The effect of Māori ethnicity misclassification on cervical screening coverage

There is a large difference in the cervical screening coverage rate between Māori and European women in New Zealand. This paper examines the extent to which this difference is due to misclassification of ethnicity.

Data from Waitemata District Health Board's two Primary Health Organisations (PHOs) was used to identify the population of Waitemata domiciled women aged 25-69 years eligible for cervical screening. Their cervical screening status was obtained from the National Cervical Screening Programme register (NCPS-R). Data from Auckland and Waitemata DHBs was used to determine the women's ethnicity in the National Health Index (NHI). Women who had withdrawn from the NCSP-R, women who were deceased and women for whom an NHI ethnicity code could not be obtained were excluded from the analysis. Ethnicity codes from the three sources (PHO registers, NCSP-R and NHI) were compared to identify women classified as non-Māori in the NCSP-R but Maori in either of the other two data sources. The effect on Maori cervical screening coverage rates of not counting these women was assessed.

Within the study population there was a total of 6718 women identified as Māori on the NCSP of whom 5242 had been screened within the last 3 years and 1476 who had not. In addition to these, there were 2075 women identified as Māori in either the PHO or NHI databases but not in the NCSP-R who had been screened within the preceding 3 years, and a further 2094 who had not been screened. There were also 797 women identified as Maori in the NHI or PHO datasets who were not on the NCSP-R (and therefore were not screened). If all screened women classified as Māori from any source were counted, Waitemata DHB's Māori screening coverage rate would rise from 49.3% to 68.8% (or to 61.0% and 63.2% respective if just PHO and NHI Māori were counted).

Misclassification of ethnicity could explain (in absolute terms) up to 19.5% of the 35.0% difference in cervical screening coverage rate between Māori and non-Māori , non-Pacific, non-Asian coverage in Waitemata District. Misclassification is likely to have similar effects on coverage estimates throughout New Zealand. Without improving the accuracy of ethnicity data in the NCSP-R it will be impossible for the country to achieve the target coverage rate of 80% among Māori.