Since 2014, Swiss home care organizations have had the possibility to use the results of regular reports on 19 quality indicators in their efforts for quality improvement. However, comparison of quality indicators between organizations should be interpreted with caution, because these indicators may also be affected by factors other than the quality of services. This problem can be addressed by different methods of risk adjustment of the quality indicators.
It is the purpose of a study founded by the Spitex association Switzerland to develop an easy administrable form of risk adjustment for the home care quality indicators, based on factors measured with RAI-Home-Care Switzerland.
Methods and Results
We use logistic regression to predict the existence of a quality problem on a client level and linear regression to predict the level of a quality indicator on an organizational level. We also use a so called agency intake profile (AIP) to predict the outcome of an indicator. The AIP is defined as the rate of clients already showing the specific quality problem at admission. Finally, we compare the impact of the different methods for risk adjustment.
The results reveal that the different methods of risk adjustment show different impacts at the level of the quality indicators and have different explanatory power for the quality indicators. Not all risk adjustment models show satisfying explanatory power, and for some indicators it is not possible to find any compelling method of risk adjustment.
For the decision about which method of risk adjustment will finally be used in the quality reports, decision makers will have to weigh the advantages and disadvantages of the different methods.