Using machine learning to assess the predictive potential of standardized nursing data for home healthcare case-mix classification

Abstract

Background The Netherlands is currently investigating the feasibility of moving from fee-for-service to prospective payments for home healthcare, which would require a suitable case-mix system. In 2017, health insurers mandated a preliminary case-mix system as a first step towards generating information on client differences in relation to care use. Home healthcare providers have also increasingly adopted standardized nursing terminology (SNT) as part of their electronic health records (EHRs), providing novel data for predictive modelling. Objective To explore the predictive potential of SNT data for improvement of the existing preliminary Dutch case-mix classification for home healthcare utilization. Methods We extracted client-level data from the EHRs of a large home healthcare provider, including data from the existing Dutch case-mix system, SNT data (specifically, NANDA-I) and the hours of home healthcare provided. We evaluated the predictive accuracy of the case-mix system and the SNT data separately, and combined, using the machine learning algorithm Random Forest. Results The case-mix system had a predictive performance of 22.4% cross-validated R-squared and 6.2% cross-validated Cumming’s Prediction Measure (CPM). Adding SNT data led to a substantial relative improvement in predicting home healthcare hours, yielding 32.1% R-squared and 15.4% CPM. Discussion The existing preliminary Dutch case-mix system distinguishes client needs to some degree, but not sufficiently. The results indicate that routinely collected SNT data contain sufficient additional predictive value to warrant further research for use in case-mix system design.

Development of a casemix classification to predict costs of home care in the Netherlands: a study protocol

“…Examples include short-term care for frail elderly and chronically ill (<3 months), care for terminally ill patients, and care for children. The decision tree for allocation to a single care category can be found in a recent study by De Korte et al 18…”

“…Table 2 lists all planned combinations. Previous work on similar data showed that there are clients with either a comparatively low care use (defined as <4 hours during a care episode) or high care use (defined as >40 hours per 4 weeks care episode or >120 hours per 13 weeks episode) 18. It is not uncommon to exclude such clients from a prospective payment system based on casemix classification and instead to reimburse their care based on, for example, FFS 34.…”

“…However, a rare characteristic may still be valuable if it has a significant effect on the outcome. Based on previous work,18 we expect a few hundred predictors to be left over if we initially select all NANDA-I or Omaha variables present in at least 2% of all clients. This also has the practical benefit of reducing the computational cost of the downstream modelling steps.…”

“…Using standardised (nursing) classification data for casemix classification offers plenty of potential for achieving these policy goals, without adding a considerable administrative burden to home care providers. Previous research suggests that such data, which are routinely registered in home care and stored in providers’ electronic health records (EHRs), can be used to predict casemix efficiently 18. However, there are also barriers that impede immediate adoption, including the use of multiple classification systems in Dutch home care (eg, NANDA-I, Omaha or InterRAI) and variations in registration practices 19.…”