EBCP CAT Assignment – Harm
Student Number: 28253809
Word Count:
CAT Assignment
Clinical Vignette
Tom is a 60-year-old man who as a result of the new bowel screening program presents to his GP for a routine colonoscopy. He tells his GP, Dr Dan, that he has no previously diagnosed medical conditions but has a family history of colorectal cancer, with his father having been diagnosed within the last year. Tom is concerned and asks Dr Dan if there is anything he can do to prevent the risk of him developing colorectal cancer himself. He tells Dr Dan that a few men at work have decided to cut down on their red meat intake as they heard there was a link between red meat and colorectal cancer. Tom loves his red meat and so asks Dr Dan if changing his diet really will decrease his risk of developing cancer.
Dr Dan proceeds with the routine colonoscopy and assures Tom he will do some research to try and alleviate his concerns.
The Information Gap
Dr Dan wants to alleviate Tom’s concerns about the risk of developing colorectal cancer and therefore wants to find a study that shows the risk of colorectal cancer in men who eat red meat compared to those who eliminate it from their diet. The conclusion of this study would allow Dr Dan to inform Tom if continuing to eat red meat would put him at a higher risk of developing colorectal cancer.
PICO Formula
Population Adults
Intervention Red Meat consumption
Comparison No Red Meat consumption
Outcome Increase in risk of colorectal cancer
The Answerable Question
Using the above PICO formula, question was formulated: “In adults, does eating red meat increase the risk of developing colorectal compared to those who do not?”
Search Strategy
The article search was performed using MEDLINE accessed through a direct connection with the Monash University Library system. It was chosen over other databases due to familiarity and ease of use. The search was achieved by first searching for individual keywords such as “Red Meat” and “Colorectal Cancer” and combining these key words with AND which produced 326 articles. These results were limited to humans, adults and those articles published in the last 15 years which produced 119 articles.
Of the 119 articles; 102 could be excluded immediately as the title did not appropriately fit the PICO components. Most of these studies focused on specific racial groups i.e. rates of colorectal cancer in Hawaiian or Japanese men which would have been too sub set for our patient.
The remaining 17 studies were mostly reviews, systematic reviews or meta-analysis which would not have effectively answered the PICO question set and were therefore excluded.
The chosen study was a prospective cohort study of 37,112 people recruited over 4 years. As a prospective cohort study, it was a relatively high level of evidence on the pyramid. Despite the study including other foods such as chicken and fish consumption, it specifically answered all aspects of the PICO question and directly addressed the question created from the information gap. A specific benefit of the chosen study was it was drawn from a Melbourne population base, making it very applicable to our patient.
Citation
Bingham, S. A. (2017). Plenary Lecture High-meat diets and cancer risk, 44(July 1998), 243–248.
Haggar, F. a, Boushey, R. P., & Ph, D. (2009). Colorectal Cancer Epidemiology : Incidence , Mortality , Survival , and Risk Factors. Clin Colon Rectal Surg, 22(4), 191–197. https://doi.org/10.1055/s-0029-1242458.
NHMRC (National Health and Medical Research Council). (2013). Eat for Health: Australian Dietary Guidelines. Archives of Pediatrics & Adolescent Medicine (Vol. 133). https://doi.org/10.1001/archpedi.1979.02130080108026
Norat, T., Lukanova, A., Ferrari, P., & Riboli, E. (2002). Meat consumption and colorectal cancer risk: Dose-response meta-analysis of epidemiological studies. International Journal of Cancer, 98(2), 241–256. https://doi.org/10.1002/ijc.10126
ARE THE RESULTS VALID?
1. How were cases defined and the controls chosen?
The cases were defined from a cohort of 41,528 residents of Melbourne recruited between the ages of 27 and 75 between the years of 1990 and 1994.
Ethics was approved by the cancer council of Victoria’s Human Research Ethics Committee before written consent to participate and access to medical records was obtained.
Exclusion criteria were implemented for anyone who had previously had colorectal cancer, diabetes, an ischemic cardiac event or angina before the baseline was obtained as their diet would not have been reflective of the general cohort. This left 37,112.
Interviews were scheduled to obtain information on other risk factors that could have impacted outcomes that were not controlled such as age, sex, alcohol consumption, physical activity, education, BMI, waist measurements and hormone replacement therapy.
Detection of the outcome was then identified from the Victorian Cancer Registry as diagnosis of adenocarcinomas of the colon and rectum in accordance with the international classification of diseases rubric and was therefore similar for both the exposed and control groups.
2. How was the exposure(s) identified across the two groups? Was this consistent across both groups?
The exposure of red meat was measured through dietary assessment over both groups. Subjects were required to complete a 121-item food frequency questionnaire originally developed as a result of a previous study into food records in 810 Melbourne residents.
Fresh red meat was defined in the questionnaire supplied to both groups as any form of pure or mixed veal, beef, lamb, pork, rabbit and game. Processed meat was included as salami, sausages, bacon, hams and manufactured lunch meats.
3. Did the investigators measure all known confounders in the cases and the controls? If so, were these adjusted for in the analyses?
Yes all known and potential confounders were measured and included if they altered the hazard ratios for the red meat consumption by more than 5%.
Sex, energy intake (kJ/d) and country of birth were included in all models and consumption of other food groups such as cereal products and fat intake were recorded and adjusted for in the analysis.
A few variables were added and then not retained in the analysis. Education, level of physical activity, calcium and folate intake, use of multivitamins, current alcohol consumption and use of hormone replacement therapy in women, BMI and waist to hip measurements were originally added but as the hazard ratios for these variables were >5% compared to models with none of these variables they were excluded from the analyses.
4. What type of case-control study was performed? Are there any limitations associated with this design? (e.g. retrospective questionnaire, record linked etc)
This was a prospective case-control study with a follow up period of 9 years. The main limitation was that red meat consumption, the main variable of the study, was measured on a self-assessment questionnaire which leaves room for recall bias.
Additionally, dietary red meat was measured on frequency of intake without portion sizes which could underestimate then level of red meat consumed.
Furthermore, family history of colorectal cancer was not taken into account, despite being a strong nonmodifiable risk for this outcome.
WHAT ARE THE RESULTS?
5. What is the odds ratio?
The event rate in the red meat exposure group was 1.3 and 0.8 in the control group. Therefore, the odds ratio was 1.625.
This means that the odds of developing colorectal cancer was 1.625 higher in the group exposed to red meat.
6. How strong is the association between exposure and outcome?
In order to assess causation for the exposure and outcome, a number of factors need to be considered including strength, consistency, specificity, temporality, biological gradient and plausibility. (Temporality and biological gradient are addressed in questions 8 and 9)
Strength – The raw odds ratio was calculated at 1.625 which is moderately high, however this figure was not adjusted for confounding factors. The study calculated a lower hazard ratio of 1.4 after adjustments.
Consistency – As this is just one study, it is hard to determine consistency, and therefore a higher statistical study such as a meta-analysis would need to be conducted.
Specificity – The specificity between the exposure and outcome would be relatively low due to large number of risk factors associated with colorectal cancer.
Plausibility – Yes the basis of the study is plausible as previous studies have associated red meat consumption with colorectal cancer (Bingham, 2017; Norat, Lukanova, Ferrari, & Riboli, 2002).
7. How precise was the estimate of the risk? (i.e what is the confidence interval for the OR?)
The confidence interval for the odds ratio was not precise.
The hazard ratio produced by the study for developing colorectal cancer as a result of red meat intake was 1.4, with a confidence interval of 95% (1-1.9).
This shows a relatively low likelihood that the outcome was indeed from the exposure.
In addition, as the P value was 0.2, the results were not statistically significant.
DO THE TESTS SATISFY SOME ‘DIAGNOSTIC TESTS FOR CAUSATION’?
8. Is it clear that the exposure preceded the onset of the outcome?
Temporality is a requirement for assessing causation. As the study excluded all patients who had previous had colorectal cancer, the cause of red meat consumption preceded the effect of colorectal cancer showing temporality.
9. Is there a dose-response gradient?
Analysis of meat consumption was arranged into quartiles of weekly frequency with all quartiles ranging between 1.4 and 1.5. As there was no quantity assessed with frequency there would be little evidence to establish a dose-dependent relationship.
10. Is there any evidence from a "dechallenge-rechallenge" strategy?
As development of colorectal cancer is an irreversible outcome, it would not be possible to use a dechallenged-rechallenged strategy for this clinical situation.
HOW CAN I APPLY THE RESULTS TO PATIENT CARE?
11. Were the study patients similar to the patient under consideration in my practice?
As my patient is a 60-year-old male in Melbourne who consumes red meat and has never previously been diagnosed with colorectal cancer, he would fit into the same category as the study patients very well.
It is worth noting that Tom does have a family history of colorectal cancer which does predispose him to developing colorectal cancer himself (Haggar, Boushey, & Ph, 2009). However, this was not one of the exclusion criteria in the study and therefore the study is still applicable to my patient.
12. What was the magnitude/seriousness of the risk?
An unadjusted calculated relative risk of 1.625 is high. This would give a very high number needed to harm of 2 which means that every second person who regularly eats red meat will develop colorectal cancer. However, the results produced by the study and therefore calculated were not precise nor significant.
13. Are there any benefits that offset the risks of exposure, or should an attempt be made to stop the exposure in the patient?
Despite the discussed risk of colorectal cancer in those who eat red meat, the Australian Dietary Guidelines do recommend (National Health and Medical Research Council, 2013) fresh red meat as part of a balanced diet. Therefore, it is always important to discuss the benefits and risks of any treatment with the patient.
As he has significant family history of colorectal cancer and there is no significant evidence in this study linking red meat to colorectal cancer there would likely be little additional harm in continuing to eat red meat. However, if Tom is particularly concerned, it could be advised to cut down in order to alleviate his anxiety.
Table: Clinically useful measures of the effects of exposure
Adverse outcome (Yes)
Adverse outcome (No)
Exposure (Yes) 451
Exposure (No)
0.8
2.1
References:
Bingham, S. A. (2017). Plenary Lecture High-meat diets and cancer risk, 44(July 1998), 243–248.
Haggar, F. a, Boushey, R. P., & Ph, D. (2009). Colorectal Cancer Epidemiology : Incidence , Mortality , Survival , and Risk Factors. Clin Colon Rectal Surg, 22(4), 191–197. https://doi.org/10.1055/s-0029-1242458.
NHMRC (National Health and Medical Research Council). (2013). Eat for Health: Australian Dietary Guidelines. Archives of Pediatrics & Adolescent Medicine (Vol. 133). https://doi.org/10.1001/archpedi.1979.02130080108026
Norat, T., Lukanova, A., Ferrari, P., & Riboli, E. (2002). Meat consumption and colorectal cancer risk: Dose-response meta-analysis of epidemiological studies. International Journal of Cancer, 98(2), 241–256. https://doi.org/10.1002/ijc.10126