epidemiology 2001 winter 2016–readings exercisesfaculty.georgebrown.ca/~tgula/epidemiology/week1...

Epidemiology 2001 Winter 2016–Readings & Exercises

Reading: John Snow pg. 2

ppt. Introduction to epidemiology pg. 5

ppt. Classifying research epidemiologically ‐ basics pg. 9

Notes on calculating Disease Frequency pg. 11

ppt . Disease frequency concepts and calculations pg. 12

ppt. Comparing disease frequencies concepts and calculations pg. 16

Notes on calculating RR, RD and odds ratio in epidemiology pg. 18

Exercise 1: practice with population based disease frequency pg. 19

Exercise 2: practice with comparing disease frequency pg. 20



Exercise 5abcd analytic research types practice scenarios pg. 23

ppt focus on experiment pg. 27

ppt cohort and case control pg. 30

Exercise 6 identifying research types practice with abstracts pg. 35

Exercise 7abcd: practice with cohort vs case control pg. 42

Template: reading Research pg. 46

Notes on validity pg. 47

ppt validity pg. 49

Exercise 8 introducing adjusting calculations for rates pg. 54

Exercise 9 : stratification exercise pg. 55

Exercise 10ab: adjusting for age exercise (optional) pg. 56

ppt on causation pg. 58

Exercise 11: assessing causation pg. 61

ppt screening pg. 62

Exercise 12‐screening practice pg. 65

Exercise 13 practice with findings from data pg. 66

Appendix 1: The great debate outline pg. 67

© Taras Gula 2015

2

Epid 2001 – introductory reading:

Source: BBC Online, 2001. John Snow 1813 - 1858 Soho was the site of a terrible and devastating outbreak of disease that killed 600 people within a quarter of a mile in the course of a few days in 1854. But thanks to a local doctor, this was one of the last great Cholera epidemics. His name was John Snow.

Outbreak Cholera is a horrible disease. Following infection, the patient at first doesn't feel particularly ill, and the first sign is diarrhoea. But this soon gets out of control and so much fluid is lost that the blood appears thick. Within a short time - often only two or three days - about half the patients will die, mainly of dehydration. It was assumed that cholera was airborne, but Snow was sure this wasn't right. As a doctor he attended many patients without getting Cholera himself. Second, he argued that the infection always seemed to affect the gut before the patient felt generally ill, and this suggested that it was ingested. He published The Mode of Communication of Cholera in 1849 - but many refused to abandon the 'miasma' (bad air) theory.

Towards the end of August 1854, he got the chance to prove his ideas in the most dramatic circumstances. During late August 1854, there were a few Cholera deaths in Soho. But during the night of August 31st and September 1st there was what Snow called "a violent increase in the malady". Fifty-six new cases were reported that night. The next day there were 143 new cases, and on 2nd September, 116. And the deaths followed swiftly: 70 on September 1st, 127 on the 2nd.

As soon as he heard how awful the outbreak was, Snow determined to investigate it. To account for such a swift and violent epidemic, he was sure the water must be contaminated, and his suspicion immediately fell on the popular pump that stood at the junction of Broad Street and Cambridge Street. He examined it on the 3rd of September, but found only minimal visible contamination. This wasn't enough evidence. He went to the Register of Deaths and got details of all the deaths from cholera in the Golden Square, Berwick St and St Anne's, Soho, districts that week. Armed with the places where people had died, Snow returned to the streets to find out what had really happened. The most obvious thing was that most of the deaths were close to the pump. (see map on next page) In fact, of the 89 who died by 2nd September, only ten lived closer to any other pump, and in five of those cases he discovered that the dead person actually preferred the water from the Broad Street pump, and sent for it specially.

... he got the chance to prove his ideas in the most dramatic circumstances.

3

Figure 1: map of deaths due to cholera in the vicinity of the Broad Street pump

On Poland Street was the Workhouse, with 535 inmates, and surrounded on three sides by houses in which Cholera deaths had occurred. Yet only five people died there. The Work House had Its own well. Snow visited Mr Huggins of the Brewery in Broad St. He told him that they, too, had their own well but that as far as he knew the men never drank water at all - they stuck to beer. There were no deaths in the Brewery. But at the percussion cap factory at No 37 they weren't so lucky. Two tubs of water were kept for the workers, and 18 of them died. The water came from the Broad Street Pump.

Figure 2. Case Fatality (per 10,000 house) table for districts in the vicinity of the outbreak. The Broad Street Pump was belonged to the Southwark and Vauxhall company.

District Water Company Supplying the District

Number of Houses

Deaths from

Cholera

Death per 10,000 Houses

1 Southwark & Vauxhall

40,046 1263 315

2 Lambeth 26,107 98 37

4

Investigation The cases that clinched it for Snow concerned two ladies who died not in Soho, but in Hampstead, over five miles away, and in Islington, where there was no outbreak. Puzzled, Snow visited the house where one had died, and was told that every day a cart took a large bottle of water from the Broad Street pump all the way to Hampstead because the lady liked the taste. A delivery of water arrived on Thursday 31st August, and she drank then and on the Friday. By Saturday she was dead. The other lady was her niece, who paid a visit, drank the water, and then died at home in Islington.

By the 7th September, Soho was deserted. Three quarters of the people had fled - which helped to slow the outbreak. But there were still 28 new cases that day. In the evening, Snow met the Board of Guardians of St James's parish, and told them what he had found. The handle was removed from the pump the next day - and the number of cases immediately started to diminish.

Investigation of the Broad Street pump revealed what had probably been going on. The well below the pump was about 28 feet deep. At 22 feet down, within yards of the well, there was a sewer. A few people reported that the water had smelt "offensive", or that it "went off" near the time of the outbreak. Snow was now certain that the well had been contaminated with infected sewage - either from the sewer or the many nearby cesspits. As the outbreak continued, the sewage became more contaminated, and so did the water.

What is important about John Snow is his recognition of the power of statistics. He didn't know what the organism was that caused Cholera, so instead he gathered what might have been thought of as 'anecdotal' evidence - stories. But the cumulative effect of his meticulously gathered data was devastating - and was the beginning of the end for Cholera in Britain.

Questions:

1. List 9 pieces of evidence that are presented in the article?

2. Was the evidence descriptive or analytic in nature? If analytic mention the exposure that is being

investigated.

... the water had smelt "offensive" ...

5

Introduction to Epidemiology powerpoint

6

Introduction to Epidemiology powerpoint – 2

7


8


9

Classifying Research Epidemiologically 1

10

Classifying research Epidemiologically ‐ 2

11

Disease Frequency: Notes on calculating population based disease frequencies

You have seen prevalence and cumulative incidence before; The calculation of Incidence Rate is new!

You have likely not seen the word ‘candidate’. A candidate is a person can have the disease.

#

# (convert to % or per 1000, per 10,000 etc.)

#

# (convert to % or per 1000, per 10,000 etc.)

#

, ( do not convert)

Incidence Rate is new to most of you – and takes into account the unstable nature of populations over

time (e.g. people leave, die, are born etc.). In order to understand it, you need to be able to calculate

person time:

Person time (PY – person years, PM ‐ person months) Each individual is tracked from when they join

the study until they develop outcome or leave the study. Each individual contributes a number of

years (or months or weeks etc.) Person time for the denominator is the sum of all individual

contributions.

Example from ppt. lecture: Calculate the Incidence rate for the following, where ‘x’ is the outcome of

interest and PY is the person years each subject contributed to the study.

12

Disease frequency concepts and calculations – 1

13

Disease frequency concepts and calculations ‐ 2

14


15


16

Comparing disease frequency concepts and calculations ‐ 1

17

Comparing disease frequency concepts and calculations ‐ 2

18

Notes on the epidemiological approach to calculating Relative Risk and Risk Difference.

One of the fundamental building blocks within Analytic epidemiology is the 2 by 2 contingency table:

Outcome

Exp

osu

re Yes No Total Risk – is essentially the probability of

an individual in the ‘candidate population’ getting ‘outome x’. Risk is usually estimated through Cumulative Incidence – average risk over a period of time.

Yes a b a+b

No c d c+d Total a+c b+d a+b+c+d

Risk Difference (RD) = Rate or risk in exposed (RE) – Rate or risk in unexposed (RU)

Confidence intervals for RR can be calculated using the excel calculator provided on www.stataras.com

Odds Ratio: The odds ratio is an estimate of RR and is calculated in special circumstances (case‐control

studies for example). You may remember from Math1112 that odds are not the same as probabilities,

though they are often confused with probabilities. The odds of an event (E) happening is simply the ratio

of the probability of E {P(E)} over the probability the E will not happen {P(EC)}

Odds of E = )(

)(CEP

EP where P ( CE ) is the probability that E won’t happen {the complement of P (E).

Example : P(E) = 4/19, what are the odds of E: start off by calculating P(EC) = 15/19

then the odds of E = 15

4

19

1519

4

4:15 i.e. for every 4 people who get the disease E there are 15 who do not.

From the Contingency table the odds ratio can be calculate very easily with a short cut:

Odds of disease among exposed = a:b

Odds of disease among unexposed = c:d

Confidence intervals for odds ratio can be calculated using the excel calculator provided on www.stataras.com

For RR and Odds ratio – if 95% C.I. includes 1 then the ratios are not statistically significant and results

cannot be generalized.

19

Exercise 1: Practice with calculating population based disease frequency

An epidemiologic investigation begun on January 1st, 1999 identified a population of 1,000 individuals.

During the year of the study, 10 new cases were found (marked with a ‘q’ on the chart below). Among

the total of 10 cases, there were 6 deaths during the year (marked as an ‘x’ on the chart below).

For the 10 cases, the diagram indicates the time of case recognition (‘q’), periods of observation during

the study, and death (‘x’) or survival (). The 990 remaining individuals in the study did not become ill

or die during the year of observation. Assume that all deaths take place on the last day of month

(marked with a ‘!’ in the top row) and that ‘case recognition’ takes place on the 1st day of the month.

1999 2000 ! Jan ! Feb ! Mar ! Apr ! May ! Jun ! Jul ! Aug ! Sep ! Oct ! Nov ! Dec ! Jan 1 q-------------------------------------------------------------------------------------------- 2 q---------------------------------x 3 q----------------------------------------------------------x 4 q-----x 5 q---------------------------------------------------------------------------------------------- 6 q----------------------------------x 7 q------------------------------------------------------------------------------ 8 q-----------------x 9 q---------------------------------------------------------------------- 10 q-------------------------------x Calculate the following: Prevalence of the disease on:

i. January 1, 1999

ii. July 1, 1999

iii. December 31, 1999

CI (for the year)

IR (for disease among the participants) =

IR (for death among the diseased) =

20

Exercise 2: Practice with comparing disease frequency

The 58th annual convention of the American Legion was held in Philadelphia from July 21 until July 24, 1976. People at the convention included American Legion delegates, their families, and other legionnaires who were not official delegates. Between July 20th and August 30th, some of those who were or had been present became ill with a type of pneumonia subsequently named Legionnaire's Disease. No one attending the convention developed the disease after august 30th. Following are the numbers of delegates and non‐delegates who developed Legionnaire's Disease during the period July 20 to August 30 (41 day period).

Developed Legionnaire’s Disease

Del

egat

e Yes No Total Yes 125 1724 1849

No 3 759 762

Total 128 2483 2611

1. Compute the "rate" of Legionnaires' Disease for:

the delegates = ___________________ and for the non‐delegates ____________________

2. Calculate the Risk difference of Legionnaires' Disease for delegates vs non‐delegates. State in words the meaning of this rate difference

3. Calculate the Relative Risk of Legionnaires' Disease among delegates compared to non‐delegates. State in words the meaning of this rate ratio.

4. Use the excel calculator to calculate the 95% Confidence interval for RR. LB = ____ UB = ____ and

make a statement of comparison.

5. Are the results statistically significant? Explain how you know and what that means about your conclusion in 3.

6. Calculate the odds ratio of developing legionnaire’s disease in delegates.

21

Exercise 3: more Practice with comparing disease frequency

Using data from the 6 cities study discussed in the textbook pp. 60‐63, answer all questions

below.

Died Person time

Res

iden

ce

Yes No Total Ohio (most polluted)

291 1060 1351 17,914

Wisconsin (least polluted)

232 1399 1631 21,618

Total 523 2459 2982 39,532

1. Compute the mortality rate among the residents of the most polluted vs least polluted city.

a. as Cumulative Incidence b. as Incidence Ratio

Ohio Ohio

Wisconsin Wisconsin

2a. Calculate RD of death for residents of the two cities using CI. State in words the meaning of this

difference

2b. Calculate the IR Difference – the difference between Incidence Rates.

2c. Which of the two is more accurate measure of the actual difference in risks? Why?

3. Calculate the 95% C.I. for Relative Risk of death in residents of Ohio vs Wisconsin. State in words the

meaning of this rate ratio.

22

Exercise 4: More Practice with Comparing disease frequency

Read the article below and answer the questions on the side panel

Additionally we know that 377 men walked less than a mile a day

(162 died)and contributed 3322 person years to the study, and

that 330 walked more than 2 miles per day (73 died) contributing

3114 person years to the study.

What is the exposure under study –

how is it defined?

What are the outcomes under study?

Find the measures of population disease frequency (i.e. all deaths).

Cum Incidence =

IR =

Calculate measures of comparison.

RR =

RD =

IRD =

What did the investigators do to ensure that the comparisons were “fair?”

23

Exercise 5a classifying research type with review of contingency table

Australian researchers conducted a study of the relationship between optimism and colon cancer survival. Their hypothesis was that colon cancer patients who had a positive outlook on life would have a lower five‐year cumulative incidence of mortality. The study included 100 recently diagnosed colon cancer patients who underwent psychological testing and were found to have a optimistic outlook on life and 100 recently diagnosed colon cancer patients who underwent the same psychological tests and were found to have a pessimistic outlook on life. By the end of five years of follow‐up, 50 of the 100 patients with the optimistic outlook and 75 of the 100 patients with the pessimistic outlook had died from colon cancer. Research Question: Does being optimistic lead to longer life in those diagnosed with colon cancer?

Q1. Which type of research is the above and what makes it so?

Q2. Set up a contingency table and answer the research question using the data provided, then discuss the relevance of the evidence for establishing cause.

24

Exercise 5b classifying research types and focus on findings

A study was done to determine whether the amount of money spent on soft drinks was related to mortality from diabetes. The investigators collected data on per capita (average per person) soft drink consumption in ten US states and examined its relationship to mortality rates from diabetes in those ten states. In order to calculate per capita sales they gathered annual data on soft drink sales from commerce records and then divided these figures by the state’s population from the most recent census. The mortality data were gathered from the vital records department in each state.

Research Question: Does consuming soft drinks lead to excess mortality?


Q2. Answer the research question using the data provided and use Sir Hill’s guidelines to discuss whether there is evidence for establishing cause – you can take a look at a partial list below.

Sir A.B. Hill’s Guidelines for establishing causality (partial list only – more in chapter

15)

Time sequence established: There is evidence that exposure preceded disease.

Strength of association: Stronger associations are more likely to be causal.

Consistency: If other investigations using different populations, different study designs show

similar results, there is strong support for causality.

Biologic credibility: Does association make sense biologically?

Dose‐response: Does disease risk increase as exposure level increases?

25

Exercise 5c classifying research types and focus on findings

Phthalates are present in many diverse products, including insect repellents, body lotions, perfumes, and

food packaging. Because animal experiments suggest that phthalates may have an adverse effect on the

male reproductive system, a group of infertility specialists decided to conduct a study on phthalate

exposure and sperm abnormalities in adult men. 100 cases with sperm abnormalities were identified

from a fertility clinic’s records and 100 controls were identified from partner’s of female patients at their

infertility clinic. 30 cases and 10 controls had high urinary phthalate levels; the remainder had normal

urinary phthalate levels.

Research Question: Does Phthalate exposure lead to sperm abnormalities?


Q2. Answer the research question using the data provided and use Sir Hill’s guidelines to discuss whether there is evidence for establishing cause. To answer research question you will need to calculate the odds ratio. Start off by filling in the tables below.

Exposure to Phthalates

Yes No Total

Spe

rm

Abn

orm

aliti

es

No Yes

Total

Sperm Abnormalities

Yes No Total

Exp

osur

e to

P

htha

late

s

No Yes

Total

Calculate the odds ratio Those males with sperm abnormalities are ____ times more likely to have been exposed to Phthalates.

Calculate the odds ratio Those males with exposure are ____ times more likely to develop sperm abnormalities.

26

Exercise 5d thinking about case control

150 patients with gall bladder cancer were identified at a clinic. 75 of those patients ate foie gras

regularly and 75 did not. See table A below

Research Question: Does foie gras cause gall bladder cancer?

A. Gall Bladder Cancer

Yes No Total

Ate

Foi

e G

ras

No Yes 75 N/A

75 N/A

Total 150 N/A

There is not enough information to do any comparison of cancer rates in exposed to non‐exposed because all of the individuals have cancer. How do we find out if exposure to foie gras is a possible cause without doing a separate cohort study?

All we can calculate is the rate of foie gras eating in cancer patients (i.e. 50% of those with cancer ate foie gras, which seems like a lot).

Q1. There are two options – do a cohort study from scratch, or go out and collect foie gras eating data on those who do not have cancer. Typically, researchers collect 4 times more ‘controls’ (i.e. non‐cancer participants) than ‘cases’ (those with cancer). Take a look at table B. The second column represents results from a separate data collection of 600 patients without gall bladder cancer.

B. Gall Bladder Cancer

Yes No Total

Ate

Foi

e G

ras

No Yes

75 277 352

75 323 398

Total 150 600 750

Rate of exposure in diseased =

Rate of exposure in undiseased =

Rate of disease in exposed =

Rate of disease in unexposed =

We can’t compare the rates of disease in exposed vs unexposed because the numbers of cancer victims yes/no (150 and 600) were chosen by the researcher. i.e. the rates of disease are likely not correct. we know for sure that in the general population the rate of gall bladder cancer is not 150/750 = 20%.

Q2. In order to be able to comparing rates of disease, we need to estimate RR by calculating the odds ratio

(take a look at page 18 for formulas).

Odds ratio = ____________

Those who ate Foie Gras regularly are ______ times more likely to get Gall Bladder cancer

27

Experimental research

28

Experimental research ‐ 2

29

Experimental research ‐ 3

30

Cohort and case control Cohort and case control continued

Cohort and case control

31

Cohort and case control ‐2

32

Cohort and case control ‐ 3

33


34


35

Exercise 6: classifying analytic research types.

Read the following research reports and abstracts carefully and identify whether the research is descriptive, cross-sectional, experimental, cohort or case control and explain what makes them so.

Q1 Effect of habitual knuckle cracking on hand function Jorge Castellanos, David Axelrod Abstract The relation of habitual knuckle cracking to osteoarthrosis with functional impairment of the hand

has long been considered an old wives' tale without experimental support. The mechanical

sequelae of knuckle cracking have been shown to produce the rapid release of energy in the form

of sudden vibratory energy, much like the forces responsible for the destruction of hydraulic

blades and ship propellers. To investigate the relation of habitual knuckle cracking to hand

function 300 consecutive patients aged 45 years or above and without evidence of neuromuscular,

inflammatory, or malignant disease were evaluated for the presence of habitual knuckle cracking

and hand arthritis/dysfunction. The age and sex distribution of the patients (74 habitual knuckle

crackers, 226 non-knuckle crackers) was similar. There was no increased preponderance of

arthritis of the hand in either group; however, habitual knuckle crackers were more likely to have

hand swelling and lower grip strength. Habitual knuckle cracking was associated with manual

labour, biting of the nails, smoking, and drinking alcohol. It is concluded that habitual knuckle

cracking results in functional hand impairment.

Choice: Explanation:

36

Exercise 6: classifying analytic research types continued Q2 Crack Research: Good news about knuckle cracking Despite the popularity of knucle cracking, most known knuckle crackers have probably been told

by some expert—whose advice very likely began, “I’m not a doctor, but ...”—that the behavior

would lead to arthritis. One M.D. convincingly put that amateur argument to rest with a study

published back in 1998 in the journal Arthritis & Rheumatism entitled “Does Knuckle Cracking

Lead to Arthritis of the Fingers?” The work of sole author Donald Unger was back in the news in

early October when he was honored as the recipient of this year’s Ig Nobel Prize in Medicine.

The Igs, for the uninitiated, are presented annually on the eve of the real Nobel Prizes by the

organization Improbable Research for “achievements that first make people laugh, and then make

them think.” In Unger’s case, I thought about whether his protocol might be evidence that he is

obsessive-compulsive. From his publication: “For 50 years, the author cracked the knuckles of his

left hand at least twice a day, leaving those on the right as a control. Thus, the knuckles on the left

were cracked at least 36,500 times, while those on the right cracked rarely and spontaneously.”

Unger undertook his self and righteous research because, as he wrote, “During the author’s

childhood, various renowned authorities (his mother, several aunts and, later, his mother-in-law

[personal communication]) informed him that cracking his knuckles would lead

to arthritis of the fingers.” He thus used a half-century “to test the accuracy of this hypothesis,”

during which he could cleverly tell any unsolicited advice givers that the results weren’t in yet.

Finally, after five decades, Unger analyzed his data set: “There was no arthritis in either hand, and

no apparent differences between the two hands.” He concluded that “there is no apparent

relationship between knuckle cracking and the subsequent development of arthritis of the fingers.”


37

Exercise 6 identifying analytic research types - continued

Q3. Coffee consumption unrelated to alertness 2 June 2010

The stimulatory effects of caffeine may be nothing more than an illusion according to new BBSRC‐funded research, which shows there is no real benefit to be gained from the habitual morning cup of coffee.

Tests on 379 individuals who abstained from caffeine for 16 hours before being given either caffeine or a placebo and then tested for a range of responses showed little variance in levels of alertness.

The study, published online in the journal of Neuropsychopharmacology, reports that frequent coffee drinkers develop a tolerance to both the anxiety‐producing effects and the stimulatory effects of caffeine. While frequent consumers may feel alerted by coffee, evidence suggests that this is actually merely the reversal of the fatiguing effects of acute caffeine withdrawal. And given the increased propensity to anxiety and raised blood pressure induced by caffeine consumption, there is no net benefit to be gained.

Peter Rogers, from the University of Bristol's Department of Experimental Psychology and one of the lead authors of the study, said: "Our study shows that we don't gain an advantage from consuming caffeine ‐ although we feel alerted by it, this is caffeine just bringing us back to normal. On the other hand, while caffeine can increase anxiety, tolerance means that for most caffeine consumers this effect is negligible."

Approximately half of the participants were non/low caffeine consumers and the other half were medium/high caffeine consumers. All were asked to rate their personal levels of anxiety, alertness and headache before and after being given either the caffeine or the placebo. They were also asked to carry out a series of computer tasks to test for their levels of memory, attentiveness and vigilance.

The medium/high caffeine consumers who received the placebo reported a decrease in alertness and an increase in headache, neither of which were reported by those who received caffeine. However, their post‐caffeine levels of alertness were no higher than the non/low consumers who received a placebo, suggesting caffeine only brings coffee drinkers back up to 'normal'.


38

Exercise 6 identifying analytic research types - continued

Q4. Randomised prostate cancer screening trial: 20 year follow-up

Gabriel Sandblom, associate professor1, Eberhard Varenhorst, professor2, Johan Rosell, statistician3, Owe Löfman, professor4, Per Carlsson, professor5 Accepted 24 December 2010 BMJ

ABSTRACT Objective: To assess whether screening for prostate cancer reduces prostate cancer specific mortality. Design: Population based randomised controlled trial. Setting: Department of Urology, Norrköping, and theSouth-East Region Prostate Cancer Register. Participants: All men aged 50-69 in the city of Norrköping, Sweden, identified in 1987 in the National Population Register (n=9026). Intervention: From the study population, 1494 men were randomly allocated to be screened by including every sixth man from a list of dates of birth. These men were invited to be screened every third year from 1987 to 1996. On the first two occasions screening was done by digital rectal examination only. From 1993, this was combined with prostate specific antigen testing, with 4 μg/L as cut off. On the fourth occasion (1996), only men aged 69 or under at the time of the investigation were invited. Main outcome measures: Data on tumour stage, grade, and treatment from the South East Region Prostate Cancer Register. Prostate cancer specific mortality up to 31 December 2008. Results: In the four screenings from 1987 to 1996 attendance was 1161/1492 (78%), 957/1363 (70%), 895/1210 (74%), and 446/606 (74%), respectively. There were 85 cases (5.7%) of prostate cancer diagnosed in the screened group and 292 (3.9%) in the control group. The risk ratio for death from prostate cancer in the screening group was 1.16 (95% confidence interval 0.78 to 1.73). In a Cox proportional hazard analysis comparing prostate cancer specific survival in the control group with that in the screened group, the hazard ratio for death from prostate cancer was 1.23 (0.94 to 1.62; P=0.13). After adjustment for age at start of the study, the hazard ratio was 1.58 (1.06 to 2.36; P=0.024). Conclusions After 20 years of follow-up the rate of death from prostate cancer did not differ significantly between men in the screening group and those in the control group.


39

Exercise 6 identifying analytic research types

Q5. Daily Candy Consumption in Children Linked to Violence in Adulthood Deborah Brauser

October 5, 2009 — Excessive consumption of confectionery (sweets/chocolates) during childhood can increase the likelihood of adult aggression, according to results from a large cohort study published in the October issue of the British Journal of Psychiatry.

In this study, "children who ate confectionery daily at age 10 years were significantly more likely to have been convicted for violence at age 34 years," write Simon C. Moore, PhD, and colleagues from the Violence and Society Research Group and Applied Clinical Research and Public Health in the School of Dentistry at Cardiff University in Wales.

Although past studies have shown a causal effect between diet and behavioral problems, including aggression, this is the first study to look at the long-term effects of childhood diet on adult violence, report the study authors.

In addition, "we think that kids who demand candy every day do not learn to delay gratification, which some think is a feature of violent individuals generally," Dr. Simon told Medscape Psychiatry.

"The main takeaway for clinicians is that small changes early in life can make a profound difference in childhood," he added.

Background

"We've been researching offending behavior in youngsters for a long time and we've found that problem behavior seems to be associated with impulsive risk taking," explained Dr. Simon. "We also noticed that the diets of these youngsters were pretty poor. Coupled with other research suggesting a link between diet and behavior, we decided to look at data to see whether confectionery consumption was linked to adult violence."

The investigators examined data from 17,415 participants in the British Cohort Study born between April 5 and April 11, 1970 in the United Kingdom. That cohort study collected data on health, education, and social and economic status from respondents at varying stages from age 5 to 42 years.

Consistent Association Found Between Confectionery and Violence

Results showed that 69.7% of the study participants who were violent by the age of 34 reported that they ate confectionery nearly every day during childhood, whereas only 42% of those who were nonviolent did.

Other significant relations between control variables and violence included the child's sex (male) and the parents' attitude toward parenting. Access to motorized transport at 34 years was found to protect against adult violence, whereas living in a rural area at 34 years increased the risk for violence.

"We found the main result quite surprising and interpret it as perhaps evidence that it is the way that confectionery is given to kids rather than the candy itself," said Dr. Simon.

He explained that a plausible mechanism for the link is that persistently using confectionery to control childhood behavior might prevent children from learning to defer gratification, which can lead to more impulsive behaviors. These are both biases that are strongly associated with delinquency.

"Irrespective of the causal mechanism, which warrants further attention, targeting resources at improving childhood diet may improve health and reduce aggression," write the study authors.

"The next step is to better understand how social, psychological, and consumption factors interact to produce problem behavior," concluded Dr. Simon.

This study was supported by a grant from the Economic and Social Research Council. The study authors have disclosed no relevant financial relationships. Br J Psychiatry. 2009;195:366-367. Choice: Explanation:

40


Q6. Three-second distraction doubles work errors Posted By Andy Henion-Michigan State On January 8, 2013 @ 11:29 am In Top Stories | 2 Comments MICHIGAN STATE (US) — Even tiny interruptions can derail your train of thought and increase mistakes, new research shows. Short interruptions—such as the few seconds it takes to silence that buzzing smartphone— have a surprisingly large effect on one’s ability to accurately complete a task, according to new research. The study, in which 300 people performed a sequence-based procedure on a computer, found that interruptions of about three seconds doubled the error rate. Straight from the Source Read the original study [1] DOI: 10.1037/a0030986 Brief interruptions are ubiquitous in today’s society, from text messages to a work colleague poking his head in the door and interrupting an important conversation. But the ensuing errors can be disastrous for professionals such as airplane mechanics and emergency room doctors, says Erik Altmann, lead researcher on the study and associate professor of psychology at Michigan State University. “What this means is that our health and safety is, on some level, contingent on whether the people looking after it have been interrupted,” says Altmann. The study, published in Journal of Experimental Psychology: General [1], is one of the first to examine brief interruptions of relatively difficult tasks. Study participants were asked to perform a series of tasks in order, such as identifying with a keystroke whether a letter was closer to the start or the end of the alphabet. Even without interruptions a small number of errors in sequence were made. Sometimes participants were interrupted and told to type two letters—which took 2.8 seconds—before returning to the task. When this happened, they were twice as likely to mess up the sequence. Altmann says he was surprised that such short interruptions had a large effect. The interruptions lasted no longer than each step of the main task, he noted, so the time factor likely wasn’t the cause of the errors. “So why did the error rate go up?” Altmann asks. “The answer is that the participants had to shift their attention from one task to another. Even momentary interruptions can seem jarring when they occur during a process that takes considerable thought.” One potential solution, particularly when errors would be costly, is to design an environment that protects against interruptions. “So before you enter this critical phase: All cell phones off at the very least,” Altmann says. Gregory Trafton of the Naval Research Laboratory and Zach Hambrick of Michigan State are co-authors of the study, which was funded by the US Nav


41


Q7. Literacy education as treatment for depression in patients with limited literacy and depression: a randomized controlled trial.

Weiss, Barry D, Francis, Laurie; Senf, Janet H; Heist, Kim; Hargraves, Rie

BACKGROUND: Individuals with limited literacy and those with depression share many characteristics, including low self-esteem, feelings of worthlessness, and shame.

OBJECTIVE: To determine whether literacy education, provided along with standard depression treatment to adults with depression and limited literacy, would result in greater improvement in depression than would standard depression treatment alone.

DESIGN: Randomized clinical trial with patients assigned either to an intervention group that received standard depression treatment plus literacy education, or a control group that received only standard depression treatment.

PARTICIPANTS: Seventy adult patients of a community health center who tested positive for depression using the 9-question Patient Health Questionnaire (PHQ-9) and had limited literacy based on the Rapid Estimate of Adult Literacy in Medicine (REALM).

MEASUREMENTS: Depression severity was assessed with PHQ-9 scores at baseline and at 3 follow-up evaluations that took place up to 1 year after study enrollment. Changes in PHQ-9 scores between baseline and follow-up evaluations were compared between the intervention and control groups.

RESULTS: The median PHQ-9 scores were similar in both the intervention and control groups at baseline (12.5 and 14, respectively). Nine-question Patient Health Questionnaire scores improved in both groups, but the improvement was significantly larger in the intervention group. The final follow-up PHQ-9 scores declined to 6 in the intervention group but only to 10 in the control group.

CONCLUSIONS: There may be benefit to assessing the literacy skills of patients who are depressed, and recommending that patients with both depression and limited literacy consider enrolling in adult education classes as an adjuvant treatment for depression.

Weiss, B. D., Francis, L., Senf, J. H., Heist, K., & Hargraves, R. (2006). Literacy education as treatment for depression in patients with limited literacy and depression: a randomized controlled trial. Journal of general internal medicine, 21(8), 823–8. doi:10.1111/j.1525-1497.2006.00531.x


42

Exercise 7a: practice with case control and cohort calculations A case-control study of the effectiveness of bicycle safety helmets Thompson RS, Rivara FP, Thompson DC New England Journal of Medicine 1989, Vol 320 No 21 p1361-7 Authors' abstract Bicycling accidents cause many serious injuries and, in the United States, about 1300 deaths per year, mainly from head injuries. Safety helmets are widely recommended for cyclists, but convincing evidence of their effectiveness is lacking. Over one year we conducted a case-control study in which the case patients were 235 persons with head injuries received while bicycling, who sought emergency care at one of five hospitals. One control group consisted of 433 persons who received emergency care at the same hospitals for bicycling injuries not involving the head. A second control group consisted of 558 members of a large health maintenance organization who had had bicycling accidents during the previous year. 7.23% of the case patients were wearing helmets at the time of their head injuries, as compared with 23.51 percent of the controls. Q1. Who/what was the unit of analysis Q2. calculations Wore Helmet

Yes No Total

Hea

d In

jury

No Yes

Total

Head Injury

Yes No Total

Wor

e H

elm

et

No Yes

Total

Calculate the odds ratio Those who had hed injuries were ____ times as likely to wear a helmet.

Calculate the odds ratio Those who wore helmets were ____ times as likely to sustain head injuries.

Findings of researchers: In regression analyses to control for age, sex, income, education, cycling experience, and the severity of the accident, we found that riders with helmets had an 85 percent reduction in their risk of head injury (odds ratio, 0.15; 95 percent confidence interval, 0.07 to 0.29). We conclude that bicycle safety helmets are highly effective in preventing head injury. Helmets are particularly important for children, since they suffer the majority of serious head injuries from bicycling accidents. Q3. Why is their result for odds ratio different?

43

Exercise 7b (based on scenario 5 from exercise 6) Daily Candy Consumption in Children Linked to Violence in Adulthood Deborah Brauser

October 5, 2009 — Excessive consumption of confectionery (sweets/chocolates) during childhood can increase

the likelihood of adult aggression, according to results from a large cohort study published in the October issue

of the British Journal of Psychiatry.

In this study, "children who ate confectionery daily at age 10 years were significantly more likely to have been

convicted for violence at age 34 years," write Simon C. Moore, PhD, and colleagues from the Violence and

Society Research Group and Applied Clinical Research and Public Health in the School of Dentistry at Cardiff

University in Wales.

The investigators examined data from 17,415 participants in the British Cohort Study born between April 5 and

April 11, 1970 in the United Kingdom. That cohort study collected data on health, education, and social and

economic status from respondents at varying stages from age 5 to 42 years.

Consistent Association Found Between Confectionery and Violence

Results showed that 69.51% of the study participants who were violent by the age of 34 reported that they ate

confectionery nearly every day during childhood, whereas only 42% of those who were nonviolent did.

"We found the main result quite surprising and interpret it as perhaps evidence that it is the way that

confectionery is given to kids rather than the candy itself," said Dr. Simon.

Additional crucial information: 0.471% of all individuals were incarcerated for violent crimes by the age of 34.

Q1. What type of study is this?

Q2. calculations

Incarcerated for Violent crime by 34

Yes No

Total

Ate

Can

dy

No Yes

Total

Calculate the appropriate measure of comparison Those who had ate candy were ____ times as likely to be incarcerated for violent crime by the age of 34.

Q3. Do you agree that candy ‘can increase the likelihood of adult aggression’? Explain your answer.

44

Exercise 7c Distinguishing case control vs. retrospective cohort study.

Case‐Control Example: start with outcome (yes =case vs no = control) and look for exposure(s).

Abstract: Stretesky, Paul., Pogrebin, Mark., Unnithan, Prabha. and Hogan, Michael; National Case‐

Control Study of Homicide Offending and Methamphetamine Use Submitted to American Society of

Criminology for presentation at 2007 Annual Meeting in Atlanta (GA), November 14‐17,2007 at Atlanta

Marriott Marquis.

Previous research suggests that methamphetamine use may lead to violence of various kinds. Our own research in Colorado indicates that a significant number of homicide offenders used methamphetamines. Building upon that potential relationship we use existing data to ask the following research question: Does methamphetamine use increase the likelihood that a person will commit a criminal homicide? We address this question using quantitative methods grounded in a case-control approach which links two originally separate samples into one large sample. The cases represent inmates in state and federal prisons interviewed by the U.S. Bureau of the Census between June and October 1997 in their Survey of Inmates in State and Federal Correctional Facilities who had committed a homicide as an adult between 1990 and 1997. The controls consist of a general sample of U.S. adults (age 18 or older) interviewed in the National Survey of Drug Use and Health in 1994, 1995, 1996, 1997. We assess the association between methamphetamine use and homicide offending through the use of multivariate logistic regression. Results indicate that a statistically significant relationship between methamphetamine use and homicide does exist. However, the methamphetamine-homicide relationship it is no different than the relationship between general drug use (e.g., use of cocaine/crack, heroin, and marijuana) and homicide offending. Thus, there appears to be nothing unique about methamphetamine use and homicide offending.

What is outcome variable?

What is/are exposure variable(s)?

Sketch a map of the study design below

45

Exercise 7d Restrospective cohort : exposure and outcomes both took place in the past.

International retrospective cohort study of neural tube defects in relation to folic acid recommendations: are the recommendations working?

Lorenzo D Botto, Alessandra Lisi, Elisabeth Robert-Gnansia, J David Erickson, Stein Emil Vollset, Pierpaolo Mastroiacovo, Beverley Botting, Guido Cocchi, Catherine de Vigan, Hermien de Walle, Maria Feijoo, Lorentz M Irgens, Bob McDonnell, Paul Merlob, Annukka Ritvanen, Gioacchino Scarano, Csaba Siffel, Julia Metneki, Claude Stoll, Richard Smithells, Janine Goujard,

Accepted 9 December 2004

Abstract

Objectives To evaluate the effectiveness of policies and recommendations on folic acid aimed at reducing the

occurrence of neural tube defects.

Design Retrospective cohort study of births monitored by birth defect registries.

Setting 13 birth defects registries monitoring rates of neural tube defects from 1988 to 1998 in Norway, Finland,

Northern Netherlands, England and Wales, Ireland, France (Paris, Strasbourg, and Central East), Hungary, Italy

(Emilia Romagna and Campania), Portugal, and Israel. Cases of neural tube defects were ascertained among liveborn

infants, stillbirths, and pregnancy terminations (where legal). Policies and recommendations were ascertained by

interview and literature review.

Main outcome measures Incidences and trends in rates of neural tube defects before and after 1992 (the year of the

first recommendations) and before and after the year of local recommendations (when applicable).

Results The issuing of recommendations on folic acid was followed by no detectable improvement in the trends of

incidence of neural tube defects.

Conclusions Recommendations alone did not seem to influence trends in neural tube defects up to six years after the

confirmation of the effectiveness of folic acid in clinical trials. New cases of neural tube defects preventable by folic

acid continue to accumulate. A reasonable strategy would be to quickly integrate food fortification with fuller

implementation of recommendations on supplements.

What is outcome variable?

What is/are exposure variable(s)?

Sketch a map of the study design below

46

Template for analysis of research articles in epidemiology

Questions to answer when reading Research and Research Abstracts

1. Write out reference (citation) in proper APA format – an example of a citation to a journal article is given below. Peckover, S., & Winterburn, S. (2003). Teaching research to undergraduate community nursing students: reflections upon curriculum design. Nurse Education in practice, 3, 104-111.

2. What is the research question (or research questions) to be answered by this study?

3. What type of epidemiological study is this? What elements make it clear that your choice is correct?

4a. What is the unit of analysis?

4b. What is/are the independent(exposure, explanatory) variable(s)? What is/are the dependent (outcome) variables? . i.e. Describe the treatment and control group (or case/control groups).

5. Describe methods employed in order to avoid bias (selection and/or observation).

6. Make a list of confounders identified by the researchers, and potential confounders that you think were overlooked.

7. Describe methods employed to avoid the effect of confounders. (in design and/or analysis phase)

8. External validity: What is the population to which the investigators intend to refer their findings? Is external validity discussed? Are their generalizations valid? Explain.

9. Data analysis tools used (list them even though many of these will be unknown to you!). Focus especially on discussions of practically significant differences and confidence intervals not just statistically significant differences (p-values).

10a. Major conclusion/findings – i.e. What is the answer to research question?

10b. Other findings/results – not always present, but often in research more than just one main research question is addressed.

47

Notes on Validity and its components

Scientists read the results of research skeptically. They do this because they carry a great responsibility as experts in their field in helping us decide which medicines to use, which foods to avoid, which pipelines to build and which materials we should build our houses from. Decisions need to be based on evidence, and unfortunately it is very rare that evidence allows us to be 100% sure that our decision is correct.

In the statistics and research methods courses that you have completed we focused mostly on the data analysis aspect of research, but we have also spent time with data collection (internal validity) and sampling (external validity). These two are broad components of research studies that need to be assessed before research gets published and accepted as ‘true’. The word validity (and value!) comes from the Latin word for strength and effectiveness (valere), and is appropriate to use when we aim to assess the utility of a study’s findings. It is important that research results have strong (significant) findings and that it be conducted in a proper (valid) way so that we can be sure that these strong findings are not based on a weak foundation.

In the abstract of the Downs &Black (1997) article/checklist that you have received the authors describe different types of validity that they used to design the instrument (27 question checklist) that can be used to evaluate research studies: Face Validity (do experts consider the study/instrument valid – i.e. are methods used to collect data, recruit participants and analyse data generally accepted as valid), Content Validity (does the study/instrument tackle all aspects of the problem that is being posed), Criterion Validity (is the study/instrument useful, can it predict something). There are many other types of validity out there many of which are appropriate only in very specific spheres of research and science.

Validity is also related to three other important terms: accuracy, precision and reliability. Reliability involves an assessment of an instrument’s consistency. For example, when placing a baby on a scale three separate times one should get the same weight each time. If the scale produces three very different results then the scale is not reliable. Accuracy and precision are best illustrated by an example suggested to me by a student: Given that a person whom you are looking for lives in a big city like Toronto may be accurate, but is not precise. You can go to Toronto, but are very unlikely to find him or her unless you get a more precise location. Given that person’s postal code or GPS coordinates of that person’s house or location is much more precise measure of location. Nevertheless, if you have the wrong postal code or the wrong GPS coordinates then you have good precision but poor accuracy. Thus for valid measures it is best to be accurate and precise!

48

Notes on Validity and its components continued

Typically one would evaluate research findings by evaluating components of the research separately; i.e. data collection, sampling/recruitment, and analysis. The textbook is organized a bit differently as it focuses on three major threats to validity: Bias, Confounding and Random Error. These are discussed in detail in chapters 10 – 12 of the textbook. Before reading those chapters (and the ppt slides in the pages to follow) make sure that you take a look at the Downs and Black (1997) checklist and make sure that you are clear about the differences between internal and external validity as described below. The textbook writers chose not to spend time on these components of validity.

1. Internal validity involves an assessment of the validity of the findings of the study itself and usually involves assessing the validity of data collection – including the instruments used (i.e. measurements, observation methods, surveys). Bias and Confounding pose threats to internal validity. Random Error is a rare threat to data collection – but is possible.

2. External validity has to do with generalizing the findings to a population beyond the sample that was in the study. It involves an assessment of the sampling method. There are two possible threats to external validity (both are related to taking a sample properly). Selection bias is one, and the other is random error. In the textbook the focus is on random error as a threat. However in Downs and Black (1997) questions 11, 12 and 13 focus on the validity of the sampling method by asking whether the sample is a good one.

3. In the textbook section on sampling, and in the discussion of external validity in Downs and Black (1997) an admission is made that random sampling is rarely conducted in epidemiological research and thus many of the inferential statistics tools (statistical significance) may not be valid. This is tremendously important to keep in mind!

49

Internal and External validity ‐ 1

50

Internal and External validity


51


52


53


54

Exercise 8: Practice with identifying types of Selection Bias.

1. Which type of selection bias (self-selection, restriction, non-response bias) is the following an example of?

“In a study on relating obesity to television viewing in children, interviews were conducted with families in order to gauge the number of hours of television watched per week. Given that the sample was a selection of volunteers we could expect families where more television is watched to be less likely to participate in the survey.”

2. Do PAP smears help prevent cervical cancer? a case control study

• Cases: diagnosed with cervical cancer at a city hospital.

• Controls : canvassed the neighbourhood from which hospital patients tend to come on foot during the day.

Is there anything wrong with using a sample of women who were home during the day as controls? Which type of selection bias would this be?

Had Pap Smear

Cer

vica

l C

ance

r

Yes No Total Yes (cases)

100 150 250

No (controls)

750 250 1000

55

Exercise 9: Stratification Example – prep for validity lecture

The goal is to tease out effects of possible confounders by looking at whether there is a change in the results once the analysis is stratified and then the measures adjusted.

Below is a contingency table representing the results of a study in which the goal was to assess the relationship between Binge Drinking (defined as having more than 5 drinks in one night) and self-perceived stress. 331 residents of Sudbury Ontario were asked whether they had had at least one binge drinking episode in 2011, and their stress level was assessed as either High or Low.

Research Question: Are people who had at least one binge drinking episode more likely to be highly stressed than those who did not?

Step 1. Calculate the overall relative risk

High Stress

Binge Drinking In Prev Year

Yes No Total Rates

Yes 39 24 63 % The risk of High Stress in drinkers

No 114 154 268 % The risk of High Stress in non drinkers

153 178 331

RR = _____ (this will be called the crude RR) Answer to research question:

Step 2. The data was split into two groups (stratified) by age. Calculate RR for each group separately.

Age: Under 40 Age: 40 and up

Binge Drinking in Prev Year

High Stress Binge

Drinking in Prev Year

High Stress

Yes No Yes No

Yes 21 17 38 Yes 18 7 25

No 26 59 85 No 88 95 183

47 76 123 106 102 208

Under 40 RR = ______ 40 and over RR = ____

Step3. Answer the RQ given that it seems that age is a confounding variable.

56

Exercise 10a (optional) Adjusting for possible confounder: adjusting for age (based on method used on page 72 of the textbook )

Example 1: We will be comparing the risk of bacteriuria in oral contraceptive (OC) users and non‐users.

Crude results

Bacteriuria

OC Use

Yes No Total Rates

Yes 27 455 482 5.60% The risk of bacteriuria in OC users

No 78 1860 1938 4.02% The risk of bacteriuria in non ‐OC users

105 2315 2420

RRCrude = 5.6/4.02 = 1.39 OC users are 1.39 times more likely to contract bacteriuria than non users.

Adjusted by 4 categories of age: It was thought that age may be a possible

confounder so the results were stratified before standardization as follows:

OC users Non‐OC users Total study

Age Group

Rate (%) of bacteriuria(N)

N (total in age group)

Rate (%) of bacteriuria (N)

N (total in age group)

N (pooled)by age

16‐19 1.19 (1) 84 3.20 (9) 281 84+281=365

20‐29 5.63 (16) 284 3.99 (22) 552 836

30‐39 6.25 (6) 96 5.46 (34) 623 719

40‐49 22.22 (4) 18 2.70 (13) 482 500

Total = 482 Total = 1938 Total = 2420

Steps to age standardization of relative risk. (pg. 72 method)

1. We can see that the age distribution of OC users is different from the age distribution

of non‐OC users

2. The logic is to take the % with bacteriuria information in OC users and adapt that to the

age distribution of the ‘total study’ population – to see what the rate would be if age was

not a factor.

3. standardized risk of bacteriuria in OC users . . . .

=0.08574

4. standardized risk of bacteriuria in non‐OC users . . . .

=0.04039

5. The standardized relative risk RRAdjusted = 2.123. quite a bit higher than the crude rate.(RR = 1.39) from

above.

Multiply the

corresponding

pooled N by each of

the rates of disease

57

Exercise 10b (optional) solutions to exercise 9 using adjusted rates

Step 4a (optional) Calculate age adjusted RR by using Mantel-Haenszel’s cohort formula (see pg. 299)

RRadjusted = .

.1.63, which is a bit higher

than the crude RR of High Stress in Binge drinkers.

Step4b. Using the page 72method Adjusted by 4 categories of age: It was

thought that age may be a possible confounder so the results were stratified

before standardization as follows:

Binge No Binge Total study

Age Group

Rate of high stress

N

Rate of high stress

N

N (pooled by age)

<40 21/38 38 26/85 85 123

>=40 18/25 25 88/183 183 208

Total = 63 Total = 268 Total = 331

Steps to age standardization of relative risk.

1. We can see that the age distribution of Bingers is different from the age distribution of non‐Bingers

2. The logic is to take the % with High‐stress information in Bingers and adapt that to the age distribution

of the ‘total study’ population – to see what the rate would be if age was not a factor.

3. age standardized risk of High Stress in Bingers / /

=0.65781

4. age standardized risk of High Stress in non‐Bingers / /

=0.41585

5. The standardized relative risk RRAdjusted = 0.65781/0.41585 = 1.5818.

a bit higher than the crude rate, but not as high as the Mantel‐Haenszel result (RRAdjusted=1.63).

Answer to RQ – It seems that no matter how you approach the calculations those that binge drink are at slightly higher risk for stress (crude RR = 1.45). When adjusting for age distribution the rates is slightly higher closer to RRAdjusted ≈ 1.6. It does not seem that age is a strong confounding variable in this example.

58

Causation ‐ 1

59

Causation ‐ 2

60

Causation ‐ 3

61

Exercise 11: Read the article below and identify which of Hill’s 9-point guideline for establishing causality has been met.

Would you start drinking 4 cups of green tea a day based on these findings?

Green Tea a Performance-enhancing Drug

Green tea may soon show up in locker rooms and doping tests after being found to boost exercise endurance in mice

up to 24% while spurring the use of fat as energy.

The finding is based on green tea extract (GTE). and is difficult to extrapolate to human athletes. Japanese researcher

Takatoshi Murase estimates that to match the observed effects, athletes weighing 75 kilograms (165 pounds) would

need to drink about four cups of green tea a day—and over several weeks. "One of our important findings," says

Murase, "was that a single high-dose of GTE or its active ingredients didn't affect performance. It is the long-term

ingestion of GTE that is beneficial."

Exercise boost

To test the theory that catechins boost endurance capacity by stimulating fat burning, the researchers used mice

swimming in an adjustable-current water pool. Some of the mice received no green tea compounds, others received

green tea extract and still others received only a catechin in green tea known as EGCG.

While acknowledging that the impact of dietary interventions on performance is controversial, the researchers note that

compounds in green tea called catechins have already been found to have various physiological effects.

Mice on no supplements could swim on average 33 minutes before they were exhausted. Mice on green tea extract

consistently performed better after the first week and by week nine those taking 0.5% green tea extract by weight could

swim on average 40 minutes compared to 33 for the controls. A similar effect was observed in mice on EGCG,

suggesting that the catechin was at least partly responsible for the benefits.

To support their theory about fat burning, the researchers found that fatty acids in blood increased slightly but

significantly in mice on the supplements. They say that their findings suggest that green tea extract enhanced the

ability of muscle to use fatty acids as an energy source.

Supplemental benefits

Previous results showed that catechins also helped in counteracting obesity from a high-fat diet, suggesting that

catechins stimulate fat oxidation.

It's thought that the fat oxidation improves exercise performance by allowing the body to get energy from fat rather than

carbohydrates during endurance activities.

To avoid potential complicating factors in other studies, the researchers controlled for possible influences of caffeine—

a known performance enhancer—and changes that might have affected the animals' buoyancy.

The study was conducted by Murase and colleagues at the Biological Sciences Laboratories of Kao Corp. in Tochigi,

Japan—a company that makes green tea beverages and has been investigating the tea's anti-obesity effects.

The researchers say their findings show that green tea extract can boost exercise capacity and support the hypothesis

that stimulating the use of fatty acids can improve endurance.

The next steps are to determine the molecular mechanism by which green tea stimulates fat burning and whether the

antioxidant properties of catechins mediate their effects on endurance capacity.

62

Screening ‐ 1

63

Screening ‐ 2

64

Screening ‐ 3

65

Exercise 12 Screening practice exercise Your boss has just developed a new screening test for a disease and you are in charge of testing its validity. You decide to evaluate the test on 1000 individuals 300 who have had diagnoses of the disease and 700 who have been found not to have the disease The screening test gave a positive result for 292 individuals. 285 of these individuals actually had the disease. a) Fill in all cells of the two by two table.

Diagnosis of Disease Total

Res

ults

of

Scr

eeni

ng T

est

Yes No

Positive

Negative

Total

b. Calculate the sensitivity of the new screening test. c. Interpret the results of sensitivity calculation in one sentence. d. Calculate the specificity of the new screening test. e. Interpret the results of specificity calculation in one sentence. f. Calculate the Probability of a false positive and interpret the results in a statement. g. Calculate the probability of a false negative and interpret the results in a statement.

66

Exercise 13: practice with findings from data Author: Steiner, Ivan, P. et al; from Canadian Journal of Emergency Medicine (2009) Title: Impact of a nurse practitioner on patient care in a Canadian emergency department

Objective: Our objective was to determine whether the addition of a broad-scope nurse practitioner (NP) would improve emergency department (ED) wait times, ED lengths of stay (LOS) and left-without-treatment (LWOT) rates. Comparisons were made to shifts where the Emergency Physician worked alone. We hypothesized that the addition of a broad-scope NP during weekday ED shifts would result in shorter patient wait times, reduced LOS, fewer patients leaving the ED without treatment, and higher numbers of patients per shift.

Questions Q1: What is/are the exposures of interest? Q2: What is/are the outcome(s) of interest? Q3: What is the overall research question that the researchers are asking? Q4: Answer the research question using data from Table 2. Come to an overall conclusion by using evidence from each of the outcome variables provided.

Q5: Describe one measure of practical significance the researchers could have calculated?

67

The great epidemiology debates

Outline of format

Task 1: As a class we choose Topics and teams for and against

Task 2: Individually you will conduct a lit review (10%) on your topic, summarizing one article supporting

and one contradicting your position. Organize these strategically with your team in order that all aspects

of your debates position are considered.

Task 3. Organize your team’s debate

Great debates format: 4 topics, 8 debating teams, up to 6 members on each team;

time: ≈ 30‐45 minutes per debate for a total of 90 minutes of commitment – you need to be there for

one pair of debates.

Date: Debates 1 and 2 in week 13 (March 31st) after the data analysis assignment; debates 3 and 4 in

week 14 (April 7th) starting at 11AM.

Each debate will have 6 parts as follows

1. opening statement: each team has up to 6 members: 2 minutes each for 2‐3 team members ( 4‐6

minutes for each team)

2. Questions from audience to each of the teams – for clarification purposes (2‐5 minutes)

3. 5‐10 minute break for composition of rebuttal

4. rebuttal: 2 team members speak for 2 minutes each (total 4 minutes)

5. Final statement: 5th or 6th member of team closes out and summarizes the evidence (2 minutes)

6. Vote for the winner and general discussion.

epidemiology 2001 winter 2016–readings exercisesfaculty.georgebrown.ca/~tgula/epidemiology/week1...

Documents