Index-based inequality in quality of care: An empirical comparison of apples and pears

Purpose: Socioeconomic inequalities have been studied for decades using a variety of methods, but limited attention has been paid to the way methodological differences influence research findings. We aimed to compare index-based measures of socioeconomic inequality in quality of care. Patients and Data: A national cohort of 110,848 unique stroke patients admitted to publicly funded hospitals in Denmark from 2004–2014 was studied. We used individual-level data from national registers and the Danish Stroke Registry. Quality of care was defined as fulfilment of process performance measures based on clinical guidelines recommendations (range 0–100%). Socioeconomic position was operationalised using information on dispo-sable family income (continuous, DKK) and education (categorical, 7 levels). Methods: Income-and education-related inequality in quality of care was assessed using concentration indices and the slope index of inequality. All indices were estimated both in absolute and relative terms. Results: Income-related inequality appeared to be generally higher than education-related inequality. Depending on the choice of index, the inequality in quality of care increased by 5% or declined by up to 43% during the study period. Unlike the concentration indices the slope index of inequality was highly sensitive to changes in how the income and educational levels were operationalised. Conclusion: Careful reporting and interpretation of inequality studies is warranted in order not to misguide decision makers. Unless the policy objective reflects an explicit focus on one specific type of inequality, the use of different inequality indices can lead to different conclusions.