Generalisability of pharmacoepidemiological studies using restricted prescription data

Area of Science:

• Epidemiology
• Health Services Research
• Oncology

Background:

• Population-based pharmacoepidemiology research relies on linking medication and disease data.
• Irish dispensing data is limited to the means-tested government-medical cards scheme, potentially affecting generalizability.
• Previous research identified gender as a predictor of medical card status.

Purpose of the Study:

• To compare women with and without medical cards at the time of ovarian cancer diagnosis.
• To investigate factors associated with medical card status in ovarian cancer patients.
• To assess the impact of medical card status on pharmacoepidemiological analyses.

Main Methods:

• Ovarian cancers diagnosed between 2001-2010 were identified from the National Cancer Registry Ireland.
• Logistic regression was used to evaluate associations between demographic and socioeconomic factors (age, region, deprivation, smoking, employment, marital status) and medical card status at diagnosis.
• Cumulative incidence of new medical card receipt post-diagnosis was assessed.

Main Results:

• Of 3396 women diagnosed with ovarian cancer, 1778 (52%) had a medical card at diagnosis.
• Medical card prevalence at diagnosis was 33% for women under 70 and 87% for women 70 and older.
• For women under 70, all evaluated factors were significantly associated with medical card status at diagnosis.
• 52% of women without a medical card at diagnosis obtained one post-diagnosis.

Conclusions:

• Medical card coverage among ovarian cancer patients mirrors the general population.
• Several factors predict medical card status, particularly in women under 70.
• External validity of pharmacoepidemiological studies using Irish dispensing data requires careful consideration due to these predictive factors.