1. No audio. Recording preparation.
Health IT Workforce Curriculum Version 4.0

2. The Culture of Health Care
Evidence-Based Practice
Welcome to The Culture of Health Care: Evidence-Based Practice. This is Lecture e.
The component, The Culture of Health Care, addresses job expectations in health care settings. It discusses how care is organized within a practice setting, privacy laws, and professional and ethical issues encountered in the workplace.
Lecture e
This material (Comp 2 Unit 5) was developed by Oregon Health & Science University, funded by the Department of Health and Human Services, Office of the National Coordinator for Health Information Technology under Award Number IU24OC This material was updated in 2016 by Bellevue College under Award Number 90WT0002.
This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. To view a copy of this license, visit

3. Evidence-Based Practice Learning Objectives
Define the key tenets of evidence-based medicine (EBM) and its role in the culture of health care (Lectures a, b).
Construct answerable clinical questions and critically appraise evidence answering them (Lecture b).
Explain how EBM can be applied to intervention studies, including the phrasing of answerable questions, finding evidence to answer them, and applying them to given clinical situations (Lecture c).
Describe how EBM can be applied to key clinical questions of diagnosis, harm, and prognosis (Lectures d, e).
Discuss the benefits and limitations to summarizing evidence (Lecture f).
Describe how EBM is used in clinical settings through clinical practice guidelines and decision analysis (Lecture g).
The objectives for Evidence-Based Practice are to:
Define the key tenets of evidence-based medicine (EBM) and its role in the culture of health care.
Construct answerable clinical questions and critically appraise evidence answering them.
Explain how EBM can be applied to intervention studies, including the phrasing of answerable questions, finding evidence to answer them, and applying them to given clinical situations.
Describe how EBM can be applied to key clinical questions of diagnosis, harm, and prognosis.
Discuss the benefits and limitations to summarizing evidence.
Describe how EBM is used in clinical settings through clinical practice guidelines and decision analysis.

4. Using EBM to Assess Questions about Harm or Etiology
Question is not whether someone exposed to agent gets ill, but whether those with illness have higher rate or amount of exposure
Ideally assessed by randomized controlled trial, but RCT may be impractical or unethical
Next best evidence comes from observational studies, which have limitations
This lecture discusses the two remaining types of basic clinical questions: harm and prognosis.
We first discuss the use of evidence-based medicine to assess questions about harm or etiology. What causes a disease? Diseases may be caused by things in the air or water, bacteria, chemicals, radiation from the sun, or they may be caused by things that we do, such as a medical intervention that has a complication.
The primary issue with assessing harm or etiology [ee-tee-ol-uh-jee] is not whether someone who is exposed to some kind of agent gets ill but whether those who have that illness have had a higher rate or amount of exposure. For example, just because someone who uses a cell phone gets brain cancer does not necessarily mean that the cell phone causes brain cancer. In fact, let’s look at an example. Suppose we think that a certain chemical causes cancer. The ideal way to assess harm is to do a randomized controlled trial, so the best thing to do would be to get a group of people together and randomize half of them to getting exposed to the chemical and half of them not getting exposed. But obviously, exposing people to a chemical that might cause cancer is unethical, so we can’t do that. We need other kinds of study designs that enable us to detect whether something causes harm. We have to go down to the next best level of evidence, which is observational studies; however, we need to be careful in how we interpret observational studies to ensure that they answer our questions.

5. Examples of Questions to Answer about Harm
Do silicone breast implants cause autoimmune diseases, such as lupus? (Gabriel et al., 1994; Vase, 2013)
Women with silicone breast implants developed connective tissue diseases and arthritis but at no higher rate than those without them
Do anti-obesity drugs (e.g., fenfluramine and phentermine, also known as fen-phen) cause heart valve abnormalities? (Gardin et al., 2000; Andrejak, 2013)
Those who used these drugs developed certain heart valve abnormalities at a higher rate than those who did not
Here are some examples of questions that have come up in the news media about whether things cause harm. Back in the mid and early 1990s, there was the issue of whether silicone breast implants caused auto-immune diseases such as lupus and rheumatoid arthritis. It was definitely true that some women who had implants developed these diseases, which are commonly called connective tissue diseases; but they didn’t develop them at a higher rate than those who had no breast implants. So, silicone breast implants don’t appear to cause these auto-immune diseases despite the many lawyers who believe to the contrary.
Another question that came up in the 1990s was whether a number of anti-obesity drugs on the market at the time, such as fenfluramine [fen-floo r-uh-meen] and phentermine [fen-ter-meen] (also known as fen-phen), caused problems such as heart valve abnormalities. It turned out that those who used the drugs actually did develop certain heart abnormalities at a higher rate than those who didn’t use the drugs.

6. Hierarchy of Evidence for Harm
Randomized controlled trial
Cohort study
Case-control study
Case series/report
When we’re assessing studies of harm, we essentially have a hierarchy of evidence. A randomized controlled trial provides the best form of evidence, and often, we can use a randomized controlled trial to determine harm. The next best study is a cohort study, followed by a case-control study, and then a case series or an individual case report—the weakest form of evidence.

7. Evidence and Its Limits
Randomized controlled trial
Ideal, but often cannot be done or would be unethical to do so
Cohort study
Prospective study without randomization
Particularly useful when poor outcomes are rare and huge sample size would be required—e.g., upper gastrointestinal hemorrhage with NSAIDs
Problematic when groups are dissimilar—e.g., people who take NSAIDS may be sicker than or otherwise different from those who do not
Although a randomized controlled trial would provide the best evidence for harm, it often can’t be done, or it would be unethical to do so. Therefore, we have to use lesser forms of evidence. Another type of study is a cohort study. This is a prospective study, where we take a group of patients who get exposed to something, and we follow them forward in time, but we don’t randomize them. Cohort studies can be useful when poor outcomes are pretty rare such that a huge sample size would be required. For example, upper gastrointestinal hemorrhage that occurs with nonsteroidal anti-inflammatory drugs, or NSAIDs [en-sayds], such as Motrin and naproxen [nuh-prok-suhn], turns out to be relatively rare. We see it in clinical practice because so many patients take these drugs, but the rate of bleeding from NSAIDs is actually relatively low. In fact, it is difficult to do a clinical trial to detect that. So, a cohort study can help us in a situation where these poor outcomes are rare. Of course, cohort studies are problematic when the groups aren’t similar. For example, if we follow a cohort of people and we compare those who take the NSAID and those who don’t, they are obviously different people. Those who take the NSAID are more likely to have ongoing medical problems, particularly those who require these drugs, so we don’t get that benefit of randomization.

8. Evidence and Its Limits Continued
Case-control study
Most common form of observational study
Retrospectively identify cases of diseases and match to otherwise similar controls, looking to see if different rate or amount of exposure
Useful when condition is very rare or has long development time
Classic case was demonstration that DES causes vaginal cancer (reviewed in Swan, 2000)
A case-control study is the most common form of observational study, and it’s the most common type of study done to assess harm. There are times when we suspect that something is harmful and we want to find out as quickly as possible. We can’t do a prospective study because it might take years to get results, so we look retrospectively. We identify cases of the disease that we think are caused by the harmful agent, and then we match them to controls, thus the name, case-control study. Then we look at the two groups, those who have the disease and those who are the controls, and we see if there’s a different rate or amount of exposure. This type of study, for example, enabled us to see that there was not a higher rate of use of breast implants in patients with lupus and rheumatoid arthritis. Similar to cohort studies, case-control studies can be useful when the condition is rare. It can also be useful when the condition has a long development time. In fact, that’s how it was determined that the drug DES [D-E-S], which was taken off the market in the 1950s, causes vaginal cancer in women.

9. Evidence and Its Limits Continued 2
Case-control study – continued
Problem is when controls create spurious association, e.g.,
Coffee drinking associated with miscarriages
2016 NIH article says yes, Slate magazine article disputes findings
pancreatic cancer (MacMahon et al., 1981), but controls were patients with other GI diseases whose symptoms were exacerbated by coffee (so they drank less)
Differences were not present when other appropriate controls were used (Zheng et al., 1993)
One of the problems with case-control studies is that they can create spurious [spyoo r-ee-uh s] associations. A case in point was a study that purported to detect a linkage between coffee consumption and pancreatic cancer. This study was published in the New England Journal of Medicine, so it carried the prestige of that journal. It turns out that those who ran the study used the wrong control group, which shows you that sometimes even in the best of journals, bad science can get through. They used people who had other gastrointestinal diseases as the control participants. Many gastrointestinal diseases have symptoms that are exacerbated by coffee, particularly gastroesophageal [gas-troh-ih-soff-uh-jee-uhl] reflux. People with these other diseases drink less coffee. So, when you use them as a control group of cases versus those with pancreatic cancer, the people with pancreatic cancer drink more coffee, but the coffee is not the cause of their cancer. They drink more coffee because they can tolerate it better. In fact, when this study was repeated with an appropriate control group over a decade later, that association was disproven, and those of us who drink coffee can rest assured that we aren’t causing pancreatic cancer in ourselves.

10. Evidence and Its Limits Continued 3
Case series/report
No comparison group
Famous example was Bendectin for nausea in pregnancy, where adverse publicity led to removal from market of safe and effective treatment
Was combination of two agents, both of which were effective and neither of which were harmful (Magee, Mazzotta, & Koren, 2002; Hale, 2012)
Re-introduced use of ingredients in a delayed-release combination pill after US FDA approval in 2013 (Nuangchamnong, 2014)
The lowest form of evidence is the case series, or even worse, the single case report. The reason a case series provides the lowest form of evidence is that there’s no comparison group. A case series can be used to generate a hypothesis that could then be tested with more rigorous study designs, but a case series or case report, in and of itself, isn’t considered to provide high-quality evidence. A classic example of poor evidence resulting from a case series is the drug Bendectin [ben-dek-tin] that came out in the 1980s as treatment for nausea in pregnancy. We should always be careful about giving any medication to pregnant women, but it turns out that Bendectin was unfairly singled out. The adverse publicity was so strong that it was taken off the market. Some women actually benefitted from the drug, because one of the ingredients in Bendectin was a vitamin. Bendectin was a combination of two agents, both of which were known to be effective and neither of which had any harmful effects, but because there was a case series of birth defects, Bendectin was taken off the market, withdrawing an effective treatment for women with nausea from pregnancy.
In 2013, the US Food and Drug Administration (FDA) approved the two ingredients – doxylamine and pyridoxine – and they are now offered in a delayed-release pill as a first-line treatment for the management of nausea and vomiting in pregnancy.

11. “Pure” Prognosis Studies Are Rare
Prognosis is “natural history” of disease
But very little “history” is “natural” in modern era with our abundance of diagnostic tests, interventions, harmful agents, and so on
Many studies measure prognosis after a test or intervention
Let’s make a few comments about prognosis [prog-noh-sis]. One of the important things to know about prognosis is that it’s actually hard to do a good prognosis study because prognosis is the natural history of a disease—how the disease actually unfolds. In this day and age, there’s very little in the history of a disease that is natural. In almost every disease, we intervene in some way, whether through diagnostic tests or treatments. Many studies end up measuring prognosis after some form of test or intervention. We may measure prognosis for a group of patients who have been diagnosed with a certain condition, or we may measure prognosis for a group of individuals who have been given a certain treatment for a particular condition. When we start to do that, the differences between a prognosis study and a treatment study start to blur a little bit.

12. Prognosis Usually Measured by a Survival Curve
The most common way that we measure prognosis is with a survival curve. This slide shows a survival curve from one type of cancer when it is diagnosed at different stages. Like many cancers, the later the stage of the cancer, the worse the prognosis. Patients who have stage one of this particular cancer have a good prognosis, but those who have stage four are likely to die within seven years.
5.5 Chart: Survival Curve (Adapted from Dunn, 2002)

13. Example Studies of Prognosis
Extremely preterm birth (Marlow, Wolke, & Bracewell, 2005)
Followed cohort of 241 children from UK and Ireland born at 25 or fewer weeks gestation
Compared with 160 classmates born at full term
41% of preterm children had “serious impairment” on cognitive assessment compared with 1.3% in control group
Untreated early, localized prostate cancer (Johansson et al., 2004; also see Yao, 2010; Gulati,. 2011)
223 men followed from 1977 to 1984
17% developed generalized disease
16% died of disease
Here are some example studies. Several prognosis studies have looked at children who were extremely preterm [pree-turm], had extremely low birth weight, and were in neonatal intensive care units. There are a number of ethical and philosophical issues about how much intervention is desirable for these extremely small children, and a prognosis study gives us an indication of how they do, especially in this case, six years out. A 2005 study followed a cohort of 241 children from the United Kingdom and Ireland who were born at twenty-five or fewer weeks’ gestation—extremely preterm. They were compared with classmates in their schools who were born at full term. The differences were substantial; forty-one percent of the preterm children had serious impairment on a variety of cognitive tests compared with one-point-three percent of their full-term classmates. Another study of prognosis was published on prostate cancer, specifically untreated, early, localized prostate cancer. This study focused on a group of men who were diagnosed between 1977 and 1984 and had regular long-term follow-up. Today, many of those men would have had surgery, but this study gives us an indication of the natural history of the disease. About seventeen percent of these men developed generalized disease—in other words, the cancer spread, or metastasized [muh-tas-tuh-sahyzd], to other areas, and sixteen percent of the men died of the disease. This is another demonstration that many men who develop prostate cancer do not die from it, and it does not even spread outside their prostate.

14. Evidence-Based Practice Summary – Lecture e
Questions about harm assess whether exposure to some natural or manmade agent causes disease and is usually answered with a case-control study
Questions about prognosis tell us the natural course of disease
This concludes Lecture e of Evidence-Based Practice. In summary, questions about harm assess whether exposure to some natural or manmade agent causes disease, and these questions are usually answered with a case-control study. Questions about prognosis tell us the natural course of disease.

15. Evidence-Based Practice References – Lecture e
Al-Dakkak, I. (2011). Tea, coffee and oral cancer risk. Evidence-Based Dentistry, 12, 23–24. Retrieved from
Alsop, J., Scott, M., & Archey, W. (2016). The mixed randomized trial: combining randomized, pragmatic and observational clinical trial designs. Journal of Comparative Effectiveness Research, 5(6), Retrieved from
American College of Obstetrics and Gynecologists. (2010). Moderate caffeine consumption during pregnancy. (Committee Opinion #462), Obstetrics & Gynecology :
Andrejak, M., & Tribouilloy, C. (2013). Drug-induced valvular heart disease: an update. Archives of Cardiovascular Diseases, 106(5),
Anonymous. (2015). Beware spurious correlations. Harvard Business Review. June. Retrieved from
Faraoni, D., & Schaefer, S. T. (2016). Randomized controlled trials vs. observational studies: why not just live together?. BMC anesthesiology, 16(1), Retrieved from
Ferreira, J. C., & Patino, C. M. (2016). Choosing wisely between randomized controlled trials and observational designs in studies about interventions. Jornal Brasileiro de Pneumologia, 42(3), Retrieved from
Gulati, R. et als. (2011). What if I don’t treat my PSA-detected prostate cancer? Answers from three natural history models. Cancer Epidemiology Biomarkers Prevention. 20(5): Retrieved from
No audio.

16. Evidence-Based Practice References – Lecture e Continued
Gabriel, S., O’Fallon, W., Kurland, L., Beard, C., Woods, J., & Melton, L. (1994). Risk of connective-tissue diseases and other disorders after breast implantation. New England Journal of Medicine, 330, 1697–1702.
Gardin, J., Schumacher, D., Constantine, G., Davis, K., Leung, C., & Reid, C. (2000). Valvular abnormalities and cardiovascular status following exposure to dexfenfluramine or phentermine/fenfluramine. Journal of the American Medical Association, 283, 1703–1709.
Hale, R. W., & Niebyl, J. (2012). Bendectin: how a safe and effective drug was removed from the market by our legal system. ACOG Clinical Review, 17(5),
Johansson, J., Andren, O., Andersson, S., Dickman, P., Holmberg, L., Magnuson, A., & Adami, H. (2004). Natural history of early, localized prostate cancer. Journal of the American Medical Association, 291, 2713–2719.
MacMahon, B., Yen, S., Trichopoulos, D., Warren, K., & Nardi, G. (1981). Coffee and cancer of the pancreas. New England Journal of Medicine, 304, 630–633.
Magee, L., Mazzotta, P., & Koren, G. (2002). Evidence-based view of safety and effectiveness of pharmacologic therapy for nausea and vomiting of pregnancy (NVP). American Journal of Obstetrics and Gynecology, 186, S256–S261.
Marlow, N., Wolke, D., & Bracewell, M. (2005). Neurologic and developmental disability at six years of age after extremely preterm birth. New England Journal of Medicine, 352, 9–19.
Nuangchamnong, N., Niebyl, J. (2014). Doxylamine succinate–pyridoxine hydrochloride (Diclegis) for the management of nausea and vomiting in pregnancy: an overview. International Journal of Women’s Health, 6: 401–409.
No audio.
Health IT Workforce Curriculum Version 4.0

17. Evidence-Based Practice References – Lecture e Continued 2
Swan, S. (2000). Intrauterine exposure to diethylstilbestrol: Long-term effects in humans. Acta Pathologica, Microbiologica et Immunologica Scandinavica, 108, 793–804.
Vase, M. Ø., Friis, S., Bautz, A., Bendix, K., Sørensen, H. T., & d'Amore, F. (2013). Breast implants and anaplastic large-cell lymphoma: a Danish population-based cohort study. Cancer Epidemiology Biomarkers & Prevention, 22(11), Retrieved from
Weng, Xiaoping, Ph.D., Odoili, R. MSPH, Li De-Kun, MD, PhD. Kaiser Permanente Division of Research. (2008). Maternal caffeine consumption during pregnancy and the risk of miscarriage: a prospective cohort study. American Journal of Obstetrics and Gynecology, 198 (3), 279.e1-279.e8. Retrieved from
Yadav, D., & Lowenfels, A. B. (2013). The epidemiology of pancreatitis and pancreatic cancer. Gastroenterology, 144(6), Retrieved from
Yao, S. & Lu-You, G. (2010). Diagnosis of localized, screen-detected, prostate cancer—Crisis or opportunity? Journal of the National Cancer Institute. 102(13): 919–920. Retrieved from
Zheng, W., et al. (1993). A cohort study of smoking, alcohol consumption, and dietary factors for pancreatic cancer (United States). Cancer Causes & Control, 4, 477–482.
No audio.
Health IT Workforce Curriculum Version 4.0

18. Evidence-Based Practice References – Lecture e Continued 3
Charts, Tables, Figures
5.5 Chart: Survival Curve. Adapted from Dunn, S. (2002). Survival curves: The Basics. CancerGuide. Retrieved from
No audio.
Health IT Workforce Curriculum Version 4.0

19. The Culture of Health Care Evidence-Based Practice Lecture e
This material was developed by Oregon Health & Science University, funded by the Department of Health and Human Services, Office of the National Coordinator for Health Information Technology under Award Number IU24OC This material was updated in 2016 by Bellevue College under Award Number 90WT0002.
No audio.
Health IT Workforce Curriculum Version 4.0