groups of mice and test two types of cheese to find out which one mice like best. I do NOT randomly assign the mice to groups. The experimental group #1 loved Swiss cheese, & the control group #2 refused to eat the cheddar. I assume confidently that mice LOVE Swiss cheese & do NOT like cheddar. What’s the problem with my conclusion? If you want to know, then read on!swiss cheese

In my mouse Controlled Trial Without Randomization, the groups were formed by convenience and Not randomly assigned. Thus, any difference in outcomes between groups might be related to some pre-existing difference between groups. My outcome of mice loving Swiss & hating Cheddar might have nothing to do with the experimental treatment.   In fact, I did not know that all my mice in the Swiss cheese group #1 hadn’t eaten in 2 days, and my mice in the cheddar group #2 had just had a full lunch. Ooops.

On the other hand if I had randomly assigned all the mice to two groups, then I could be relatively confident that all little differences between group members were evenly distributed to both groups, so that the groups were equivalent. My two mouse-groups would have probably ended up with a pretty even distribution of both hungry and not-so-hungry mice.   Then if my Swiss cheese group devoured the Swiss and my cheddar group rejected the cheddar, I could be more certain that mice love Swiss and dislike cheddar.

Happy evidence hunting!

Research design

Ebola or Other Outbreak: When We Can’t or Shouldn’t Experiment

Leave a comment

 What do we do to study the cause of disease when we cannot or should not expose people to disease risk (i.e., manipulate the independent variable). For example, while we want to understand Ebola transmission and outcomes, legally and ethically we cannot & should not expose people to Ebola risk factors.   We cannot do Ebola experiments on people.

Thus, we have to observe what happens when nature takes its course. One common research design in which we let disease/nature take its course is a case-control study. What is a case-control study?

Here’s a quick explanation.   The researcher looks for people who have (or had) the disease and then looks back in time at their history of exposure to risk factors for the disease. Those who have been exposed and who did not (or not yet) get the disease are the control subjects. If risk factors for the disease are not well-known then it may be difficult to find control subjects because we would have a hard time telling who was exposed.

Case-control and other studies in which we look back at what happened in the past are called retrospective studies. (In contrast, most nursing studies are prospective studies—in other words they start at the present and move forward. For example, if we were doing research on Ebola symptom management, we would try out symptom management strategies on persons with Ebola and measure into the future how well those strategies work.)

A great flow diagram and clear explanation of case control studies is at http://www.ciphi.ca/hamilton/Content/content/resources/explore/fb_case_v_cohort.html . Check it out!

Critical thinking practice: If you were to design a case-control study related to information in the excerpt below, answer these questions:

  • Who would be the case subjects?  
  • Who would be the control subjects?
  • What are the risk factors?
  • Why would the study be retrospective?

“Ebola virus, a member of the Filoviridae group, is transmitted by direct contact with blood, secretions, or contaminated objects and is associated with high case-fatality rates (28). Investigations of outbreaks in Africa suggest that Ebola infection may be more severe during pregnancy and that mortality rates are higher. Pregnant women infected with Ebola more often have serious complications, such as hemorrhagic and neurologic sequelae, than do nonpregnant patients (31). Unlike risk for death from Lassa fever, which is highest during the third trimester of pregnancy, risk for death from Ebola is similar during all trimesters (33).” (Jamieson et al, 2006, http://wwwnc.cdc.gov/eid/article/12/11/06-0152_article)