Lindsay Ann Abrigo

Week 5 Assignment – Smoking and Lung Cancer Case Study

April 11, 2013

  1. What makes the first study a case-control study?

The first study is considered a case-control study because a case-control study is used when the population under investigation is large and not well defined. It is used generally when there is a communitywide outbreak where participants are generally selected based on the presence or absence of the disease outcome and investigators will then work backwards from the effect to find the cause. The patients who were hospitalized for the disorders other than lung cancer represented the controls, which were the comparison group necessary for analysis in a case-control study.

  1. What makes the second study a cohort study?

The case-subjects represent a small, well-defined population, or cohort, and they are categorized accordingly based on their exposure to a specific risk factor, which in this case would be smoking cigarettes. In this study, the rate of death from lung cancer for smokers was compared with the rate for nonsmokers. This is considered a cohort study, specifically a prospective cohort study, because individuals were followed overtime to determine the occurrence of disease.

  1. Why might hospitals have been chosen as the setting for this study?

Hospitals may have been chosen as the setting for this study because there is a higher likelihood to find lung cancer patients in hospitals receiving care and treatment; there is greater potential to match controls with case-subjects; it is more convenient and provides more controls; there is a higher probability for accurate diagnosis of lung cancer; hospital records will be readily available to investigators; and patients may be more willing to participate.

  1. What other sources of cases and controls might have been used?

Investigators could have used other sources as a means of choosing cases and controls. For selecting cases, they could have used cancer registries, death certificates, insurance files, doctors’ offices, and medical records. For selecting controls, they could have resorted to neighbors, friends, and acquaintances of the case-subjects; patients who may see the same doctors as the case-study patients but have other illnesses; and/or a random sampling from the population that is not ill.

  1. What are the advantages of selecting controls from the same hospitals as cases?

The advantages of selecting controls from the same hospitals as cases mean that the controls from the same hospital are likely to come from the same population as the case-subjects. This would ultimately control for variables such as socioeconomic status, access to health care, and environmental conditions. Other advantages include increased convenience for all involved,increased participation by those patients who are being treated and cared for in the hospital, and similar medical record maintenance methods.

  1. How representative of all persons with lung cancer are hospitalized patients with lung cancer?

The hospitalized patients with lung cancer are very representative of all persons with lung cancer because they are likely patients who have a past history of frequent smoking and were also likely to be hospitalized because of the adverse health effects.

  1. How representative of the general population without lung cancer are hospitalized patients without lung cancer?

The hospitalized patients without lung cancer are not very representative of the general population who do not have lung cancer because hospitalized patients are more than likely to smoke than the general population.

  1. How may these representativeness issues affect interpretation of the study’s results?

If case-subjects and controls do not closely represent the intended population, the results of the study will be biased, meaning that the relationship between the exposure and the health outcome will be over- or underestimated. The control group provides the prevalence of exposure for people without the health problem under study. Because the control population in this study tends to smoke more than the population it represents, the exposure prevalence among controls will be higher than expected. The net effect of this inconsistency is to underestimate the true risk of lung cancer associated with smoking.

  1. From this table, calculate the proportion of cases and controls who smoked.

Proportion of case-subjects who smoked: 1,350/1,356 = 0.99557522 = 99.5%

Proportion of controls who smoked: 1,296/1,357 = 0.9550479 = 95.5%

  1. What do you infer from these proportions?

According to these proportions, I can infer that the prevalence of smoking is really similar in both groups (99.5% vs. 95.5%) and also extremely high in both groups (>95%)

  1. A) Calculate the odds of smoking among the cases.

Odds of smoking for case-subjects: 1,350/7 = 192.857143 = 192.9/1

B) Calculate the odds of smoking among the controls.

Odds of smoking for controls: 1,296/61 = 21.2459016 = 21.2/1

  1. Calculate the ratio of these odds. How does this compare with the cross-product ratio?

Odds Ratio: (1,350/7)/(1,296/61) = 192.857143/21.2459016 = 192.9/21.2 = 9.0990566 = 9.1

  1. What do you infer from the odds ratio about the relationship between smoking and lung cancer?

From the odds ratio about the relationship between smoking and lung cancer, one can infer that the odds of being a smoker is 9.1 times higher for case-subjects (those who smoke cigarettes) than for controls (nonsmokers). One can also infer that the risk of lung cancer is 9 times higher for cigarette smokers than for nonsmokers.

Table 2 shows the frequency distribution of male cases and controls by average number of cigarettes smoked per day.

  1. Compute the odds ratio by category of daily cigarette consumption, comparing each smoking category to nonsmokers.

1 – 14 cigarettes: (565/7)/(706/61) = (565 x 61)/(706 x 7) = 34,465/4,942 = 6.97389721 = 7.0

15 – 24 cigarettes: (445 x 61)/(408 x 7) = 27,145/2,856 = 9.50455182 = 9.5

25+ cigarettes: (340 x 61)/(182 x 7) =20,740/1,274 = 16.2794349 = 16.3

All smokers: (1,350 x 61)/(1,296 x 7) = 82,350/9,072 = 9.07738095 = 9.1

  1. Interpret these results.

The results show that the odds of acquiring lung cancer increases as the daily number of cigarettes smoked also increases.

  1. What are the other possible explanations for the apparent association?

Other possible explanations for the apparent association between smoking and lung cancer may include information or selection bias, confounding, an error on the part of the investigator, or simply chance, although the data indicates that chance is not likely.

  1. How might the response rate of 68% affect the study’s results?

A response rate of 68% can affect the study’s results by introducing selection bias into the study and can potentially lead to conclusions that are systematically different from the truth. In this particular case, the bias would cause an underestimation of the risk of lung cancer that is associated with smoking.

  1. Compute lung cancer mortality rates, rate ratios, and rate differences for each smoking category. What do each of these measures mean?

Mortality rate per 1,000: Lung cancer mortality rates increase as the amount of cigarettes smoked per day also increases.

1 – 14: 22/38,600 = 0.00056995 x 1,000 = 0.56994819 = 0.57

15 – 24: 54/38,600 = 0.00138817 x 1,000 = 1.38817481 = 1.39

25+: 57/25,100 = 0.00227092 x 1,000 = 2.27091633 = 2.27

All smokers: 133/102,600 = 0.0012963 x 1,000 = 1.2962963 = 1.30

Total: 136/145,400 = 0.00093535 x 1,000 = 0.93535076 = 0.94

Rate ratio: When comparing the categories 1 – 14 and 0 cigarettes a day, the rate ratio is 8.1. As a result, people who smoke 1 – 14 cigarettes a day are 8.1 times more likely than nonsmokers to develop lung cancer. In comparing the categories 15 – 24 and 0 cigarettes a day, the rate ratio is 19.8, thus people who smoke 15 – 24 cigarettes a day are 19.8 times more likely than nonsmokers to develop lung cancer. People who smoke 25+ cigarettes a day are 32.4 times more likely than nonsmokers to develop lung cancer.

1 – 14: 0.57/0.07 = 8.14285714 = 8.1

15 – 24: 1.39/0.07 = 19.8571429 = 19.8

25+: 2.27/0.07 = 32.4285714 = 32.4

All smokers:1.30/0.07 = 18.5714286 = 18.6

Rate difference: The excess death attributed to smoking increases as the quantity of cigarettes smoked per day also increased. It increased from 0.50 to 2.20, and in sum causes 1.23 excess deaths per 1,000 people per year.

1 – 14: 0.57-0.07 = 0.50

15 – 24: 1.39-0.07 = 1.32

25+: 2.27-0.07 = 2.20

All smokers: 1.30-0.07 = 1.23

  1. What proportion of lung cancer deaths among all smokers can be attributed to smoking? What is the proportion called?

AR% = (1.30 ! 0.07) / 1.30 = 0.946 x 100% - 94.6%

Approximately 95% of the deaths due to lung cancer in the exposure group (all smokers) may be attributable to smoking. This proportion is known as the attributable risk percent and can also be interpreted as the proportion of lung cancer deaths that could have been prevented among smokers if they had not smoked.

  1. If no one had smoked, how many deaths from lung cancer would have been averted?

In sum, 95% of the lung cancer deaths among smokers are attributable to smoking. If no one had smoked, 95% of the 133 deaths, which would be approximately 126, due to lung cancer would have been averted.

The cohort study also provided mortality rates for cardiovascular disease among smokers and nonsmokers. The following table presents lung cancer mortality data and comparable cardiovascular disease mortality data.

  1. Which cause of death has a stronger association with smoking? Why?

The rate ratio is the primary means of measuring association. As such, it indicates a stronger association between smoking and lung cancer (18.5) than between smoking and cardiovascular mortality (1.3). Smokers are more likely to die from lung cancer than from cardiovascular disease, as suggested by the data.

  1. Calculate the population attributable risk percent for lung cancer mortality and for cardiovascular disease mortality. How do they compare? How do they differ from the attributable risk percent?

PAR% (lung cancer) = (0.94 ! 0.07) ÷ 0.94 = 0.925 x 100% = 92.5%
PAR% (cardiovascular disease) = (8.87 ! 7.32) ÷ 8.87 = 0.174 x 100% = 17.4%

Therefore, 92.5% of all deaths due to lung cancer and 17.4% of all deaths due to cardiovascular disease in the study population are attributable to smoking. From a prevention perspective, the PAR% for a given exposure can be interpreted as the proportion of lung cancer cases in the entire study population that would have been prevented if exposure had not occurred.

  1. How many lung cancer deaths per 1,000 persons per year are attributable to smoking among the entire population? How many cardiovascular disease deaths?

Lung cancer: 0.925 x (0.94) = 0.87 deaths per 1,000 person years

Cardiovascular disease: 0.174 x (8.87) = 1.54 deaths per 1,000 person years

The number of smoking-related deaths per 1,000 people per year is greater for cardiovascular disease than for lung cancer, even though the relative risk (rate ratio) is considerably lower. Thus, if no one smoked, more deaths would be prevented from cardiovascular disease than from lung cancer.

The following table shows the relationship between smoking and lung cancer mortality in terms of the effects of stopping smoking.

  1. What does these data imply for the practice of public health and preventive medicine?

The data in the table above shows that smokers have the highest risk of death from lung cancer, and nonsmokers have the lowest. Although the risk for people who have stopped smoking decreases over time after quitting, the risk is still almost three times that of nonsmokers. These data implies, that from a public health and preventive medicine perspective, smoking cessation efforts are worthwhile because of the reduction of risk, but preventative efforts are most valuable and most effective because of the lower risk associated with not smoking at all.

  1. Compare the results of the two studies. Comment on the similarities and differences in the computed measures of association.

Results from the studies are very consistent, exhibit a strong association between cigarette smoking and lung cancer, and include evidence of a dose-response relationship.

Compared with the rate ratios from the cohort study, the odds ratios in the case-control study consistently underestimate the strength of association. This discrepancy is most likely due to the fact that the controls in this study were also very likely to be smokers.

  1. What are the advantages and disadvantages of case-control vs. control studies?

Case-ControlCohort

Sample SizeSmallLarge

CostsLessMore

Study timeShortLong

Rare diseasesAdvantageDisadvantage

Rare exposureDisadvantageAdvantage

Multiple exposuresAdvantageDisadvantage

Multiple outcomesDisadvantageAdvantage

Progression, spectrum of illnessDisadvantageAdvantage

Disease ratesCannot measureAdvantage

Recall biasPotential problemLess of a problem

Loss to follow-upAdvantagePotential problem

Selection biasPotential problemLess of a problem

  1. Which type of study (cohort or case-control) would you have done first? Why? Why do a second study? Why do the other type of study?

A case-control study is quicker, easier, and less costly. If the case-control study results reveal a significant relationship between exposure and disease, then it is appropriate to do a second study to confirm the findings. Although a cohort study is more difficult and expensive to mount, as well as slower to yield results, it can confirm the results of previous studies, allow for the calculation of disease rates, and enable investigators to assess the natural progression from exposure to disease.

  1. Which of the following criteria for causality are met by the evidence presented from these two studies?

YESNO

Strong associationX

Consistency among studiesX

Exposure precedes diseaseX

Dose-response effectX

Biologic plausibilityX