Feasibility of 30-day hospital readmission prediction modeling based on health information exchange data
Health Information Exchange (HIE) allows important healthcare information to be transmitted within and between states, as established by the Health Information Technology for Economic and Clinical Health (HITECH)Act of 2009. Such information is rarely used to assess patient’s risk of readmission, and a reason for this is that it is unclear whether it is accurate or feasible to do so.
A literature review was done with PubMed, Google Scholar, and Embase. Some inclusion criteria were that the articles must be peer reviewed, focused on 30 day readmission time, and list predictor variables for readmission (either all-cause or specific cause). Exclusion criteria included a focus on pediatric or psychiatric readmission, or lack of variables in a prediction model. Data completeness of RRPM- HL7 HL7 FHIR segments were acquired from three Health Information Organizations (HIOs).
From 309 search results, only 32 articles were included in the review. From these, 297 predictors for readmission were found. These predictors were “mapped” or sorted into HL7 segments. Three HIOs were analyzed regarding their coverage level of these HL7 segments, and the three had varying levels of coverage for these.
Although the three HIOs which were analyzed had varying levels of coverage, the information accessed through HIEs can be overall very useful for healthcare quality assurance purposes, such as calculating a patient’s risk of readmission.
As a public health professional, it is very exciting to see the possibilities that population health information, such as that acquired through HIEs, can bring to improve healthcare quality and patient safety. This article was a very interesting read, although I would have liked for the methods to be a little more specific, such as having the names of the HIOs which participated in the study, and having more detail about how they were analyzed. 
- Swain and Kharrazzi 2015. Feasibility of 30-day hospital readmission prediction modeling based on health information exchange data. http://www-ncbi-nlm-nih-gov.ezproxyhost.library.tmc.edu/pubmed/26412010