Odds Ratio: Odds ratio (OR) is sometimes reported instead of Relative Risk. This normally gives the odds or likelihood of something occurring in one group compared with another group and will be interpreted in the similar way as Relative Risk data. ie.

Question 1. Which of the subsequent best describes the research question for this study?

How effective are hands-free mobile telephones in reducing the annual incidence of car accidents?

Question 2. Which of the following best illustrates the study design used in this scenario?

a. Longitudinal cohort study
b. Cross-sectional descriptive study
c. Case-control study
d. Non-randomised experimental study
e. Pre-post study

Question 3. What level of evidence does this study design represent?

a. One
b. Two
c. Three
d. Four
e. Five

The researchers used the available survey data and summarised the nature of the car accidents and if a mobile phone was being used by the driver either during (or immediately before) the accident. This information is presented in Table 1. Use the data presented in Table 1 to answer the following questions.

Question 4.

(i) What were the total numbers of usable (complete data) survey responses in 1997 and 2007, respectively; that are reported in Table 1?

(ii) Using plain language, describe the number of respondents to the 1997 survey that reported using a mobile phone during (or immediately before) the accident? How did this number compare to the 2007 survey respondents?

(iii) Identify any ONE threat to the internal validity of the data presented in this table.

The researchers investigated whether people who used a mobile phone while driving had a greater risk of having an accident compared to those who did not use a phone when driving.

Question 5. Use the data in Table to answer the following questions.

(i) Refer to the combined data from the 1997 and 2007 surveys and explain in plain language the relative risk of having a car accident when driving while talking on a hand-held phone compared to not using a phone, including an explanation of the 95% confidence interval and statistical significance of the finding.

(ii) Refer to the combined the data from the 1997 and 2007 surveys and explain in plain language the relative risk of having a car accident when driving while talking on a hands-free phone compared to not using a phone, including an explanation of the 95% confidence interval and statistical significance of the finding.

(iii) Refer to the 1997 survey data, the 2007 survey data and the combined data from the 1997 and 2007 surveys and explain in plain language the relative risk of having a car accident when driving while talking on any type of phone compared to not using a phone, including an explanation of the 95% confidence interval and statistical significance of each finding.

In order to control for potential confounding factors, the researchers performed a logistic regression on the combined data from 1997 and 2007. Gender, age, years with a driver's licence, and driving exposure measured as annual kilometres were entered as independent variables. Logistic regression is a type of statistical analysis where the independent variables are dichotomous - this means participants fall into one category or another (e.g. Gender is either male or female; age is <25 years or ≥25 years; Years with a driver's licence is <10 years or ≥10 years; and annual km driven is <15 000 km or ≥ 15,000 km). The likelihood of a particular outcome (odds ratio /OR) is then calculated based on each category. The results in Table 3 represent the likelihood (OR) that a mobile phone was used for talking when driving a car (ever) based on the independent variable. The referent category for each variable is shown in brackets in table 3; an OR of 1.0 is assumed for each referent category. Independent variables that were found to have a statistically significant impact on the likelihood of ever driving while talking on a mobile phone are indicated by asterisks.

Question 6. (i) Using the data presented in Table, explain in plain language the impact of each independent variable on the likelihood of ever having talked on a mobile phone when driving. Use both the odds ratio and the p values to explain the findings.

Question 7. Using the information from Fig, answer the following questions:

(i) What percentage of respondents (who provided usable data) believed that they could have certainly avoided the crash if they had not being driving while talking on a hand-held mobile phone?

(ii) What percentage of respondents (who provided usable data) believed that they could probably have avoided the crash if they had not being driving while talking on a hands-free mobile phone?
Answer:

(iii) What percentage of respondents (who provided usable data) believed that they probably could not have avoided the crash if they had not being driving while talking on any type of mobile phone?