Question

A multi-disciplinary team in the US formulates a hypothesis that links intimate partner violence (IPV) during pregnancy and damage to the fetal brain. They define the outcome of interest as the presence of abnormal findings including blood clot in the fetal brain detected by ultrasound during pregnancy. All infants included in the study later undergo neurological examination after birth to confirm prenatal (before birth) ultrasound observations. To determine whether a woman was actually a victim of intimate partner violence, the investigators used a questionnaire. Respondents were initially asked whether they had ever experienced any violence from their spouse. Based on more detailed questions the violence experienced was categorized into the following subtypes:

1. Physical violence included instances of pushing/shoving, throwing objects, slapping, twisting arm, punch, hitting with an object, kicking, dragging, attempting to strangle or burn, threatening with a weapon, attacking with a weapon;
2. Emotional violence included verbal or physical public humiliation, verbal threat to the woman or her family;
3. Sexual violence referred to being forced to have sex or being forced to perform sexual acts.

The questionnaire was originally validated and found to capture 4 of every 5 cases of true IPV in a random sample comprising 500 pregnant women.

The US investigators proceeded with their longitudinal study and enrolled 300 pregnant women among whom 30% screened positive on ultrasound. Of the total 150 pregnant women with negative history of IPV, 30 had abnormal fetal brain ultrasound findings. Using odds ratio, calculate the association between IPV during pregnancy and intra-uterine fetal brain damage.

1. Using odds ratio, calculate the association between IPV during pregnancy (as assessed by questionnaire) and intra-uterine fetal brain damage (as assessed by ultrasound)
2. Is IPV during pregnancy associated with fetal brain damage?
3. Explain your response for question 2.
4. Based on the information provided, construct the true 2-by-2 table, working backwards from the misclassified table (misclassification of IPV). (You are given information regarding the sensitivity of the questionnaire. Since you are not given any information on specificity, assume it is 100%)

5.Now with your new true 2-by-2 table without misclassification of IPV, calculate the association between IPV during pregnancy and intra-uterine fetal brain damage (use the OR as your measure of association)

6. Now with your new true 2 by 2 table without misclassification of IPV, is IPV during pregnancy associated with fetal brain damage?

      7. What is the effect of disease misclassification on the odds ratio?

Answer 1

Data:

IPV	Brain Damage	No Brain Damage	Total
Yes	60	90	150
No	30	120	150
Total	90	210

1.

Odds Ratio = 60*120/90*30

= 2.667

2.

Yes

3.

Women with IPV during pregnancy are 2.667 times more likely to have a fetus with brain damage as Odds ratio describes the likelihood of happening an event in one group as compared to another group.

4.

Sensitivity = 4/5 = 0.8

Actual no. of women with IPV = 150/0.8

= 187.5

~ 188

But some data is missing regarding how many of the women with a negative history of IPV (which now are 112) will have brain damage findings.

Incomplete table which I could produce:

IPV	Brain Damage	No Brain Damage	Total
Yes			188
No			112
Total	90	210

Please obtain the required data and fill in the table accordingly and other parts can be solved as I have solved above.

Formula odds ratio:

Please upvote if you have liked my answer, would be of great help. Thank you.