Using ambulatory care sensitive hospitalisations to analyse the effectiveness of primary care services in Mexico

• We analyse the avoidable hospitalisation (ACSH) rate in Mexico during 2001–2011.
• The ACSH rate in Mexico boomed after health insurance expansion.
• ACSH rate may not be a valid effectiveness indicator when health coverage expands.
• ACSH rate may identify differences in primary care between and within states.
• The heterogeneity found in the ACSH rates suggests regional differences in quality.

Ambulatory care sensitive hospitalisations (ACSH) have been widely used to study the quality and effectiveness of primary care. Using data from 248 general hospitals in Mexico during 2001–2011 we identify 926,769 ACSHs in 188 health jurisdictions before and during the health insurance expansion that took place in this period, and estimate a fixed effects model to explain the association of the jurisdiction ACSH rate with patient and community factors. National ACSH rate increased by 50%, but trends and magnitude varied at the jurisdiction and state level. We find strong associations of the ACSH rate with socioeconomic conditions, health care supply and health insurance coverage even after controlling for potential endogeneity in the rolling out of the insurance programme. We argue that the traditional focus on the increase/decrease of the ACSH rate might not be a valid indicator to assess the effectiveness of primary care in a health insurance expansion setting, but that the ACSH rate is useful when compared between and within states once the variation in insurance coverage is taken into account as it allows the identification of differences in the provision of primary care. The high heterogeneity found in the ACSH rates suggests important state and jurisdiction differences in the quality and effectiveness of primary care in Mexico.