Doctoral Thesis: A longitudinal analysis of Oral Contraceptive Pill (OCP) use

Supervisors: A/Professor Gillian Heller & Dr Anne Young

University: Macquarie University

This project used data from the ALSWH, in particular the first three surveys of the Younger cohort of women. The three main objectives of this project were:
• to describe the users of contraception at Survey 1 in terms of their socio-demographic factors, health status and health risk behaviours
• to describe the long-term users of the Oral Contraceptive Pill (OCP) in terms of their socio-demographic factors, health status and health risk behaviours and
• to determine the long-term association between OCP use and health related quality of life among younger women.
A contraceptive status variable was created at each of the three surveys and categorised each woman into one of seven groups as shown in the table below.

Table 1.1. Contraceptive status for younger women who completed surveys.

Area of residence was significantly associated with contraceptive use (p<.0001), with the percentage of young women using any form of contraception at Survey 1 being higher in the rural and remote areas of the country. After adjusting for area of residence, Young women at Survey 1 using contraception were more likely to be Australian-born or from an English-speaking background (ESB), be in a defacto relationship, be older, not currently studying and had more difficulty managing on their income. Women who consumed alcohol at a risky levels and those who were current smokers were the most likely to be users of contraception.
In the second stage of the project a dichotomous indicator variable of current OCP use (Yes/No) at each of the three surveys was created. These three variables were then combined to define eight patterns of OCP usage. For ease of interpretation and due to some small sample sizes, these eight patterns were collapsed into five groups shown in the table below. The sample of women included in the analysis was restricted to women who were long-term ALSWH participants, had complete contraception data and had not had a baby (n=5161).
The patterns of use of OCP were associated with different socio-demographic characteristics and health related behaviours. For example, women who were long-term OCP users were more likely than other women to have more education; manage on their income without difficulty; be in a stable relationship (de facto or married); be in the healthy weight range; have moderate or high levels of physical activity and drink alcohol at low risk levels (not harmful to their health).

Table 1.2. OCP usage of younger women.

The third part of the project investigated whether changes in health-related quality of life between Surveys 1 and 3 were associated with levels of OCP usage. Longitudinal analysis was carried out using the method of generalized estimating equations (GEE) to analyze longitudinal measures of health-related quality of life (measured with three of the SF-36 subscales at Surveys 1, 2 and 3) in relation to OCP usage.
The results of this study showed that long-term users of OCP had equal or better quality of life according to the General Health (GH), Mental Health (MH) and Vitality (VT) subscales of the SF-36, than other women their age, once adjustments were made for area of residence, comorbidity, level of education and smoking. The inclusion of the OCP user variable in each of the three models was significant thus indicating that there is a general effect of OCP use. That is, there is on average a difference in MH, GH and VT between the different groups of women defined by their OCP use.
There are some limitations to this study which include missing data (due to women failing to answer some questions in the survey), and inconsistencies with the data (e.g. a participant may not have followed the instructions correctly or may have contradicted themselves across time). Another limitation is that the data are obtained through self-report at only three timepoints in a 7-year period and this analysis makes the assumption that their self-reported behaviours have continued, which they may or may not have done. Also, due to initial recruitment and retention bias, a number of minority groups may be underrepresented in this study. Finally, another limitation is that variables to reflect transitions over time could have been investigated more thoroughly for their inclusion as covariates in the model. More complex longitudinal models to incorporate changes over time for these young women could be explored. The potential association of long-term OCP use and adverse outcomes such as cancer could not be studied yet due to the lag time in the development of these conditions.
Until now, there has been very little evidence regarding young Australian women’s OCP usage and its association with long-term health. The results of this study indicate that among women who have not had a baby, use of OCP may carry some benefits in terms of quality of life and this study has not shown any detrimental effects of OCP use among these younger women.