The current set of analyses was commissioned by the Australian Aged Care Quality Agency (AACQA). In November 2018, the AACQA approached the AIPCA to carry out an analysis of CER data:
We are seeking a sense of the variance of and between services and the overall variance of the sector wide results. As discussed, please apply either parametric (where possible) or non-parametric methods to perform the analyses.
The data set provided included ratings from interviews held between 9 May 2017 and 4 July 2018. This data set included data from:
The modal number of people interviewed in each home was 15, with a range from 2 to 36.
New South Wales provided the largest number and proportion of homes, followed by Victoria. Very small numbers of homes in the ACT and the Northern Territory were included.
The numbers of large and medium-sized homes were similar, with a minority of homes classified as small (6.1%).
Almost two-thirds of interviewees were selected at random, and 88.2% were residents rather than representatives.
Levels of missing data were low (less than 1.5% for all questions). The question with the most missing data was the last question: Q10: I am encouraged to do as much as possible for myself.
The two questions with the lowest levels of agreement were:
Q6: Do you like the food here?
Q7: If I’m feeling a bit sad or worried, there are staff here who I can talk to.
Spread of responses to questions varied between questions, with very small spread for question 2 (Do you feel safe here?)
Formal measures of spread proved not very useful. Instead, percentage of agreement was compared across groups. This analysis showed:
Men tended to give less positive responses than women
People with dementia said they were not encouraged to do as much as possible for themselves.
People who were not independently mobile gave much lower ratings to all questions than those who were.
Small-sized facilities received higher ratings on every question than medium-sized facilities, who again scored better than large facilities.
Care recipients tended to give higher ratings than representatives, but lower ratings for food.
Services differed on their modal responses and spread of responses.