Question

Investigators wishes to examine if delaying the age of having first child is associated with higher incidence of breast cancer in young women 18-40 years old.

1. Propose an epidemiological study design to explore this question and justify why you chose this design?
2. Assuming you are one of the investigators – please provide a short description about the participants in the study, what data you will need to collect to answer your research question and at which time point you collect this data
3. What other study design you can consider for this research question and explain why you chose it  

Answer 1

Epidemiological study design :

a. For the present study, an interview based account of first incidence of breast cancer in women within the age of 18 - 40 and above the age of 40 can be conducted. For this, data about their marital status, age, age at first child birth and age at first diagnosis of cancer can be conducted. For control, women in similar age groups as the experiment group can be taken who are admitted for causes other than cancer. Then cancer risk of experimental group and the control group can be compared to reach a definitive conclusion.

This study is chosen because we can get error free data from the hospitals on their diagnosis and also we can easily get access to control groups. This method is simple and can be easily analysed.

b. The breast cancer cases must include women who were hospitalised for the first diagnosis of breast cancer at or above the age of 40. For control the patients hospitalised for any other disease can be considered. The control and experiment groups must be interviewed about their life-history. Details about their date of birth and the date of birth of first child must be obtained. Total number of births given by each patient should be obtained. In this, abortions may be neglected as first child births. Still child birth of 5 month or older foetus can be considered. Single and married women should be included.

Births at 5 months or above the 5 months age can be grouped under the class of parous. Thus women with a single chid can be under parity of 1 and women with 2 children can be under parity 2 and so on. Other cases are said to be non-parous which include abortions or child birth below 5 months age.

Analysis :

The risk for women of any given parity relative to that of non-parous women can be calculated using the formula ad/bc where a is the number of cancer cases of that parity, b is the number of controls in the same parity, c and d are the number of non-parous cases and controls, respectively. The number of cases of parous women can be compared with the total number of cases (including number of parous and number of non-parous). The number of births happening between 18-40 years should be analysed separately. The number of births occurring after the age of 40 years should be analysed separately against cancer risk.

c. Another study design for this could be the replacement of interview based data collection with direct data collection of screening tests of the experimental and control groups. We can take the cancer burden or relative risk calculated from screening tests of women who were diagnosed with cancer through subsequent tests after screening and women who weren't diagnosed with cancer during screening or after subsequent tests. The women will belong to the age of 18 - 40 or above 40 years.

This choice of study design is more accurate and reliable since interviews can be unreliable and include several personal bias. In addition, the data from screening and subsequent diagnostic tests collected will be strong and free from errors.