Inequality in Quality? The selection and use of quality indicators to investigate ethnic disparities in the quality of hospital care, Aotearoa New Zealand.

Abstract

There are well documented differences in health outcomes between Māori and New Zealand (NZ) Europeans. Jones (2002) describes differential treatment within the health system as one determinant of ethnic inequalities: is it possible that New Zealand’s health services contribute to the differences in health status between Māori and NZ Europeans?

Aim and objectives: This thesis describes an investigation into the quality of care for Māori compared with NZ Europeans in public hospitals nationally. The objectives of this study were:

1. To identify measures applicable to this study context with validity as indicators of the quality of health care.

2. To employ this/these measure(s) to compare the quality of inpatient hospital care between NZ Māori and NZ European patients, with consideration of confounding and mediating factors in order to estimate the net effect of ethnic group on the quality indicator.

3. To offer recommendations in light of the findings of this study.

Methods: Literature review and three ‘study context’ criteria were used to select two indicators to represent inpatient quality of care - unplanned readmission/death within thirty days of discharge (‘readmission’) and patient satisfaction.

Phase One of the research used data from the National Minimum Data Set to calculate and compare the rate of readmission for Māori and NZ European inpatients at NZ public hospitals. Characteristics of the two ethnic groups were compared with age-sex adjusted proportions, and variation in the likelihood of readmission with patient and clinical factors was explored with rate ratios. The odds of readmission for NZ Māori compared to NZ European patients (n=89,090) were calculated from a logistic regression model, with variables representing age, comorbidity, index procedure, hospital volume and socio-economic position included.

In Phase Two, Māori and NZ Europeans recently discharged from one of three NZ hospitals were approached to complete the Client Satisfaction Questionnaire-8 (CSQ-8). Descriptive analyses explored the characteristics of the respondents (n=1103) according to ethnic group and mean satisfaction score. A linear regression model including variables for age and health status estimated the difference in mean CSQ-8 score for Māori compared to NZ European respondents.

Results: The Phase One analyses found 16% higher odds of readmission for NZ Māori compared to NZ European patients (odds ratio (OR) 1.16, 95% CI 1.08 – 1.24; adjusted for age, index procedure, comorbidity, hospital volume, and deprivation), and 19% higher odds (OR 1.19, 95% CI 1.11 – 1.27) when the model did not include a deprivation term. Readmission was also associated with older age (OR 1.33; 95% CI 1.19-1.48, for >79 yrs compared with 18-39 yrs), higher comorbidity (OR 2.08; 95% CI 1.89-2.31 for Charlson score 3+ compared with 0) and higher hospital volume (OR 0.81; 95% CI 0.76-0.86 for lowest volume facility compared with highest). Measurement error of quality of care by readmission was the primary source of bias in this phase; sensitivity analyses suggest the contribution of ‘poor quality’ to the increased odds of readmission for Māori may be small. That is, unmeasured factors may have a comparatively greater role than quality of care in the ethnic difference of this outcome.

The Phase Two multivariable model showed comparable satisfaction for Māori and NZ European respondents, with the difference in mean scores only -0.02 (95% CI -0.36 - 0.57). However, bias from differential non-response is possible – participation for Māori was 37% compared to 60% for NZ Europeans. These results may also be affected by differential or non-differential measurement error. That is, CSQ-8 score may have lower validity as a measure of health care quality in this setting and population.

Conclusions: A valid measurement of quality by readmission or satisfaction is difficult, as both are highly vulnerable to error. In particular, ethnic differences in readmission may be predominantly influenced by factors other than the inpatient quality of care. However, given supporting evidence and the plausibility of quality as a component cause for health outcomes inequalities, it is likely that the increased odds of readmission for Māori compared to NZ Europeans is in part due to poorer quality of care. This study recommends protocols be developed to guide the calculation and interpretation of readmission as a proxy for quality, and suggests further research to explore the measurement of patient satisfaction in the NZ setting.